Paroxetine Controlled Release in the Treatment of Menopausal Hot Flashes: A Randomized Controlled Trial

Context Standard therapy for hot flashes has been hormone replacement with estradiol or progestational agents, but recent data suggest that antidepressants inhibiting serotonin reuptake may also be effective.

Objective To evaluate a selective serotonin reuptake inhibitor (paroxetine controlled release [CR]) in treating the vasomotor symptoms displayed by a general cross-section of menopausal women.

Design and Setting Randomized, double-blind, placebo-controlled, parallel group study conducted across 17 US sites, including urban, suburban, and rural clinics.

Patients A total of 165 menopausal women aged 18 years or older experiencing at least 2 to 3 daily hot flashes and must have discontinued any hormone replacement therapy for at least 6 weeks. Women were excluded if they had any signs of active cancer or were undergoing chemotherapy or radiation therapy.

Intervention After a 1-week placebo run-in phase, study participants were randomized to receive placebo or receive 12.5 mg/d or 25.0 mg/d of paroxetine CR (in a 1:1:1 ratio) for 6 weeks.

Main Outcome Measures Mean change from baseline to week 6 in the daily hot flash composite score (frequency × severity).

Results Fifty-six participants were randomly assigned to receive placebo and 51 to receive 12.5 mg/d and 58 to receive 25.0 mg/d of paroxetine CR. The mean reductions in the hot flash frequency composite score from baseline to week 6 were statistically significantly greater for those receiving paroxetine CR than for those receiving placebo. By week 6, the mean daily hot flash frequency went from 7.1 to 3.8 (mean reduction, 3.3) for those in the 12.5-mg/d and from 6.4 to 3.2 (mean reduction, 3.2) for those in the 25-mg/d paroxetine CR groups and from 6.6 to 4.8 (mean reduction, 1.8) for those in the placebo group. Mean placebo-adjusted reduction in hot flash composite scores were −4.7 (95% confidence interval, − 8.1 to −1.3; P = .007) comparing 12.5-mg/d paroxetine CR with placebo; and −3.6 (95% confidence interval, −6.8 to −0.4; P = .03) comparing 25.0-mg/d paroxetine CR with placebo. This corresponded to median reductions of 62.2% for those in the 12.5-mg/d and 64.6% for those in the 25.0-mg/d paroxetine CR groups compared with 37.8% for those in the placebo group.

Conclusion Paroxetine CR may be an effective and acceptable alternative to hormone replacement and other therapies in treating menopausal hot flash symptoms.

For many years, hormone replacement therapy (HRT) with combined estrogen/progestin has been the standard therapy for women experiencing menopausal symptoms. However, increased risks of long-term adverse clinical outcomes in a recent prospective study conducted by the Women's Health Initiative (WHI) on long-term HRT use, are likely to change clinical practice significantly.¹ The long-term benefits of HRT regimens have been called into question; thus, alternative treatments are needed.

Hot flashes are the most common complaint among women entering menopause and, for many women, may continue to occur for up to 5 years (although about 20% of women may have them for up to 15 years).² Approximately 75% of perimenopausal women will experience some hot flashes, with 10% to 20% of those enduring severe symptoms.³ Based on these figures, today more than 25 million women in the United States alone may have experienced symptoms, of which 4 million women reported severe symptoms.⁴ Despite this prevalence, the physiology of hot flashes is not fully understood although a disturbance in normal thermoregulatory function is thought to be the main underlying cause. The primary symptom is a subjective and transient sensation of heat that usually lasts 4 to 10 minutes, which may be accompanied by differing degrees of flushing, palpitations, anxiety, irritability, and panic, a rare occurrence.²

Women who undergo chemotherapy for breast cancer or who are subsequently prescribed antiestrogens such as tamoxifen may have a 2- to 3-fold increased risk of developing hot flashes.^5-7 However, there has been a reluctance to use the current standard therapy, estrogen, in treating menopausal symptoms of women with a prior breast cancer or in treating those having a high risk of developing breast cancer. The reluctance to use estrogen to treat patients with breast cancer has also extended to progestogens, despite their proven efficacy (low-dose megestrol acetate has been shown to reduce hot flashes by about 80%).⁸ Furthermore, the recent data from the WHI indicate that this concern may be justified. In this 5.2-year follow-up study of more than 16 000 healthy postmenopausal women receiving combined estrogen (0.625 mg/d) and progestin (medroxyprogesterone acetate, 2.5 mg/d), the overall health risks appeared to exceed the benefits. The absolute excess risks per 10 000 person-years were 7 more coronary events, 8 more strokes, 8 more pulmonary embolisms, and 8 more invasive breast cancers vs reductions of 6 fewer colorectal cancers and 5 fewer hip fractures.¹ These data suggest that, in the long-term, replacement strategies offer little improvement on normal ovarian aging other than to ameliorate vasomotor symptoms and vulvovaginal atrophy.⁹

The perceived limitations of HRT, coupled with the lack of efficacy and adverse effects observed with nonhormonal therapies, have led clinicians to search for other treatment options. Recent studies of venlafaxine and fluoxetine in women with a prior history of breast cancer have suggested that certain antidepressants with the ability to inhibit serotonin reuptake may significantly reduce vasomotor symptoms of menopause.^10-12 The clinical benefit of these antidepressants is not as great as that observed for estrogen; however, the benefit appears to be greater than what is offered from other nonhormonal pharmacological approaches¹³ and from nonpharmacologic approaches such as vitamin E.¹⁴

The pilot study that my colleagues and I conducted¹² demonstrated that 5 weeks of treatment with immediate-release paroxetine (20 mg/d) reduced the hot flash composite score of breast cancer survivors by 75%, suggesting that further investigation was warranted. We conducted what is, to our knowledge, the first study of a selective serotonin reuptake inhibitor (SSRI) (paroxetine controlled release [CR]) in treating the menopausal vasomotor symptoms displayed by a group of women who were not primarily breast cancer survivors. We selected paroxetine CR for this study because it is a better tolerated formulation that has lower rates of early discontinuation due to adverse events.¹⁵

Patients recruited to the study were menopausal women aged 18 years or older who had been: amenorrheic for at least 12 consecutive months, amenorrheic for 6 months but met the biochemical criteria for menopause (follicle-stimulating hormone >40 mlIU/mL and estradiol <20 pg/mL [69.34 pmol/L]), or had undergone bilateral oophorectomy at least 6 weeks before screening. To be included in the study, patients also must have experienced a minimum of 2 to 3 daily hot flashes or at least 14 bothersome hot flashes per week and must have discontinued any HRT at least 6 weeks before screening. Psychotropic drugs must have been discontinued for a specified period prior to screening: 2 weeks for tricyclic antidepressants, selective noradrenaline reuptake inhibitors, SSRIs (other than fluoxetine), lithium and oral neuroleptics, all sedatives and hypnotics; 4 weeks for fluoxetine and monoamine oxidase inhibitors; 12 weeks for depot neuroleptics.

Women were excluded from the study if they presented with signs of an active cancer or were receiving current chemotherapy or radiation therapy. Treatment with selective estrogen receptor modulators (SERMs, eg, tamoxifen) was permitted on the condition that therapy had been initiated at least 3 months before screening and that the dose remained unchanged throughout the study. Other grounds for exclusion were: an active psychiatric disorder, concurrent major depression, intolerance to SSRIs, and substance dependence.

Institutional review board–approved written informed consent was obtained from all patients, and the study was conducted in accordance with the Declaration of Helsinki, 1996 (South Africa amendment). Before starting the study medication, medical histories were taken and physical examinations were performed for each woman. Other evaluations included vital signs, electrocardiogram, Mini International Neuropsychiatric Inventory (MINI),¹⁶ and the Beck Depression Inventory II (BDI-II).¹⁷ Patients with clinically significant mood or anxiety symptoms were excluded to permit the independent examination of the effect of paroxetine CR on vasomotor symptoms.

This was a double-blind, placebo-controlled, parallel group study conducted in 17 US sites, including urban, suburban, and rural clinics. Recruitment techniques included newspaper advertisements, presentations to women's groups, and professional referral networks. After the initial screening visit, patients entered a 1-week single-blind, placebo run-in phase to obtain baseline diary data (hot flash frequency and severity) to ensure that each woman met the minimum eligibility criteria, and to screen out any potentially high-placebo responders and thereby exclude those women who would not necessarily benefit from pharmacological intervention. A baseline visit was scheduled to confirm eligibility within 2 days of completing the run-in phase, after which patients were randomized to receive placebo or to receive 12.5 mg/d or 25.0 mg/d of paroxetine CR (in a 1:1:1 ratio) for the 6-week, double-blind treatment phase (Figure 1). Study visits were scheduled for 1, 3, and 6 weeks. Symptom-assessment questionnaires were administered at each visit, and adverse events and vital signs were monitored. At the end of the study period, patients were treated at the discretion of their health care professional.

Daily hot flash diaries, as developed and validated previously^8,14,18,19 were used to document the frequency and severity of hot flashes at baseline and during weeks 1 through 6 of the study. Hot flash composite scores were calculated from the product of the daily frequency and severity ratings.¹⁸

Questionnaires (administered at weeks 1, 3, and 6) were used to assess symptoms and problems commonly related to the menopause and potential adverse reactions to study treatment. The BDI-II (21 items) was used both to screen for patients with major depression and to monitor depression during the study. A second instrument, the MINI, was used as a screening tool to prevent patients with primary psychiatric disorders (requiring treatment) from entering the study.

In terms of symptom measurement, the Greene Climacteric Scale (GCS, 21 items) was used to assess core menopausal symptom severity including hot flashes, night sweats, somatic symptoms, and sexual interest. A separate measure, the sleep disturbance visual analog scale (VAS) was used to assess sleep disturbance in the preceding week. Anxiety (eg, feeling dizzy or lightheaded, nervous or having difficulty breathing) was measured via the Beck Anxiety Inventory-II (BAI-II, 21 items). General disability was monitored using the Sheehan Disability Scale total score and scores on the work, social life and leisure activities, and family life and home responsibilities subscores. The Clinical Global Impression (CGI) global improvement item was used to measure the overall improvement from baseline in patients' health.

The primary objective of this study was to compare the mean change from baseline to week 6 in the daily hot flash composite score among women taking 25.0 mg/d of paroxetine CR with those taking placebo. Secondary objectives of this study were to compare the mean change from baseline to week 6 in the daily hot flash composite score of those taking 12.5 mg/d of paroxetine CR with those taking placebo, and to assess the safety and tolerability of paroxetine CR in the treatment of hot flashes associated with menopause. The composite score was calculated by assigning a number to the severity of the hot flash (mild = 1, moderate = 2, severe = 3, very severe = 4) and multiplying by the daily number of hot flashes experienced at that severity level.¹⁸ The 4 resulting numbers were added to give a daily score, and each mean daily score was calculated over the 7 days preceding the latest dose of the study medication.

Other secondary end points included the mean weekly change in hot flash score for weeks 1 through 6, the proportion of hot flash score responders (≥50% reduction in score at study end), the mean change from baseline in questionnaire score (BDI-II, BAI-II, GCS, sleep disturbance VAS, Sheehan Disability Scale), and the proportion of CGI responders (patients achieving a score of 1 or 2 on the clinician-rated CGI global improvement item). Paroxetine CR tolerability was assessed throughout the study by vital sign and adverse event monitoring.

A sample size of 150 women was necessary to give the study 85% power to detect a difference of 3 points between 25.0 mg/d paroxetine CR and placebo (type I error = .05) in the mean change from baseline in daily hot flash composite score, with an SD of 5 points. For the generation of the randomization list, a randomized block of size 6 for the 3 treatments (to achieve a balance) in a 1:1:1 ratio was used. Stratification was not used in this trial. Each site registered in the trial received drugs for a maximum of 36 patients based on this randomization list (6 blocks of random numbers corresponding to the 3 treatments were reserved per site.) The efficacy analyses were performed on all patients who received at least 1 dose of the study medication. Where appropriate, data were assessed for normality and homogeneity, and the last observation carried forward procedure was used to impute missing data. Primary analyses were adjusted for treatment and site and were based on a 2-sided hypothesis at the 5% significance level. No adjustment was made for multiple comparisons. All continuous efficacy variables were analyzed using parametric analysis of variance with treatment and site effects. The proportion of responders was analyzed by a logistic regression model with treatment and site effects. All analyses were performed using SAS statistical software, version 6.12 (SAS Inc, Cary, NC).

From October 2001 to March 2002, 225 women were screened to enter the placebo run-in phase of the study, and 171 entered the placebo run-in period. Of these, 6 women did not meet the required daily number of hot flashes for study entry; thus, 165 were randomized to receive treatment (Figure 1). There were no exclusions on the basis of placebo effect during the run-in. In all, 139 (84.2%) of 165 of the participants randomized completed the 6-week treatment phase. Withdrawals included 14 women who discontinued study medication due to adverse effects (2 placebo, 12 paroxetine CR), and 1 woman receiving placebo discontinued due to lack of efficacy.

The 3 treatment groups were well matched in terms of patient characteristics (Table 1). Unlike previous studies, this was a general population of women, with only 12 (7.3%) with a history of breast cancer. The aim of this study was to examine the effect of paroxetine CR on vasomotor symptoms alone, which is supported by the fact that only 17 patients met the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for psychiatric disorders, as assessed by the MINI (6 paroxetine CR, 11 placebo). In all, 95.2% had 12 or more years of education and 81.2% experienced hot flashes for 12 months or more. Only 7 patients were receiving concurrent tamoxifen treatment and 5 reported concurrent raloxifene use.

There were mean reductions from baseline to week 6 in daily hot flash composite score of 8.4 and 7.2 for those taking 12.5 mg/d and 25.0 mg/d of paroxetine CR, respectively (Figure 2). Both treatment groups showed a significant benefit of paroxetine CR over placebo. The mean last observation carried forward placebo-adjusted reductions for paroxetine CR were −4.7 (95% confidence interval [CI] −8.1 to −1.3; P = .007 vs placebo) and −3.6 (95% CI, −6.8 to −0.4; P = .03 vs placebo) for the 12.5-mg/d and 25.0-mg/d groups, respectively. After 6 weeks of treatment, the hot flash composite score was reduced by 62.2% and 64.6% for the lower- and higher-dose paroxetine CR groups compared with a 37.8% reduction for placebo. In a separate analysis, this treatment difference remained even after adjustment for age, disease history (breast cancer or psychiatric disorders), or antiestrogen use.

Compared with placebo, the mean weekly change in hot flash composite score showed that significant improvements that continued to study end were evident within a week of treatment for those taking 25.0 mg/d of paroxetine CR. However, for those taking 12.5 mg/d of paroxetine CR, the mean reduction in the hot flash composite score compared with placebo was statistically significant in weeks 1, 3, 5, and 6 (Figure 3). The mean baseline hot flash composite scores were 16.5 (paroxetine CR 12.5 mg/d), 13.6 (paroxetine CR 25.0 mg/d), and 14.2 (placebo). Although the mean reduction from baseline at week 6 appears greater for those receiving 12.5 mg/d of paroxetine CR, the difference in baseline scores means that a smaller percentage reduction was seen among those in the lower dosage group.

By week 6, the mean daily hot flash frequency decreased from 7.1 to 3.8 (mean reduction, 3.3) for those in the 12.5-mg/d and from 6.4 to 3.2 (mean reduction, 3.2) for those in the 25-mg/d paroxetine CR groups and from 6.6 to 4.8 (mean reduction, 1.8) for those in the placebo group. The adjusted mean differences for 12.5 mg/d and 25.0 mg/d of paroxetine CR are −1.55 (95% CI, −2.75 to −0.34; P = .01) and −1.50 (95% CI, −2.66 to, −0.34; P = .01), respectively. Of 104 women with baseline and week 6 values who received paroxetine CR, more than half experienced a 50% or more reduction in hot flash frequency and severity and were considered responders. The response rates were 58.3% for those receiving 12.5 mg/d and 62.5% for those receiving 25.0 mg/d of paroxetine CR compared with 42.9% of those receiving placebo. The odds of being a responder while taking 12.5 mg/d of paroxetine CR were almost twice that of those taking placebo (odds ratio [OR], 1.95; 95% CI, 0.86-4.40; P = .111). Furthermore, the odds of achieving a response were more than 2.5 times greater for women receiving 25.0 mg/d of paroxetine CR than for those receiving placebo (OR, 2.56; 95% CI, 1.15-5.68; P = .02).

The improvements in hot flash symptoms among those taking paroxetine CR were independent of any significant changes in mood or anxiety symptoms, as shown by symptom-assessment questionnaire scores (Table 2). By week 6 the GCS total score improved by 2.0 points for those in the 12.5-mg/d and 3.3 points for those in the 25.0-mg/d paroxetine CR groups, mainly due to a significant improvement in vasomotor symptoms. Scores on the GCS item for sexual interest were minimally changed for all treatment groups after 6 weeks of treatment. In addition, more than 50% of patients in each of the paroxetine CR treatment groups were considered responders as assessed by the CGI global improvement rating. The odds of being a CGI responder were 4.39 (95% CI, 1.78-10.84) times greater for those in the 12.5-mg/d and 4.33 (95% CI, 2.16-10.54) times greater for those in the 25.0-mg/d paroxetine CR groups than for those in the placebo group (P = .001).

A total of 30 patients (53.6%) receiving placebo reported an adverse event during the study, compared with 63 patients (58.3%) randomized to receive paroxetine CR. Events experienced by patients receiving active treatment were in line with the known tolerability profile of paroxetine CR. The most frequently reported events for paroxetine CR were headache, nausea, and insomnia (Table 3) with fewer reports overall from patients receiving the lower dose of paroxetine CR. In both dose groups, the majority of adverse events reported were mild or moderate in severity (approximately 89%). Ten (20%) of 50 patients receiving 12.5 mg/d and 18 (31%) receiving 25.0 mg/d of paroxetine CR experienced adverse events that were considered possibly or probably related to their medication. Of these patients, a total of 12 (4 in the 12.5-mg/d and 8 in the 25.0-mg/d paroxetine CR groups) did not complete the study.

We have previously demonstrated that the immediate-release formulation of paroxetine may be effective in treating hot flashes and their associated symptoms in women with a history of breast cancer. The purpose of the current study was to determine the extent to which paroxetine CR can relieve hot flashes in a general population of menopausal women. A substantial reduction in hot flash symptoms was observed following 6 weeks of paroxetine CR treatment. Hot flash composite scores were reduced by 62.2% in those receiving 12.5 mg/d and 64.6% in those receiving 25.0 mg/d of paroxetine CR. Although head-to-head comparisons are not available, our results compare favorably with the findings of similar studies for fluoxetine (50% reduction over 4 weeks at 20 mg/d) and venlafaxine (61% reduction over 4 weeks at 75 mg/d and 150 mg/d; 37% over 4 weeks at 37.5 mg/d).^10,11 In our study, absolute reductions in hot flash composite score were similar regardless of paroxetine CR dose level, suggesting that 12.5 mg/d is an adequate and well-tolerated starting dose for most women.

At study end, 63 (60.5%) of 104 women who received paroxetine CR achieved a 50% or greater reduction in their hot flash composite scores. At the start of the study, women were experiencing an average of 6.5 hot flashes per day. By week 6, the average number of daily hot flashes was 3.8 for those in the12.5-mg/d and 3.2 for those in the 25.0-mg/d paroxetine CR groups and 4.8 for those in the placebo group. In addition, 29%, 30%, and 19.6% of women in the respective groups did not experience any hot flashes during week 6. A mean reduction in the hot flash composite score and treatment response showed a consistent pattern among those taking 25.0 mg/d of paroxetine CR. It is reassuring to note that significant improvements over placebo were also seen for paroxetine CR on the CGI and GCS (vasomotor symptoms subscore) questionnaires, since they assess symptoms that overlap with those detected by the primary efficacy measure. In addition, the CGI is a clinician-rated measure in contrast to the patient-assessed hot flash score. The odds of being rated as improved or very much improved on the CGI were around 4 times greater for patients receiving paroxetine CR than for those receiving placebo. Taken together with the findings of the WHI, these data suggest that paroxetine CR may have a place in the treatment armamentarium for women experiencing hot flashes.

As in studies of related agents,^10,11 these central efficacy findings are largely independent of any effect that paroxetine CR may have on mood or anxiety symptoms because patients entering the study were not permitted to have clinically significant disorders of this kind. Furthermore, levels of depression and anxiety were much lower among those participating in our study than what was reported in patients who entered a recent study of fluoxetine in the treatment of hot flashes,¹⁰ and the majority of patients in our study who had such symptoms could be graded at the minimal level by the BDI-II (93%) and BAI-II (78%).

In general, paroxetine CR was well tolerated. Rates of commonly reported adverse events were generally lower than those reported in a recent study examining the use of paroxetine CR in treating a depressed population.¹⁵ Although both doses of paroxetine CR were associated with a similar magnitude of hot flash reduction, the lower dose was better tolerated. Because the slow release of this formulation was associated with a low drop out rate due to adverse events in this depression trial, it could potentially improve compliance and may be particularly suited for mid-term to long-term treatment of menopausal women. The majority of adverse events in our study were mild or moderate and paroxetine CR appeared to have little or no effect on sexual function over 6 weeks. This is in agreement with a previous pilot study of this agent.¹²

The main weakness of this study is that no direct comparisons were made in terms of associated menopausal symptoms such as insomnia. Furthermore, there were low proportions of black and Asian women in the study population, and it is likely that the recruitment methods used for the study did not target particular ethnic groups effectively. Since black women are known to exhibit more severe menopausal symptoms (particularly vasomotor symptoms) and Asian women experience less severe symptoms, it would be interesting to compare treatment responses among different racial groups.²⁰

Although surveys such as the Study of Women's Health Across the Nation (SWAN) have greatly expanded our knowledge of the demographic and lifestyle factors affecting the symptoms of menopause, little is still known about the physical basis of hot flashes.²⁰ Hot flashes are thought to occur as the result of an alteration in the central nervous system thermoregulatory set point caused by falling estrogen levels. However, the basis for the beneficial effect of SSRIs on vasomotor symptoms remains unknown. Evidence from animal studies suggests that serotonin (5-HT) plays an important role in thermoregulation and that the temperature increases associated with hot flashes could be linked to an overloading of serotonin receptor sites in the hypothalamus.^21-23 Two 5-HT receptor subtypes, 5-HT1a and 5-HT2a, have been closely associated with temperature control in mammals.²⁴ These receptors appear to have opposite effects on temperature regulation, with 5-HT2a mediating hyperthermic effects and 5-HT1a mediating hypothermic effects.¹³ It is likely that a balance between these 2 receptors is important in maintaining optimal thermoregulation. Notably, the expression and activity of 5-HT receptors can be modulated by gonadal hormones and adrenal corticosteroids, and this may be the functional link between some of the hormonal systems linked to hot flashes and serotonin.^25,26 Further work is needed to clarify the mechanisms behind hot flashes and to explain fully the mode of action of current therapies.

This study provides new information on the treatment of hot flashes. Based on these results and the preceding pilot trial of Stearns and colleagues,¹² paroxetine CR is a well-tolerated alternative to HRT and other therapies in the treatment of hot flash symptoms. However, the duration of benefit of paroxetine CR in treating menopausal hot flashes and its applicability as a first-line or second-line treatment remain unknown. Furthermore, the question remains of whether women who are resistant to 1 antidepressant will respond to another. The optimal dose for treating hot flashes also remains to be determined and may be lower than that recommended in depression.