Doi:10.1016/s0140-6736(00)03403-6

Early report
Venlafaxine in management of hot flashes in survivors of breastcancer: a randomised controlled trial Charles L Loprinzi, John W Kugler, Jeff A Sloan, James A Mailliard, Beth I LaVasseur, Debra L Barton, Paul J Novotny, Shaker R Dakhil, Kate Rodger, Teresa A Rummans, Bradley J Christensen IntroductionHot flashes are a substantial problem in menopausal Background Hot flashes can be troublesome, especially when women. Oestrogen therapy is the mainstay of treatment for hormonal therapy is contraindicated. Preliminary data have this symptom and generally controls hot flashes well.
suggested that newer antidepressants, such as venlafaxine, Hot flashes may be more troublesome in women who can diminish hot flashes. We undertook a double-blind, have survived breast cancer1,2 than in other women for placebo-controlled, randomised trial to assess the efficacy of several reasons. First, many women treated for breast venlafaxine in women with a history of breast cancer or reluctance to take hormonal treatment because of fear of menopause from chemotherapy. Second, many survivors of breast cancer are given tamoxifen, the most prevalent side-effect of which is hot flashes. Third, women with a history Methods Participants were assigned placebo (n=56) or of breast cancer have generally been denied oestrogen venlafaxine 37·5 mg daily (n=56), 75 mg daily (n=55), or therapy, at least in North America, because of concerns 150 mg daily (n=54). After a baseline assessment week, about potentiating recurrence of breast cancer.
patients took the study medication for 4 weeks. All venlafaxine Many agents have been investigated as potential means treatment started at 37·5 mg daily and gradually increased in for alleviating hot flashes in survivors of breast cancer. The the 75 mg and 150 mg groups. Patients completed daily hot- best-described non-oestrogenic treatments for hot flashes flash questionnaire diaries. The primary endpoint was average are progestagens. For example, low doses of megestrol daily hot-flash activity (number of flashes and a score acetate result in a reduction of about 80% in hot flashes, combining number and severity). Analyses were based on the compared with a decrease of about 20% with placebo.3 At women who provided data throughout the baseline and study present, there are no convincing data that megestrol acetate has any substantial positive or negative effect on breast-cancer morbidity or mortality. Therefore, this therapy can Findings 191 patients had evaluable data for the whole study reasonably be used after a thorough discussion of potential period (50 placebo, 49 venlafaxine 37·5 mg, 43 venlafaxine risks and benefits with the patient. Nonetheless, some 75 mg, 49 venlafaxine 150 mg). After week 4 of treatment, patients and physicians are concerned about the use of any median hot flash scores were reduced from baseline by 27% hormone in survivors of breast cancer. Thus, non- (95% CI 11–34), 37% (26–54), 61% (50–68), and 61% hormonal means to alleviate hot flashes in these patients are (48–75) in the four groups. Frequencies of some side-effects needed. Various non-hormonal therapies, including vitamin E, clonidine, Bellergal (phenobarbital, ergotamine, constipation) were significantly higher in the venlafaxine 75 mg and levorotatory alkaloids of belladonna), and methyldopa, and 150 mg groups than in the placebo group.
have been examined4–7 but have limited efficacy or adverseside-effects.
On the basis of positive anecdotal experience, we treatment for hot flashes, though the efficacy must be undertook a pilot study of the antidepressant venlafaxine in balanced against the drug’s side-effects. Confirmation of the cancer patients with hot flashes. The results supported the results of this 4-week study awaits the completion of three ongoing randomised studies to assess the effects of other We undertook this study to assess more definitively the related antidepressants for the treatment of hot flashes.
efficacy and toxicity of various doses of venlafaxine for thetreatment of hot flashes in survivors of breast cancer. The hypothesis was that venlafaxine would be effective in alleviating hot-flash activity. Planned subgroup analysesincluded an investigation into whether there was a dose-response relation, with control for potential confounding Mayo Clinic and Mayo Foundation, Rochester, MN 55905(Prof C L Loprinzi T A Rummans MD, B J Christensen BS Pharm); Illinois OncologyResearch Association Community Clinical Oncology Program (CCOP), Peoria, IL (J W Kugler MD); Missouri Valley Cancer Consortium, Omaha, NE (Prof J A Mailliard MD); Ann Arbor Regional Patients eligible for this trial were women who had a history CCOP, Ann Arbor, MI (B LaVasseur RN); Carle Cancer Center CCOP, of breast cancer or who were concerned about taking Urbana, IL (D L Barton RN); Wichita Community Clinical Oncology, oestrogen for fear of breast cancer. Inclusion criteria were: Wichita, KS (S R Dakhil MD); and Ochsner CCOP, New Orleans, LA troublesome hot flashes, occurring at least 14 times per week; flashes severe enough for the patient to desire Correspondence to: Dr Charles L Loprinzi, Division of Medical therapeutic intervention, and present for at least a month Oncology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, before study entry; age older than 18 years; life expectancy at least 6 months; and performance status of 0–1 on the Eastern Cooperative Oncology Group scale.
THE LANCET • Vol 356 • December 16, 2000 For personal use only. Not to be reproduced without permission of The Lancet.
Central Cancer Treatment Group randomisation office and study statisticians had access to individual treatmentassignments during the course of the study.
After randomisation, but before starting the assigned medication, each patient was asked to compete a daily diary hot-flash questionnaire for one baseline week. This questionnaire was then completed daily for the 4 weeks the patient received study medication or placebo. This questionnaire was similar to those shown to be reliable and valid in previous clinical trials involving more than 950 Patients were also asked to complete two single-item global quality-of-life questions12 and a Beck depressioninventory13 at the end of each of the 5 study weeks. Bloodpressure was measured weekly during the study period. Astudy nurse contacted each patient once a week to encourage appropriate protocol participation, to inquire about hot flashes, and to assess side-effects.
If the diastolic blood pressure rose above 95 mm Hg (confirmed on more than one reading), the patient waswithdrawn. The study code was broken, and the patient was allowed to taper the drug dose. 6-monthly toxicity (unmasked) and efficacy (masked) statistical summary reports were reviewed by the North Central CancerTreatment Group external data-monitoring committee.
The weekly questionnaires inquired about potential side- effects of venlafaxine: loss of appetite, sleepiness, nausea, dizziness, tiredness (fatigue), mouth dryness, abnormal sweating, constipation, insomnia, nervousness, and moodchanges.
The daily hot-flash questionnaire asked about the numbers of mild, moderate, severe, and very severe hot aromatase inhibitors were allowed if they had been started 4 flashes per day (24 h period). Descriptions of hot-flash weeks before the beginning of the study and were scheduled definitions from women who had taken part in previous to continue for the next 5 weeks. Concomitant therapies studies were provided to each patient in the questionnaire not allowed were: antineoplastic chemotherapy, androgens, oestrogens, progestagens, antidepressants, clonidine, andBellergal. Exclusion criteria were use of venlafaxine in the past; any antidepressant treatment within the preceding Methods used to analyse the data were similar to those used 2 years; pregnancy; breastfeeding; use of other medications for our previous hot-flash studies.3–5,11 to treat hot flashes within the previous 2 weeks; endpoint was a bivariate construct of average daily hot-flash activity: the number of hot flashes and a score combining blood pressure >95 mm Hg, systolic blood pressure the number and severity of hot flashes. Comparisons of hot- >160 mm Hg, or both). Women with childbearing flash activity between treatment groups used two-sided potential had to use adequate contraceptive measures. All testing. Comparisons between treatment weeks that patients were required to give written informed consent as involved dose escalation used average within-patient differences. Secondary endpoints, compared by the same Individual patients were stratified according to age (<50 vs ୑50 years), current tamoxifen use, duration of hot-flash symptoms (<9 vs ୑9 months), and the average frequency of hot flashes per day (2–3 vs 4–9 vs ୑10). Stratificationwas achieved by a method of dynamic allocation thatbalances the marginal distributions.10 Patients were then randomly assigned one of four treatments: extended-releasevenlafaxine 37·5 mg daily for 28 days; extended-release venlafaxine 37·5 mg daily for 7 days, then 75 mg daily for21 days; extended-release venlafaxine 37·5 mg daily for 7days, 75 mg daily for 7 days, then 150 mg daily for 14 days; To ensure that treatment allocation was concealed from patients and medical professionals, and to ensure that patients received the right medication dose on each day, blister packs were issued, with three tablets each day for allpatients. The tablets consisted of 37·5 mg extended-release Figure 2: Mean decreases in hot-flash scores venlafaxine, 75 mg extended-release venlafaxine, or placebo For whole period: placebo vs 75 mg and 150 mg, p<0·0001; placebo vs37·5 mg, p=0·008; 75 mg vs 37·5 mg, p=0·03; 150 mg vs 37·5 mg, of identical appearance (tablets provided by Wyeth-Ayerst p=0·13. Note that doses increased gradually in the 75 mg and 150 mg Laboratories, Philadelphia, PA, USA). Only the North THE LANCET • Vol 356 • December 16, 2000 For personal use only. Not to be reproduced without permission of The Lancet.
Median decrease (%) from baseline to week 4 (95% CI) *p<0·001 for comparison with placebo.
Table 1: Median decrease in hot-flash frequencies and scores Table 2: Patients with hot-flash scores of various percentages methods, included changes in mood (Beck depression Figure 3: Toxicity comparisons during treatment week 4 inventory) and quality of life from baseline to end of treatment. The proportion of patients who withdrew Only effects with significant differences among study groups are shown.
*p<0·05 for comparison with placebo group. †Significantly different from prematurely was compared by ␹2 test. Correlation between other venlafaxine groups (p=0·0006 for 37·5 mg, p=0·02 for 75 mg).
covariates (mood change, quality of life) was assessed withSpearman correlation coefficients. Repeated-measures been taking the assigned maximum dose for at least 2 weeks), the median decrease in hot-flash scores was procedures were used to carry out a conditional analysis of significantly greater in all three venlafaxine groups than in the treatment effect (on the primary endpoint) in the the placebo group (p<0·0001). Only ten of 50 (20% [95% CI 10–34]) patients in the placebo group reported a Missing data were handled in several ways as a sensitivity reduction of more than 50% in hot-flash activity compared analysis of the robustness of results in relation to missing with 22 of 49 (45% [31–60]), 27 of 43 (63% [47–77]), and data. Less than 10% of possible data were missing, and the 27 of 49 (55% [40–69]) in the three venlafaxine groups results were consistent across a series of analyses by various (table 2). Changes within an individual in hot-flash activity in the groups that had dose escalation mirrored the results Daily hot-flash scores were calculated by assigning a across treatment groups, in that there were significant number (1 to 4) to each severity: mild, moderate, severe, decreases in activity with a change from the baseline week and very severe. These numbers were multiplied by the to 37·5 mg venlafaxine (p=0·01) or from 37·5 mg to 75 mg daily frequencies of each type of flash and the four products (p=0·01), but not from 75 mg to 150 mg (p=0·74; figure were added together to give the daily score.
2). Efficacy was similar in patients who were and those who We calculated that a sample size of 50 patients per group were not receiving tamoxifen (data not shown).
would provide 80% power to detect differences in average Toxicity data are best illustrated in terms of the changes hot-flash activity of 0·6 SDs (1·2 hot flashes per day, a score in treatment week 4 compared with the baseline week of 3 units, or a 21% fall from baseline) with a type 1 error because the groups assigned increasing doses had then been taking the highest dose for at least 2 weeks. There were nosignificant differences between groups for tiredness, dizziness, nervousness, mood changes, sweating, sleepiness, 229 patients joined this study between Feb 22 and July 27, or sleeping troubles. Significant differences were seen in 1999 (figure 1). Seven patients withdrew before taking any terms of mouth dryness, decreased appetite, nausea, and study medication and one patient was found to be ineligible. Hot-flash data for the baseline week were available for 207 patients (94%) and evaluable data over the whole study period for 191 (86%). The 30 patients who did not provide usable hot-flash data had stopped the study medications, did not properly complete or return the diaryforms, or both.
At study entry the four groups were well balanced in terms of age, tamoxifen use (69% of patients), the durationof hot-flash symptoms, the patients’ estimated frequencies of hot flashes, race, baseline uniscale quality-of-life values, and baseline Beck depression inventory scores. During the baseline week, hot-flash frequencies (average 8·0), scores(average 13·3), or both did not differ significantly among Changes in hot-flash scores for the four study groups are shown in figure 2 and table 1. After 4 weeks of study treatment (when the groups assigned titrated doses had THE LANCET • Vol 356 • December 16, 2000 For personal use only. Not to be reproduced without permission of The Lancet.
both serotonin reuptake (similarly to many of the other newer antidepressants) and norepinephrine reuptake Venlafaxine’s effects are thought to be related to theserotonin-reuptake inhibition at lower doses, but to a combination of serotoninergic and noradrenergic or apredominance of noradrenergic effects at higher doses.
Because hot flashes are considered to be a side-effect associated with older tricyclic antidepressants, this propertyof venlafaxine may explain the ceiling effect for hot flashes.
The main side-effects observed with the 75 mg daily dose of venlafaxine were mouth dryness, anorexia, and nausea.
In general, these side-effects were tolerable.
Evidence on the role of antidepressants in sexual dysfunction is mixed, and the relation is further complicated by the effect of depression. Placebo-controlledstudies reported in the Physician’s Desk Reference state that the frequency of sexual dysfunction with venlafaxine is about 2%.17 An overview of these studies reported placebo- adjusted frequencies of 2% for decreased libido and 2% for There were no significant differences between groups. unspecified female sexual dysfunction events withvenlafaxine in doses of 75 to 375 mg daily.18 In one study, constipation (figure 3). More toxic effects occurred with the with doses of up to 160 mg daily, there was no mention of 150 mg dose than with the 75 mg dose. Nausea was sexual dysfunction in the list of side-effects or the reasons temporary in most cases and largely resolved with time for drug discontinuation.19 Another study reported nausea as the only significant adverse event with a dose of up to Although we did not specifically inquire about changes in 182 mg daily.20 Our findings confirm that venlafaxine doses libido, item number 21 of the Beck depression inventory of 37·5–150 mg daily did not reduce libido, at least during addresses this issue. Libido scores improved from baseline the first month of therapy. We hypothesise that the for all study groups over the 4-week study period (figure 5).
improvement in hot flashes may have decreased night We did not observe any substantial changes in diastolic sweats and improved sleep, thus decreasing fatigue and or systolic blood pressure for any of the study groups (data Are the methods used in this study sound? Can patient- Depression scores improved by an average of 1·6, 2·4, completed questionnaires about subjective symptoms such 4·8, and 3·2 points for the placebo and 37·5 mg, 75 mg, as hot flashes provide scientifically reliable information? and 150 mg venlafaxine groups, respectively. Similarly, at Psychometric studies have shown that they can.21,22 Patients the end of the study, 16 of the 48 (33%) patients on can provide valid, reliable data about their subjective placebo had depression scores consistent with at least mild experience.23,24 Reports on a wide range of oncology depression compared with 11 of 40 (23%), nine of 43 endpoints related to quality of life have also shown the (21%), and 13 of 49 (27%) in the three venlafaxine groups scientific integrity of these methods.25,26 In figure 2 of our (p=0·59). Although the trend of the depression scores report, the data points for the three treatment groups paralleled the results of hot-flash activity, the relation was weak (Spearman’s ␳=0·29, p=0·0004). Repeated-measures superimposed. These points represent data from 54–56 ANOVA showed that baseline depression was unrelated to patients who were all receiving 37·5 mg daily of venlafaxine the reduction in hot-flash activity (p=0·54).
during that week. The similarity of the results shows the Overall quality of life increased from baseline to reproducible results of this experimental method. Further treatment week 4 by an average of 3 points in the three support for the method used comes from comparison of the venlafaxine groups and decreased 3 points in the placebo results of one of our previous studies of clonidine for hot group (p=0·02). There were no significant differences in flashes5 with those from another group of investigators.27 these changes among the venlafaxine groups.
These independent studies reported remarkably similareffects of clonidine. Cross-study comparison of placebos in our hot-flash studies shows a 20–30% reduction in hot This trial suggests that venlafaxine can alleviate hot flashes flashes over 4 weeks in groups receiving placebo. Thus, we and that the most appropriate dose for this indication is are convinced that the method is scientific, allowing an 75 mg daily, which was more effective than 37·5 mg daily objective measure of subjectively reported hot flashes.
but was as effective as, and less toxic than, the 150 mg dose.
The efficacy of venlafaxine against hot flashes does not Thus, we recommend that treatment should start with a seem to be specific to this single antidepressant. Preliminary daily dose of 37·5 mg and be increased if necessary to results suggest that several of the other newer 75 mg, but not higher. Some patients had substantial antidepressants can also alleviate hot flashes.28,29 Data from decreases in hot-flash activity with 37·5 mg venlafaxine randomised trials currently investigating these drugs should daily. For those patients, there may not be any need to better elucidate the efficacy and toxicity profiles of the With hormone treatments for hot flashes, effects may Although there was a weak correlation between the take weeks to become apparent. By contrast, the effect of decreases in hot flashes and improved scores on the Beck venlafaxine occurred in a matter of days, in most cases.
depression inventory, the effect of venlafaxine on hot Thus, an increase in dose after a week is reasonable if flashes seems to be through a mechanism other than its known effect on depression symptoms. Patients who had The dose-dependent efficacy of venlafaxine may be normal baseline depression scores had a similar reduction directly related to its pharmacology. Venlafaxine affects in hot flashes to those with high scores.
THE LANCET • Vol 356 • December 16, 2000 For personal use only. Not to be reproduced without permission of The Lancet.
From the available information, venlafaxine can be Loprinzi CL, Pisansky TM, Fonseca R, et al. Pilot evaluation of recommended for treatment of hot flashes for patients in venlafaxine hydrochloride for the therapy of hot flashes in cancer whom oestrogen is not desired. The drug also seems to be surivors. J Clin Oncol 1998; 16: 2377–81.
Quella SK, Loprinzi CL, Sloan J, et al. Pilot evaluation of venlafaxine effective against hot flashes in men who have undergone for the treatment of hot fashes in men undergoing androgen ablation androgen deprivation therapy for prostate cancer.9 therapy for prostate cancer. J Urol 1999; 162: 98–102.
10 Pocock SL, Simon R. Sequential treatment assignment with blancing for prognostic factors in the controlled clinical trial. Biometrics 1975; 31:
Charles Loprinzi was the principal investigator, responsible for the development and conduct of this study and the preparation of the report. 11 Quella SK, Loprinzi CL, Barton DL, et al. Evaluation of soy John Kugler helped with study design, chaired the study group, and phytoestrogens for the treatment of hot flashes in breast cancer contributed to the report. Jeff Sloan was responsible for statistical analysis.
survivors: an NCCTG trial. J Clin Oncol 2000; 18: 1068–74.
James Mailliard, Beth LaVasseur, Shaker Dakhil, and Kate Rodger were 12 Sloan JA, Loprinzi CL, Kuross SA, et al. Randomized comparison of involved in study development, accrual of patients, and review of the paper.
four tools measuring quality of life in patients with advanced cancer.
Paul Novotny was involved with the study development, data analysis, and J Clin Oncol 1998; 16: 3662–73.
drafting of the report. Teresa Rummans was involved with study 13 Beck AT, Steer RA, Garbin MG. Psychometric properties of the Beck development and drafting of the report. Bradley Christensen was involved Depression Inventory: 25 years of evaluation. Clin Psychol Rev 1988; 8:
with study development, distribution and supply of protocol drugs, and 14 Finck G, Barton DL, Loprinzi CL, Quella SK, Sloan JA. Definitions of hot flashes in breast cancer survivors. J Pain Symptom Manage 1998; 16:
327–33.
North Central Cancer Treatment GroupAdditional participating institutions include: Duluth CCOP, Duluth, MN 15 Sanchez C, Hyttel J. Comparison of the effects of antidepressants and (J E Krook); Iowa Oncology Research Association CCOP, Des Moines, IA their metabolites on reuptake of biogenic amines and on receptor (R F Morton); Rapid City Regional Oncology Group, Rapid City, SD binding. Cell Mol Biol 1999; 19: 467–89.
(L P Ebbert); Mayo Clinic Scottsdale CCOP, Scottsdale, AZ (T Fitch); 16 Westenberg HG. Pharmacology of antidepressants: selectivity or CentraCare Clinic, St Cloud, MN (H E Windschitl); Geisinger Clinical multiplicity? J Clin Psychol 1999; 60 (suppl 17): 4–8.
Oncology Program, Danville, PA (S Nair); Grand Forks Clinic, Grand 17 Physician’s desk reference. Montvale, NJ: Medical Economics Data, Forks, ND (J A Laurie); Siouxland Hematology-Oncology Associates, Sioux City, IA (J C Michalak); Metro-Minnesota Community Clinical Oncology 18 Preskorn S. Comparison of the tolerability of bupropion, fluoxetine, Program (P J Flynn); Quain and Ramstad Clinic, Bismarck, ND (F Addo); imipramine, nefazodone, paroxetine, sertraline and venlafaxine. J Clin Cedar Rapids Oncology Project CCOP, Cedar Rapids, IA (M Wiesenfeld); Psychiatry 1995; 6: 12–21.
Meritcare Hospital CCOP, Fargo, ND (R Levitt); Saskatchewan Cancer 19 Cunningham L, Borison R, Carman J, et al. A comparison of Foundation, Saskatoon, Saskatchewan, Canada (M T Tirona); Sioux venlafaxine, trazodone, and placebo in major depression. J Clin Community Cancer Consortium, Sioux Falls, SD (L K Tschetter).
Psychopharmacol 1994; 14: 99–106.
20 Schweizer E, Feighner J, Mandos L, et al. Comparison of venlafaxine and imipramine in the acute treatment of major depression in This study was conducted as a collaborative trial of the North Central outpatients. J Clin Psychiatry 1994; 55: 104–08.
Cancer Treatment Group and Mayo Clinic and was supported by Public 21 Wilson IB, Cleary PD. Linking clinical variables with health-related Health Service grants CA-25224, CA-37404, CA-35113, CA-63849, quality of life. JAMA 1995; 273: 59–64.
CA-63848, CA-35195, CA-35272, CA-35269, CA-35103, CA-35101, 22 Bayley KB, London MR, Grunkemeier GL, Lansky DJ. Measuring the CA-60276, CA-52352, CA-37417, and CA-35448.
success of treatment of patient terms. Med Care 1995; 33:
AS226–35.
23 Patrick D, Deyo R. Generic and disease-specific measures in assessing health status and quality of life. Med Care 1998; 27 (suppl):
Couzi RJ, Helzlsouer KJ, Fetting JH. Prevalence of menopausal symptoms among women with a history of breast cancer and attitudes 24 Bruera E, Kuehn N, Miller MJ, Selmser P, Macmillan K. The toward estrogen replacement therapy. J Clin Oncol 1995; 13: 2737–44.
Edmonton symptom assessment system (ESAS): a simple method for Vassilopoulou-Sellin R, Zolnski C. Estrogen replacement therapy in the assessment of palliative care patients. J Palliat Care 1991; women with breast cancer: a survey of patient attitudes. Am J Med Sci 7: 6–9.
1992; 304: 145–49.
25 Clark A, Fallowfield L. Quality of life measurements in patients with Loprinzi CL, Michalak JC, Quella SK, et al. Megestrol acetate for the malignant disease: a review. J R Soc Med 1986; 79: 165–69.
prevention of hot flashes. N Engl J Med 1994; 331: 347–52.
26 Aaronson NK. Methodological issues in assessing the quality of life of Barton DL, Loprinzi CL, Quella SK, et al. Prospective evaluation of cancer patients. Cancer 1991; 67 (suppl): 844–50.
vitamin E for hot flashes in breast cancer survivors. J Clin Oncol 1998; 27 Pandya KJ, Raaberters RF, Flynn PJ, et al. Oral clonidine in 16: 495–500.
postmenopausal patients with breast cancer experiencing tamoxifen- Goldberg RM, Loprinzi CL, O’Fallon JR, et al. Transdermal clonidine induced hot flashes: a University of Rochester Cancer Center for ameliorating tamoxifen-induced hot flashes. J Clin Oncol 1994; 12:
community clinical oncology program study. Ann Intern Med 2000; 132:
Bergmans MGM, Merkus JM, Corbey RS, et al. Effect of bellergal 28 Stearns V, Isaacs C, Crawford J, et al. A pilot trial assessing the efficacy retard on climacteric complaints: a double-blind, placebo-controlled of paroxetine hydrochloride (Paxil) in controlling hot flashes. Breast study. Maturitas 1997; 9: 227–34.
Cancer Res Treatment 1997; 46: 23–33.
Nesheim BI, Saetre T. Reduction of menopausal hot flashes by 29 Loprinzi CL, Quella SK, Sloan JA, et al. Preliminary data from a methyldopa: a double-blind crossover trial. Eur J Clin Pharmacol 1981; randomized evaluation of fluoxetine (Prozac) for treating hot flashes in 20: 413–16.
breast cancer survivors. Breast Cancer Res Treat 1999; 57: 34.
THE LANCET • Vol 356 • December 16, 2000 For personal use only. Not to be reproduced without permission of The Lancet.

Source: http://files.thiagorego.webnode.com.br/200000074-415cc42572/HTX%20fogachos%20vanlafaxina.pdf