Articles cacy of venlafaxine, medroxyprogesterone acetate, and cyproterone acetate for the treatment of vasomotor hot fl ushes in men taking gonadotropin-releasing hormone analogues for prostate cancer: a double-blind, randomised trial Jacques Irani, Laurent Salomon, Rostand Oba, Philippe Bouchard, Nicolas MottetSummary Background Hot fl ushes are the most common complaints reported by men undergoing androgen suppression Published Online treatment for prostate cancer. We designed a randomised double-blind trial to compare the effi cacy of three drugs, December 7, 2009 each of which has proven eff ective for preventing hot fl ushes in previous studies.
See Online/Refl ection and Methods Men with prostate cancer with an indication for androgen suppression were enrolled in the study at Reaction 106 urology centres in France between April 14, 2004, and April 20, 2007. All patients were treated for 6 months with DOI:10.1016/S1470- leuprorelin (11·25 mg). At month 6, patients who spontaneously asked for treatment, or those who presented with 2045(09)70349-3 14 hot fl ushes or more during the week before the visit, were randomly assigned to either venlafaxine 75 mg daily, Urology Unit, University medroxyprogesterone acetate 20 mg daily, or cyproterone acetate 100 mg daily. All patients received two Hospital, Poitiers, France
(Prof J Irani MD); Urology Unit, indistinguishable pills in the morning and one in the evening from week 1 to week 8, and one indistinguishable pill University Hospital, Créteil, in the morning from week 9 to week 10, to comply with the double-blind design. Random assignment with a block France (Prof L Salomon MD); size of three was done centrally, by fax, and each patient was given a randomisation number. The allocation sequence Medical Department, Takeda was stratifi ed by centre. Assessment was done at inclusion, at randomisation, and then at 4 weeks, 8 weeks, and Laboratories, Puteaux, France
(R Oba MSc); Endocrinology 12 weeks after randomisation. Participants completed a daily hot-fl ush diary for 1 week, and a quality of life Unit, Saint Antoine University questionnaire before each visit throughout the study. The primary outcome was the change in median daily hot-fl ush Hospital, Paris, France score between randomisation and 1 month. All patients who received at least one study treatment dose were included (Prof P Bouchard MD); and in the effi cacy analysis. This trial is registered with ClinicalTrials.gov, number NCT01011751. Urology Unit, Clinique Mutualiste, Saint Etienne, France (N Mottet MD) Findings Of the 919 men initially enrolled, 311 were randomly assigned to one of the study treatments at 6 months: Correspondence to: 102 to venlafaxine, 101 to cyproterone, and 108 to medroxyprogesterone. 309 patients were included in the effi cacy Prof Jacques Irani, Urology Unit, analysis, since two were excluded for protocol deviations (one in the cyproterone and one in the medroxyprogesterone University Hospital, 2 rue de la group; both were excluded because they were already undergoing treatment with serotonin reuptake inhibitor Miletrie, 86 000 Poitiers, France [email protected] antidepressants at randomisation). The change in median daily hot-fl ush score between randomisation and 1 month was –47·2% (IQR –74·3 to –2·5) in the venlafaxine group, –94·5% (–100·0 to –74·5) in the cyproterone group, and –83·7% (–98·9 to –64·3) in the medroxyprogesterone group. The decrease from baseline was signifi cant for all three groups (p<0·0001). Pairwise comparison of treatment groups adjusted by the Bonferroni method confi rmed that the decreases in hot-fl ush score were signifi cantly larger in the cyproterone and medroxyprogesterone groups than in the venlafaxine group, regardless of the interval considered (p<0·0001 in all cases). There was no signifi cant diff erence between the cyproterone and medroxyprogesterone groups (p>0·2 in all cases). Serious side-eff ects occurred in four, seven, and fi ve patients in the venlafaxine, cyproterone, and medroxyprogesterone groups, respectively, of which none, one (dyspnoea), and one (urticaria) were considered related to the drug, respectively. Interpretation After 6 months of treatment with leuprorelin, venlafaxine, cyproterone, and medroxyprogesterone proved to be eff ective in reducing hot fl ushes. However, the hormonal treatments cyproterone and medroxyprogesterone were signifi cantly more eff ective than venlafaxine. As cyproterone is a recognised treatment in prostate cancer, and its use could interfere with hormonal therapy, medroxyprogesterone could be considered to be the standard treatment for hot fl ushes in men undergoing androgen suppression for prostate cancer. Funding Takeda Laboratories, Puteaux, France. Introduction
using gonadotropin-releasing hormone (GnRH)
Androgen suppression is the most common treatment analogues.1 Leuprorelin, like most GnRH analogues, is for advanced prostate cancer, and can be achieved administered subcutaneously two to four times a year surgically or, as is most common in the west, by blockade
depending on the injected dose, and this schedule is
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles
associated with a good tolerance of the treatment. fl ushes, more than half report a signifi cant eff ect on their However, side-eff ects related to androgen depletion are quality of life.4,5inevitable, the most frequent of which are erectile
Most treatments for hot fl ushes have been assessed in
dysfunction and the occurrence of hot fl ushes.
postmenopausal women being treated for breast cancer.
Hot fl ushes are characterised by a subjective sensation Apart from non-medical treatments, the effi
of a rise in temperature in the face and trunk and by the has not been shown in prospective trials,6–8 the currently occurrence of a cutaneous vasodilatation predominantly available therapeutic options can be divided into hormonal on the face, throat, and extremities, usually followed by and non-hormonal treatments. Low-dose oestrogens have profuse sweating. Hot fl ushes fi rst occur a few months proven to be eff ective in women, but side-eff ects such as after the onset of GnRH agonist treatment, and each raised frequency of cardiovascular events and potential episode usually lasts less than 5 min. The patho-
prostate tumour growth or gynaecomastia are highly
physiological mechanism of hot fl ushes is now well undesirable in men.9 Hormonal treatment can also be defi ned: the steep decline in the serum level of sexual achieved with steroidal anti-androgens.10,11 Cyproterone hormones, following treatment with a GnRH agonist, acetate is probably the steroidal anti-androgen used most leads to the release of hypothalamic catecholamine, widely in Europe, with a reasonable effi
particularly norepinephrine, that fl ood the thermo-
profi le. Progestagens, such as megestrol acetate or
regulation centre located in the upper hypothalamus. medroxyprogesterone, have also been shown to be eff ective This results in an abnormal and poorly regulated in men treated for prostate cancer.7,12 Amongst peripheral vasodilatation and the occurrence of hot non-hormonal treatments, the most commonly used are fl ushes followed by perspiration.2
selective serotonin-re-uptake inhibitor antidepressants
Up to 80% of patients undergoing treatment with such as venlafaxine, which although mainly investigated
GnRH analogues report hot fl ushes, and up to 27% in women13 has also been investigated in men.14–16report hot fl ushes as being the most troublesome side-
To our knowledge, head-to-head comparisons between
eff ect of treatment.3 Among patients who suff er from hot
treatments have not been made in men treated with
Figure 1: Trial profi le *10 disease progression; 11 adverse events; 8 deaths; 24 cases of non-compliance; 26 refusals to continue; 4 other.
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles
androgen-suppression therapy for prostate cancer. We
Venlafaxine Cyproterone Medroxyprogesterone
therefore designed a multicentre, randomised, double-
group(N=102) group (N=100) group (N=107)
cyproterone acetate, and medroxyprogesterone acetate,
for the treatment of hot fl ushes in patients with prostate
cancer who were being treated with GnRH analogues. Patients
Patients were enrolled from 106 urology centres in France
between April 14, 2004, and April 20, 2007. Eligible patients
Total number of hot fl ushes in the week
had to have histologically confi rmed prostatic adeno-
preceding randomisation visit (mean [SD])
carcinoma with localised or locally advanced disease, a
Karnofsky index score of 70% or more, and could not have
received any hormonal therapy. The inclusion criteria
were extended to patients with metastatic disease
following an amendment soon after the beginning of the
study. Patients treated with drugs related to the study
Table 1: Characteristics at randomisation according to treatment group
medications (such as selective serotonin-re-uptake inhibitors and steroid hormones) or that were potentially eff ective for the treatment of hot fl ushes (clonidine, each visit, patients had to complete daily self-evaluation gabapentine, veralipride, or β-alanine) were not included questionnaires—the Mayo Clinic hot-fl ush diary,17 in the study. Other exclusion criteria were contraindications
which is a validated symptom-measuring tool used to
to any of the study drugs (mainly corticosteroids and assess the frequency and severity grade of hot fl ushes. monoamine oxidase inhibitors), diabetes, severe kidney, Based on the defi nitions we provided them, patients liver, and cardiovascular disease, and psychiatric had to categorise the severity of their hot fl ushes as progressive disease. A history of thromboembolism was mild (less than 3 min, light physical symptoms, no also a criterion for exclusion. The trial was approved by emotional symptoms and no action needed), moderate the ethics committee of the coordinating centre, and (up to 5 min, moderate physical symptoms, mild regulatory bodies in all participating centres. All patients anxiety, some irritability and loss of concentration, need provided written informed consent before inclusion.
to use a fan, loosen clothing, and remove bedding), severe (up to 10 min, physical symptoms described as
Procedures
feeling hotter or very hot, heavy perspiration, dizziness,
This study was designed as a prospective, randomised, nausea, shortness of breath, weakness, and extreme double-blind trial, with the main objective of comparing discomfort; moderate anxiety and irritability; need to the effi
cacy of venlafaxine, medroxyprogesterone acetate, loosen clothing, change clothing, and change bedding), or
and cyproterone acetate for the prevention of vasomotor very severe (up to 30 min, very signifi cant physical and hot fl ushes in men with prostate cancer treated with GnRH
emotional symptoms and the need to change clothing,
analogues. All included patients received a quarterly bedding, towel off , take a shower, and take rest). Patients 11·25 mg depot of leuprorelin, administered subcutan-
were asked to make a daily record of the number of hot
eously. Flutamide 750 mg daily was given orally during the
fl ushes and their severity (mild=1; moderate=2; severe=3;
fi rst month of treatment to prevent testosterone fl are-up. and very severe=4). A daily score was calculated by After 6 months, patients who experienced 14 or more hot multiplying the number of hot fl ushes by the average fl ushes in the week before the visit, or those who severity. The “daily hot fl ush score” represented the mean experienced discomfort from hot fl ushes suffi
cient for value of these scores during the week preceding the visit.
them to spontaneously request treatment, were randomly
Additionally, participants completed the self-administered
assigned to receive either venlafaxine delayed release European Organisation for Research and Treatment of 75 mg/day, medroxyprogesterone 20 mg/day, or cypro-
terone acetate 100 mg/day, in addition to their third (version 3) on each visit, and informed the investigator leuprorelin injection. All patients received two indistin-
about their compliance with the treatment and their
guish able pills in the morning and one in the evening from
week 1 to week 8, and then one pill in the morning and (poor, average, good). The quality-of-life questionnaire one in the evening from week 9 to week 10, to take account
QLQ-C30 includes 30 questions covering fi ve functional
of the gradual discontinuation of venlafaxine, which was scales (physical, role, emotional, cognitive, and social); given at half-dose (37·5 mg/day) in weeks 9 and 10.
three symptom scales (fatigue, pain, and nausea and
Patients had follow-up visits 4 weeks, 8 weeks, and vomiting); and a global health and quality-of-life scale. Six
12 weeks after randomisation. During the week before single-item symptom measures cover dyspnoea, sleep
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles
in median hot fl ush score after 4 weeks of treatment could
be detected with an α risk of 5% and a β risk of 20%. We
estimated that 14% of the study population would not yield
evaluable data, and therefore decided to randomly assign 117 patients to each group (351 in total).
All patients who received at least one study treatment
data were compared using Fisher’s exact test or, when
not applicable, χ² test. Quantitative data were compared
by a Kruskall–Wallis test. Analysis of covariance was used
when comparing the change in hot-fl ush score between
randomisation and another time period. In case of a signifi cant diff erence between the three treatment
groups, all pairwise diff erences were tested using the
student’s test or, when applicable, the Wilcoxon test with
level of signifi cance adjustment using the Bonferroni
method (αĽ=α/3). This statistical procedure included the QLQ-C30 scores. Because of the numerous comparisons
Figure 2: Median daily hot fl ush score according to randomisation group
involved in the QLQ-C30 analysis and the subsequent
α risk, only signifi cant diff erences noted for at at least two timepoint comparisons were considered. Tests were
disturbance, appetite loss, constipation, diarrhoea, and bilateral, and considered as signifi cant at the level of fi nancial impact. All scales are scored from 0 to 100. High α=0·05. Statistical analyses were done with SAS 8.2 and scores for a functional scale represent high or healthy AdClin 3.1.1. This trial is registered with AFSSAPS levels of functioning, and similarly for global health or French National Registry, number F-LEU-100/ 03.11.21, quality of life. High scores for a symptom scale or item and ClinicalTrials.gov number NCT01011751. represent a high severity of symptom or problem.18
The primary endpoint was the change in median daily Role of the funding source
hot fl ush score after 4 weeks of treatment (between The study sponsor was involved in study design and data randomisation and 1 month) according to the hot-fl ush collection. An advisory board had overall scientifi c diary. Secondary endpoints included frequency of hot responsibility for study and protocol design, and advised fl ushes at each visit and relative score changes, percentage
the sponsor as to the conduct of the trial. After the end of
of patients with a decrease in hot fl ushes (complete or the trial, the board received all data analyses they more than 50%), eff ect of the treatment on quality of life,
requested. The sponsor had no role in the data analysis,
interpretation, or writing of the report. JI had full access to the raw data and had fi nal responsibility for the
Randomisation and masking
decision to submit the manuscript for publication.
An independent contract research organisation (CREAPHARM, Saint-Médard-en-Jalles, France) gener-
ated the randomisation list using a randomisation table, 919 men were enrolled in the study, with 311 (33·8%) and assigned the patients to the trial groups. Random reporting hot fl ushes requiring treatment 6 months after assignment with a block size of three was done centrally,
androgen suppression. The ratio of randomised/included
which was communicated to investigators by fax, and patients was lower than planned at the beginning of the each patient was given a randomisation number. The study, and despite amendments to increase the number allocation sequence was stratifi ed by centre. Participants,
of investigators, the number of inclusions, and the
investigators, sponsor personnel, and outcome assessors
duration of the study, we could not reach the primary goal
were blinded to treatment assignment, since the pills of 351 patients taking into account a 14% dropout rate. used for each of the three study treatments were
From Oct 19, 2004, to Sept 14, 2007, 311 patients were
randomly assigned to either the venlafaxine group (n=102), the cyproterone group (n=101), or the medroxyprogesterone
Statistical analysis
group (n=108; fi gure 1). Two patients were excluded for
cacy data from previous reports were only available for
protocol deviations (one in the cyproterone group and one
medroxyprogesterone acetate and venlafaxine. We used in the medroxyprogesterone group, both because they successive iterations based on the values selected for the were already taking serotonin re-uptake inhibitor averages, standard deviations, and α and β risks, to antidepressants at randomisation); thus 309 patients calculate that 92 patients per treatment group—276 in received at least one study treatment dose and were total—were needed to ensure a diff erence between groups
cacy analysis. The fi nal study visit was
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles Venlafaxine group Cyproterone group Medroxyprogesterone group
Median relative change from randomisation to 1 month (IQR)
Median relative change from 4 weeks to 8 weeks (IQR)
Median relative change from randomisation to 8 weeks (IQR)
Median relative change from randomisation to last available score (IQR)
Relative changes are signifi cant for the three drugs at all time intervals (p<0·0001) except for venlafaxine between 1 month and 2 months (p=0·4; signed rank Wilcoxon test). Table 2: Relative median daily hot-fl ush score changes according to the hot-fl ush diary Cyproterone group vs Cyproterone group vs Medroxyprogesterone group vs medroxyprogesterone group venlafaxine group venlafaxine group Randomisation to 4 weeks
–44·3 (–61·4 to –27·2, <0·0001)
–44·2 (–59·3 to –29·1,<0·0001)
4 weeks–8 weeks
14·8 (–56·5 to 86·1,0·6676)
–83·2 (–204·2 to 37·8,0·0122)
Randomisation: last available score
–52·4 (–70·3 to –34·5, <0·0001)
–52·2 (–69·8 to –34·6,<0·0001)
p value adjusted by Bonferroni method of two-by-two comparisons of treatment groups (covariance analysis). Table 3: Daily mean hot-fl ush score relative changes according to the hot fl ush diary
done on Dec 6, 2007. 264 patients (88 in the venlafaxine The changes in median hot-fl ush score from randomisation group, 81 in the cyproterone group, and 95 in the to 4 weeks; 4 weeks to 8 weeks; randomisation to 8 weeks, medroxyprogesterone group) attended all three scheduled and randomisation to the last available score are shown in follow-up visits and had no missing data (fi gure 1).
table 2. Pairwise comparison of treatment groups adjusted
At randomisation, 62 (20·1%) patients described their by the Bonferroni method confi rmed that the decreases in
hot fl ushes as mild, 134 (43·4%) described them as hot-fl ush scores were signifi cantly larger in the cyproterone moderate, and 113 (36·6%) described them as severe. and medroxyprogesterone groups than in the venlafaxine The median (mean) weekly number of hot fl ushes and group, regardless of the time interval looked at. There was daily hot-fl
ush score in the week preceding the no signifi cant diff erence between the cyproterone and
randomisation visit was 33·0 (44·4) and 7·6 (10·5), medroxyprogesterone groups (table 3). Overall, after respectively. The three treatment groups were comparable 4 weeks of treatment, 219 patients (70·9%) had an in size, and well matched in terms of age, body-mass improvement of at least 50% in their hot-fl ush scores, and index, blood pressure, Karnofsky index score, and the 70 patients (22·7%) reported a complete absence of hot proportion of mild, moderate, or severe hot fl ushes fl ushes (table 4). Improvements for both criteria were (table 1). However, there were a higher number of hot signifi cantly lower in the venlafaxine group than in the fl ushes and a higher median hot-fl ush score in the other two groups (p=0·0006; table 4). medroxyprogesterone acetate group in the week leading
QLQ-C30 questionnaires were not evaluable for fi ve, 20,
up to the randomisation visit (table 1).
40, and 53 patients at randomisation, 4 weeks, 8 weeks,
At 4 weeks, the median weekly number of hot fl ushes for and 12 weeks, respectively. Functional domains were
the study group was 9·0 (mean 15·7), with 21·0 (26·6), 4·0
scored high at all visits (median 83·3 [mean 85] out of
(1·8), and 7·5 (12·7) for the venlafaxine, cyproterone, and 100). The following functional and symptom scales were medroxyprogesterone groups respectively. The median not signifi cantly diff erent between any of the follow-up daily hot-fl ush score was 1·6 (mean 3·6) for the whole visits and randomisation: physical; role; cognitive and tudy group, 4·0 (6·5)for the venlafaxine group, 0·9 (1·8) social functioning; fatigue; pain; nausea and vomiting; for the cyproterone group, and 1·3 (2·7) for the global health status; insomnia; appetite loss; constipation; medroxyprogesterone group (fi gure 2). The decrease in diarrhoea; and fi nancial diffi
median daily hot-fl ush score between randomisation and emotional functional scale was signifi cantly better in the 4 weeks was signifi cant for all three groups (p<0·0001). venlafaxine group compared with the two other treatment
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles
thrombosis, hypertension), investigations abnormalities
Venlafaxine Cyproterone group Medroxyprogesterone Total
(abnormal blood glucose, increased blood pressure,
group (N=102) group (N=107)
increased weight), and nervous system disorders (balance
confusional state, sleep disorder). Other adverse events
included infections, neoplasms, psychiatric disorders, respiratory disorders, and breast tenderness (table 5).
Table 4: Improvement in daily hot-fl ush score after 4 weeks
There were 75 adverse events in 100 patients in the cyproterone group, 66 in 102 patients in the venlafaxine group, and 60 in 107 patients in the medroxyprogesterone
group. Cyproterone led to a higher number of vascular
Medroxyprogesterone groupVenlafaxine group
adverse events than did the other treatments, but this
was not statistically signifi cant (table 5). The number of patients who presented with adverse events related to study drug or which led to treatment discontinuation
between randomisation and 12 weeks was comparable (table 5). 16 patients presented with serious adverse
events, two of which were related to the study drugs: one case of dyspnoea caused by cyproterone acetate and one case of urticaria caused by medroxyprogesterone acetate. Discussion
Our study shows that men with signifi cant hot fl ushes during androgen suppression responded better to
cyproterone acetate and medroxyprogesterone acetate
than to venlafaxine within the 12-week study period.
Androgen suppression using GnRH analogues is
Figure 3: Mean emotional functioning scores Among the EORTC QLQ-C30 domains, the emotional functional scale score (0–100) changes were signifi cantly
widely prescribed in prostate-cancer treatment.1 Hot
diff erent between the three treatment groups between randomisation and 4 weeks (p=0·0147) and between
fl ushes are the most common side-eff ect of this treatment,
randomisation and 8 weeks (p=0·0047). Venlafaxine achieved the highest scores at 4 weeks and 8 weeks.
and can be very uncomfortable for patients.3 However,
The blue, red and green dotted lines represent the linear tendency between randomisation and 12 weeks for
hot fl ushes do not always occur, and when they do their
venlafaxine, cyproterone, and medroxyprogesterone groups, respectively.
eff ect is not always considered by the patient as worthy of
groups between randomisation and 4 weeks and between treatment. In our study, after 6 months of androgen-randomisation and 8 weeks (fi gure 3). The dyspnoea suppression treatment with leuprorelin, 311 patients single-item symptom score was also signifi cantly diff erent were randomised either because of the frequency of their between randomisation and 4 weeks (p=0·0353) and hot fl
ushes or their level of discomfort, whereas
between randomisation and 8 weeks (p=0·0126), with the 386 patients did not meet the randomisation criteria best scores being achieved in the venlafaxine group.
(fi gure 1). The proportion of patients entering the study
cacy was rated as good in 171 was much lower than the 65% expected on the basis of
(57·0%) of 300 patients at 4 weeks, 171 (61·5%) of previous reports. It became apparent during the study 278 patients at 8 weeks, and 151 (57·0%) of 265 patients at
that a signifi cant number of patients with hot fl ushes do
12 weeks. Cyproterone acetate and medroxyprogesterone not seek treatment: only 204 patients among the acetate were considered to have good effi
cacy at all visits, 919 enrolled (22·2%) spontaneously asked for treatment.
with 60·0–84·1% patients rating treatment effi
cacy as In our view, this justifi es the recommendation not to
good. Signifi cantly fewer patients in the venlafaxine group
systematically prevent hot fl ushes in all patients on
cacy as good (28·4–33·7%; p<0·0001).
initiation of an androgen-suppressing treatment.
Compliance was good in all treatment groups (291 of
Diff erent treatments have been assessed for the
303 patients; >96%). Six deaths were reported during the
treatment of hot fl ushes, including complementary
study, three during the fi rst 6 months, and three between
or alternative treatments,6 acupuncture,8 clonidine,19
randomisation and 12 weeks. None of the deaths were gabapentine,13 selective serotonin re-uptake inhibitors, and considered to be related to study treatments. During the hormonal treatments. Many have been assessed hot-fl ush treatment period, 201 adverse events occurred individually against a placebo in randomised trials, mostly (table 5). The most frequent adverse events related to in postmenopausal women or women being treated for study drug were gastrointestinal disorders (abdominal breast cancer. Fewer trials have been done in men who are pain, constipation, diarrhoea, and nausea), general undergoing hormonal therapy for prostate cancer. We disorders (asthenia, face oedema, feeling cold, sense of designed this study to assess the drugs with the most oppression), vascular disorders (deep and superfi cial vein convincing previous evidence of good response rates and
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles Venlafaxine group Cyproterone Medroxyprogesterone group (N=100) group (N=107)
Total number of adverse events related to study drugs
Patients with one or more serious adverse event
Patients with one or more adverse event related to study drug
Patients with one or more serious adverse event related to study drug
Patients with one or more vascular disorders related to study drug (n)
Patients with one or more adverse event leading to discontinuation
All data are n (%) unless stated otherwise. *There is no signifi cant diff erence between the three treatment groups. Table 5: Adverse events in the 3 months after randomisation by treatment group*
tolerance,7,9 and our study confi rmed the fi ndings of and 4 weeks and between randomisation and 8 weeks previous studies, since all three drugs induced a signifi cant (fi gure 3). This could be seen as a positive side-eff ect of improvement in hot fl ushes compared with baseline venlafaxine caused by its antidepressant properties. (randomisation). Progestagens proved their eff ectiveness
An intra-patient improvement in hot-fl ush frequency
and satisfactory tolerance.20,21 We used medroxyprogesterone
and intensity is not enough to draw a defi nite conclusion
acetate rather than megestrol acetate, since no low-dose about drug effi
cacy, owing to possible confounding by the
formulation of megestrol acetate is available in France; the
placebo eff ect.9,18 Most of the phase 3 studies of the three
only marketed form consists of strongly dosed tablets study drugs to date have been placebo-controlled, but (160 mg). The dose of 20 mg/day of medroxyprogesterone there are very few data on direct comparisons between acetate seemed to have a good benefi t/risk profi le, since drugs with an established effi
the response rates seem independent of the dose according
cacy and tolerance of the three drugs head to head,
cacy of cyproterone acetate with the aim of producing directly comparable data to
in treating hot fl ushes was shown many years ago.10,24
help guide therapeutic choices, since there are no clear
However, this steroidal anti-androgen has mainly been recommendations in this fi eld. We also chose not to used as a treatment for prostate cancer. We decided to use include a placebo group, since we considered that enough the dose of 100 mg/day, owing to the lack of previous data data were available for the study drugs versus placebo. assessing doses lower than 100 mg/day for hot-fl ush
We found that the hormonal treatments (cyproterone
treatment. The use of antidepressants such as venlafaxine acetate and medroxyprogesterone acetate) were better at to treat hot fl ushes is a recent development. The venlafaxine
reducing hot fl ushes and are more eff ective than
dose we used in our study was 75 mg/day, after previous venlafaxine, with little diff erence in tolerance between the studies showed that higher doses gave the same response treatments. Our results confi rm those of the North with a less satisfactory tolerance.25 It is possible that starting
Central Cancer Treatment Group Trial N99C7, in which
at 75 mg/day rather than at a lower dose could have led to depomedroxy progesterone acetate was shown to be better more cases of early drug discontinuation; however, overall at controlling hot fl ushes than venlafaxine in menopausal compliance was good in all treatment groups (>96%). The women.26 We believe that the randomised, double-blind number of patients who presented with adverse events design of our study enables us to conclude with reasonable between randomisation and 12 weeks that led to confi
dence that cyproterone acetate and medroxy-
discontinuation were comparable in the three treatment progesterone acetate are more eff ective than venlafaxine groups (table 5). The functional and symptom scales at 12 weeks for treating hot fl ushes in men who are treated assessed by the QLQ C-30 were not signifi cantly diff erent with GnRH analogues for prostate cancer. There was no between visits and compared with randomisation except signifi cant diff erence between cyproterone acetate and for the emotional functional scale, which was signifi cantly medroxyprogesterone acetate, although there was an improved in the venlafaxine group between randomisation
imbalance in the number of patients spontaneously
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9 Articles
asking for hot-fl ush treatment before randomisation and 7
Loprinzi CL, Barton DL, Sloan JA, et al. Mayo Clinic and North
in the total number of hot fl ushes in the week preceding
Central Cancer Treatment Group hot fl ash studies: a 20-year experience. Menopause 2008; 15: 655–60.
randomisation that worked against medroxy progesterone
Frisk J, Spetz AC, Hjertberg H, Petersson B, Hammar M.
acetate. Additionally, cyproterone acetate (at higher doses)
Two modes of acupuncture as a treatment for hot fl ushes in men
is a recognised hormonal treatment in prostate cancer,
with prostate cancer-a prospective multicenter study with long-term follow-pp. Eur Urol 2009; 55: 156–63.
and its use for the treatment of hot fl ushes could 9 Thompson CA, Shanafelt TD, Loprinzi CL. Andropause: symptom
potentially interfere with a planned hormonal therapy.
management for prostate cancer patients treated with hormonal
We therefore consider that medroxyprogesterone acetate
ablation. Oncologist 2003; 8: 474–87.
should be considered the standard treatment for hot 10 Eaton AC, McGuire N. Cyproterone acetate in treatment of
post-orchidectomy hot fl ushes. Double-blind cross-over trial.
ushes in men treated with androgen-suppression
Lancet 1983; 2: 1336–37.
11 Sakai H, Igawa T, Tsurusaki T, et al. Hot fl ushes during androgen
However, there are several caveats that should be borne
deprivation therapy with luteinizing hormone-releasing hormone agonist combined with steroidal or nonsteroidal antiandrogen for
in mind when using our data to inform treatment
prostate cancer. Urology 2009; 73: 635–40.
recommendations. First, we only have data for a short 12 Higano CS. Side eff ects of androgen deprivation therapy: follow-up period: there are no data on the stability of our
monitoring and minimizing toxicity. Urology 2003; 61 (2 suppl): 32–38.
results in the long term. Second, the potential eff ect of 13 Loprinzi CL, Sloan J, Stearns V, et al. Newer antidepressants and
steroidal anti-androgens on prostate cancer should be
gabapentin for hot fl ushes: an individual patient pooled analysis.
taken into account. Cyproterone acetate has been used
J Clin Oncol 2009; 27: 2831–37.
for decades to treat prostate cancer, either alone or in 14 Quella SK, Loprinzi CL, Sloan J, et al. Pilot evaluation of
venlafaxine for the treatment of hot fl ushes in men undergoing
combination with androgen suppression. Patients treated
androgen ablation therapy for prostate cancer. J Urol 1999;
with a combination of GnRH analogues and cyproterone
162: 98–102.
acetate or medroxyprogesterone acetate should be aware 15 Loprinzi CL, Barton DL, Carpenter LA, et al. Pilot evaluation of
paroxetine for treating hot fl ushes in men. Mayo Clin Proc 2004;
of the potential for withdrawal syndrome, and consider
79: 1247–51.
stopping cyproterone acetate or medroxyprogesterone 16 Naoe M, Ogawa Y, Shichijo T, Fuji K, Fukagai T, Yoshida H. Pilot acetate as a fi rst step when prostate-specifi c antigen
evaluation of selective serotonin reuptake inhibitor antidepressants in hot fl ash patients under androgen-deprivation therapy for
prostate cancer. Prostate Cancer Prostatic Dis 2006; 9: 275–78. Contributors
17 Aaronson NK, Ahmedzai S, Bergman B, et al. The European
JI, PB, and NM designed the study. JI and RO collected the data. JI, LS,
Organization for Research and Treatment of Cancer QLQ-C30:
PB, and NM were involved in the analysis and interpretation of the data.
a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 1993; 85: 365–76.
All authors critically revised the manuscript and gave their fi nal approval for its publication.
18 Sloan JA, Loprinzi CL, Novotny PJ, Barton DL, Lavasseur BI,
Windschitl H. Methodologic lessons learned from hot fl ash studies.
Confl icts of interest J Clin Oncol 2001; 19: 4280–90.
JI has received honoraria for the study advisory board and for postgraduate
19 Buijs C, Mom CH, Willemse PH, et al. Venlafaxine versus
education lectures from Takeda, Ipsen, and Astellas. PB has received
clonidine for the treatment of hot fl ushes in breast cancer patients:
research grants and honoraria for consultancies from Schering Plough,
a double-blind, randomized cross-over study. Breast Cancer Res Treat
PregLem, Wyeth, HRA Pharma, and Pierre Fabre Research. RO is an
2009; 115: 573–80.
employee of Takeda. NM has received research grants from Takeda and
20 Loprinzi CL, Michalak JC, Quella SK, et al. Megestrol acetate for the
Caprion, and has served on advisory boards for Takeda, Caprion, and
prevention of hot fl ushes. N Engl J Med 1994; 331: 347–52.
Thallion. He received honoraria for postgraduate education lectures from
21 Quella SK, Loprinzi CL, Sloan JA, et al. Long term use of megestrol
Takeda, Ipsen, and Astellas. LS declared no confl icts of interest.
acetate by cancer survivors for the treatment of hot fl ushes. Cancer 1998; 82: 1784–88. Acknowledgments
22 Charig CR, Rundle JS. Long-term side eff ect of orchiectomy in
We thank Stéphanie Chretin and Françoise Bugnard for assistance in the
treatment of prostatic carcinoma. Urology 1989; 33: 175–78.
statistical analysis and interpretation of the data.
23 Schiff I, Tulchinsky D, Cramer D, Ryan KJ. Oral
References
medroxyprogesterone in the treatment of postmenopausal
Sennfält K, Carlsson P, Varenhorst E. Diff usion and economic
symptoms. JAMA 1980; 244: 1443–45.
consequences of health technologies in prostate cancer care in
24 Cervenakov I, Kopecny M, Jancar M, et al. Hot fl ush, an unpleasant
Sweden, 1991–2002. Eur Urol 2006; 49: 1028–34.
symptom accompanying antiandrogen therapy of prostatic cancer
Kouriefs C, Georgiou M, Ravi R. Hot fl ushes and prostate cancer:
and its treatment by cyproterone acetate. Int Urol Nephrol 2000;
pathogenesis and treatment. BJU Int 2002; 89: 379–83. 32: 77–79.
3 Sharifi N, Gulley JL, Dahut WL. Androgen deprivation therapy for
25 Loprinzi CL, Kugler JW, Sloan JA, et al. Venlafaxine in management
prostate cancer. JAMA 2005; 294: 238–44.
of hot fl ushes in survivors of breast cancer: a randomised controlled trial. Lancet 2000; 356: 2059–63.
Spetz AC, Hammar M, Lindberg B, Spångberg A, Varenhorst E. Prospective evaluation of hot fl ushes during treatment with
26 Loprinzi CL, Levitt R, Barton D, et al. Phase III comparison of
parenteral estrogen or complete androgen ablation for metastatic
depomedroxyprogesterone acetate to venlafaxine for managing
carcinoma of the prostate. J Urol 2001; 166: 517–20.
hot fl ushes: North Central Cancer Treatment Group Trial N99C7. J Clin Oncol 2006; 24: 1409–14.
Nishiyama T, Kanazawa S, Watanabe R, Terunuma M, Takahashi K. Infl uence of hot fl ushes on quality of life in patients with prostate
27 Burch PA, Loprinzi CL. Prostate-specifi c antigen decline after
cancer treated with androgen deprivation therapy. Int J Urol 2004;
withdrawal of low-dose megestrol acetate. J Clin Oncol 1999;
11: 735–41. 17: 1087–88.
Moyad MA. Complementary/alternative therapies for reducing hot
28 Sella A, Flex D, Sulkes A, Baniel J. Antiandrogen withdrawal
fl ushes in prostate cancer patients: reevaluating the existing indirect
syndrome with cyproterone acetate. Urology 1998; 52: 1091–93.
data from studies of breast cancer and postmenopausal women. Urology 2002; 59 (suppl 1): 20–33.
www.thelancet.com/oncology Published online December 7, 2009 DOI:10.1016/S1470-2045(09)70338-9
KLOR-CON ® particularly careful monitoring of the serum potassium concentrationpotassium is administered too rapidly intravenously, potentially fatalhyperkalemia can result (see Contraindications and Warnings ). It is Interaction with Potassium-Sparing Diuretics: Hypokalemia should important to recognize that hyperkalemia is usually asymptomatic andPowder (Potassium Chloride for Oral
Articles