NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Nelson HD, Haney E, Humphrey L, et al. Management of Menopause-Related Symptoms. Rockville (MD): Agency for Healthcare Research and Quality (US); 2005 Mar. (Evidence Reports/Technology Assessments, No. 120.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Management of Menopause-Related Symptoms

Management of Menopause-Related Symptoms.

Show details


To address Key Questions 1 and 2, 6,342 abstracts were reviewed to identify relevant studies evaluating menopause related symptoms and their associations with the menopausal transition. The review focused on prospective studies of cohorts of midlife women transitioning through the stages of menopause. Forty-nine studies conducted among 14 cohorts met inclusion criteria. Seven cohorts were based in the U.S. (Massachusetts Women's Health Study, 20–25 Seattle Midlife Women's Health Study,26–28 Ohio Midlife Women's Study,29 National Health Examination Follow-up Study [NHANES],30 Study of Women's Health Across the Nation [SWAN],31–39 University of Minnesota/Tremin Trust Longitudinal Study [includes a small subset from other countries],40 and Pennsylvania Ovarian Aging Study41, 42). Seven cohorts were based outside the U.S. (Gothenburg, Sweden,43–45 Australian Longitudinal Study on Women's Health,46, 47 Medical Research Council [MRC], U.K.,48–51 Melbourne Women's Midlife Health Project, Australia, 52–62 Manitoba Project on Women and Their Health in the Middle Years, Canada,63 Copenhagen, Denmark,64–66 and Eindhoven, Netherlands.67, 68). Cohorts are described in Table 2, results are listed by symptoms in Table 3, and full details are provided in Appendix 6-1.

Table 2. Descriptions of cohort studies.

Table 2

Descriptions of cohort studies.

Table 3. Results of cohort studies of menopause related symptoms.

Table 3

Results of cohort studies of menopause related symptoms.

Major limitations of cohort studies include dissimilar methods for evaluating and reporting symptoms and for assessing menopausal change. Some studies based results on cross-sectional data reported at serial time points rather than individual tracking of women over time. Some studies failed to adjust or stratify for potentially important variables such as age, race, BMI, life events, or history of depression when attempting to attribute symptoms to change in menopausal stage. Although most included studies were population-based, in many cases, enrolled women were additionally selected from the initial recruited cohort and may have been less representative of the general population.44, 52, 53, 55, 57–59 Also, many studies were based on cohorts recruited from community populations and are more representative of volunteers than entire communities.

A total of 51 cross-sectional studies meeting similar inclusion criteria were obtained to provide additional prevalence data including 29 studies from the 14 study cohorts21–25, 27, 30, 33–40, 42, 45, 48–50, 56, 60–66, 69 and 22 studies from other populations.70–91 Results of cross-sectional studies are summarized in Table 4.

Table 4. Results of cross-sectional studies of menopause related symptoms.

Table 4

Results of cross-sectional studies of menopause related symptoms.

Key Question 1. Symptoms Associated with Menopause

What is the evidence that the symptoms more frequently reported by middle-aged women are attributable to ovarian aging and senescence?

Vasomotor Symptoms

Four studies from four population-based cohorts evaluated the association between vasomotor symptoms and menopausal stage, including three rated good-quality46, 51, 53and one fair.26 Women transitioning from premenopause to either perimenopause or postmenopause had increased vasomotor symptoms in the three good-quality studies.46, 51, 53

The prevalence of hot flashes was reported in 33 cross-sectional studies.25, 27, 33, 35–37, 40, 48, 60, 62, 64, 66–67, 69–76, 78–85, 87–90 Prevalence rates ranged from 14 percent to 51 percent for premenopausal women, were approximately 50 percent for perimenopausal women, and between 30 percent to 80 percent for postmenopausal women, depending on the age and population studied. In one study, approximately 29 percent of women age 60 years reported hot flashes, suggesting only a modest decline in frequency with increasing time from menopause.66

Vaginal Dryness

Vaginal dryness was associated with menopause, and became more common with the transition from premenopause to postmenopause, in the one good-quality cohort study examining this relationship.53 Thirteen cross-sectional studies37, 40, 48, 69, 71, 72, 74–76, 80, 83–85 reported prevalence rates ranging from 4 percent to 22 percent for premenopausal women, 7 percent to 39 percent for perimenopausal women, and 17 percent to 30 percent for early postmenopausal women. Two cross-sectional studies showed either no association,84 or a weak association,83 between vaginal dryness and menopausal stage.

Sleep Disturbance

Three cohort studies evaluated the impact of menopausal status on sleep; two were rated good-quality46, 53 and one fair.26 In the good-quality studies, women had more difficulty sleeping as they progressed through the menopausal stages,46, 53 and sleeping difficulty was associated with the presence of hot flashes.53 The fair-quality study reported no association.26

Eighteen cross-sectional studies evaluated the prevalence of sleep disturbance at different menopausal stages.25, 35–38, 48, 60, 69, 78–82, 84, 85, 87, 90, 92 Of these, eight reported prevalence rates ranging from 16 percent to 42 percent for premenopausal women, 39 percent to 47 percent for perimenopausal women, and 35 percent to 60 percent for postmenopausal women (including women with either surgical or natural menopause and hormone therapy users).37, 60, 78, 80–82, 85, 90 Some studies,79, 82, 87 but not others,38, 90 reported an association between hot flashes and sleep disturbance. When hot flashes/night sweats and insomnia were considered together in one study, 27 percent of premenopausal women, 46 percent of perimenopausal women, and 38 percent of postmenopausal women reported disturbances.25

Four cross-sectional studies reported increased adjusted odds ratios for difficulty sleeping for menopausal women compared to premenopausal women37, 48, 69, 92 In one study, an age-adjusted significant correlation between sleep problems and menopausal status was reported, but increases among perimenopausal and postmenopausal women were small, and social class was a stronger predictor.69 A factor analysis of sleep-related problems found sleep was not associated with menopausal status.84 In another study, sleep disordered breathing increased across the menopausal stages after multivariate adjustment and stratification by BMI and age.92

Mood Symptoms

Fourteen studies from 11 cohorts evaluated the impact of menopausal stage on mood.20, 24, 26, 28–30, 41, 44, 47, 51, 54, 55, 63, 68 Of these, seven were rated good-quality,29, 41, 47, 51, 54, 55, 68 four fair,20, 26, 28, 63 and three poor. 24, 30, 44 Symptoms included anxiety,29 depression,20, 24, 28–30, 41, 63, 68 dysphoric or negative mood,26, 54 and positive mood.55 Development of a mental disorder,44 psychological symptom reporting,51 and general mental health47 were also assessed.

Twelve of the 14 studies, including five good-quality, four fair, and three poor, found no influence of menopausal stage on mood symptoms, development of a mental disorder, or general mental health.20, 24, 26, 28–30, 44, 47, 51, 54, 55, 63 Two good-quality studies demonstrated a change in mood symptoms with the menopausal transition.41, 68 A well-designed, nested case-control study of age-matched menopausal and premenopausal women from the Pittsburgh cohort found no differences between the groups in depression, anger, anxiety, or stress symptoms.93

Twenty-four cross-sectional studies reported relevant data.22–24, 33, 35, 36, 48, 50, 60, 63, 64, 66, 67, 69, 70, 72, 76, 80–85, 90 The prevalence of depressed mood ranged from 8 percent to 37 percent for premenopausal women, 11 percent to 46 percent for perimenopausal women, 8 percent to 47 percent for women with natural menopause, and 8 percent to 38 percent for women with surgical menopause.23, 66, 67, 80, 85

Cognitive Disturbances

No cohort studies longitudinally assessed cognitive symptoms. Eight cross-sectional studies showed no differences in the prevalence rates of memory or concentration problems by menopausal stage.37, 48, 67, 69, 80, 81, 84, 85

Somatic Complaints

Somatic complaints were evaluated in four studies conducted among three cohorts rated good-quality46, 47, 53 or fair-quality.26 Perimenopausal women had increased somatic symptoms compared to premenopausal women in one cohort.46, 47 Symptoms included back pain, severe tiredness, stiff or painful joints, bodily pain, and worse general health. Other studies reported reduced breast tenderness,53 or no association with somatic or neuromuscular symptoms.26

Fourteen of 18 cross-sectional studies evaluating changes in somatic symptoms showed no significant increases as women transitioned from premenopause to postmenopause in adjusted analytic models.36, 37, 48, 64, 66, 67, 69, 71, 72, 81–85, 87 Three studies showed slight increases in somatic symptoms during perimenopause that decreased during postmenopause,78, 89 or remained stable.67 Another study showed significantly fewer somatic symptoms for women at menopause compared with those at premenopause.90

Urinary Complaints

Three good-quality cohort studies evaluated the impact of menopausal stage on urinary symptoms.46, 53, 56 In one study, women who were perimenopausal at baseline were more likely to report leakage of urine two years later than women who remained premenopausal.46 Two studies from the Melbourne cohort found no association of menopausal status with urinary symptoms.53, 56

Fourteen cross-sectional studies37, 45, 48, 49, 56, 60, 67, 71, 75, 77, 84–87 provided relevant data. Prevalence rates of urinary incontinence or leakage ranged from 10 percent to 36 percent for premenopausal women, 17 percent to 39 percent for perimenopausal women, 14 percent to 36 percent for women undergoing natural menopause, and 18 percent to 36 percent for women undergoing surgical menopause.37, 46, 56, 67, 86

Uterine Bleeding Problems

No studies meeting inclusion criteria addressed uterine bleeding problems, most likely because currently accepted definitions of menopause rely historically on changes in uterine bleeding.

Sexual Dysfunction

Sexual dysfunction was evaluated in two fair-quality studies from the Melbourne cohort.57, 58 Sexual function was variably measured and included decreased responsivity, decreased libido, and/or decreased frequency of sexual activity. In that cohort, women demonstrated decreases in some or all of the sexual parameters as they transitioned from early to late perimenopause or postmenopause.57, 58

Thirteen cross-sectional studies, including one from the Melbourne cohort, also evaluated sexual function.22, 40, 42, 48, 61, 65, 66, 69, 83, 84, 86, 89 Five studies showed the prevalence of sexual disinterest, dysfunction, or discomfort to increase from premenopause to perimenopause to postmenopause,48, 61, 67, 69, 86 and four showed no change.40, 65, 84, 89

Reduced Quality of Life

Four cohort studies addressed the impact of menopausal stage on quality of life, three rated good-quality47, 52, 59 and one poor.30 Two studies from the Melbourne cohort had conflicting results.52, 59 One study showed declines in well-being across progressive menopausal stages (lowest for the one to two years postmenopause) that improved when women were more than two years postmenopause.52 A second study showed increased well-being from early to late perimenopause and postmenopause.59 In another good-quality study, women who remained perimenopausal over a two year follow-up period reported significant decreases in quality of life compared to women who remained premenopausal, although significant decreases were not reported by women in other stages.47 A poor-quality cohort study found that well-being was not associated with change in menopausal status.30

Three cross sectional studies also evaluated quality of life.22, 34, 74 Satisfaction with life,22 and impaired functioning34 did not vary with menopausal stage in multivariate regression modeling of data from the Massachusetts Women's Health Study. In another study, various aspects of general well-being were lower among postmenopausal women compared to premenopausal women.74


  • Vasomotor symptoms: Evidence from population-based cohort and cross-sectional studies supports the association between vasomotor symptoms and menopausal stage. Studies are consistent in reporting increasing prevalence rates of vasomotor symptoms as women transition from premenopause to either perimenopause or postmenopause, affecting 50 percent or more of women. Studies suggest that vasomotor symptoms persist for several years after menopause for some women.
  • Vaginal dryness: Vaginal dryness is associated with menopause and prevalence rates increase as women transition through the menopausal stages. Estimates indicate that up to one third of perimenopausal and postmenopausal women experience vaginal dryness.
  • Sleep disturbance: Although results of studies are mixed, two good-quality cohort studies indicate that women have more difficulty sleeping as they transition through menopausal stages, and this may be due to vasomotor symptoms. Up to 40 percent to 60 percent of perimenopausal and postmenopausal women experience sleep disturbance, a slight increase from prevalence rates of premenopausal women.
  • Mood symptoms: The majority of studies from a large literature report no associations between menopausal stage and mood symptoms, development of a mental disorder, or general mental health. Studies of prevalence rates report wide ranges that are similar across menopausal stages.
  • Cognitive disturbances: The few studies evaluating cognitive disturbances in menopause have inadequate measures.
  • Somatic complaints: Most studies report no association of somatic symptoms with menopause, although somatic symptoms were increased among perimenopausal women compared with premenopausal women in one cohort and two cross-sectional studies.
  • Urinary complaints: Urinary leakage increased among perimenopausal women compared with premenopausal women in one study, and another reported no associations. Studies of prevalence rates report wide ranges that are similar across menopausal stages.
  • Uterine bleeding problems: No studies meeting inclusion criteria addressed uterine bleeding problems, most likely because currently accepted definitions of menopause rely historically on changes in uterine bleeding.
  • Sexual dysfunction: Women from two study cohorts reported declines in some or all of the measured sexual parameters as they transitioned through menopausal stages. Results of cross-sectional studies are mixed.
  • Reduced quality of life: Results of available cohort and cross-sectional studies are conflicting.

Key Question 2. Factors Influencing Symptoms

When do the menopausal symptoms appear, how long do they persist and with what frequency and severity, and what is known about the factors that influence them?

Included studies do not characterize the severity and duration of specific symptoms, and frequency is described by prevalence data in Key Question 1.

Race and Ethnicity

Two cohort studies38, 41 addressed the influence of race and ethnicity on menopausal symptoms, while many adjusted for it in multivariable models. Depression was reported more often among African American women compared to Caucasian women,41 and sleep difficulties were reported more often among Caucasian and Hispanic women and less often among African American, Chinese, and Japanese women.38

Seven cross-sectional studies conducted in the large SWAN cohort also evaluated symptoms by race or ethnicity.33–37, 39, 70 In SWAN, vasomotor symptoms were reported more often among African American women and less often among Hispanic, Chinese, and Japanese women compared with Caucasian women.33, 37 Japanese, Chinese, African American, and Hispanic women reported fewer mood symptoms compared to Caucasian women,33, 35, 36 although “impaired social function” was reported among Hispanic women in two studies34, 70 and among African American women in one study.34 Hispanic women also reported more bodily pain than Caucasian women.34 A cross-sectional study conducted outside the SWAN cohort found that Mexican American women had similar prevalence rates of vasomotor symptoms and vaginal dryness as Caucasian women.72

Age at Onset of the Menopause Transition

Long term studies evaluating menstrual cycles over many years among large cohorts of women showed the median onset of menstrual irregularity to begin between the ages of 45 and 47.5 years with a mean duration of the menopausal transition lasting 5 years.94–96 Factors associated with earlier menopause include smoking,97–99 shorter cycle length during menstruating years,100 and shorter stature and leaner body weight.101 Factors associated with later age at menopause include higher gravidity or parity98, 99 and higher BMI.99

A cross-sectional study reported increased rates of hot flashes among women with menopause prior to age 53 after adjustment for other factors.88 Another found that earlier age of inception of perimenopause was related to a longer perimenopause stage.21

Body Mass Index (BMI)

Two good-quality cohort studies evaluated BMI and its effect on symptoms among menopausal or perimenopausal women and found that BMI was a significant predictor of urinary incontinence,56 but not positive mood.55 A cross-sectional study reported a slight increase in vasomotor symptoms (15 percent) among women with a BMI of 27 or greater, although, vasomotor symptoms were less common at the extremes of BMI.37 Lower BMI was independently associated with increased hot flashes in another cross-sectional study.88 Three cross-sectional studies22, 42, 74 found no significant differences for increased BMI. Other studies evaluating menopausal symptoms adjusted for BMI in multivariate models but did not report the contribution of BMI to symptoms.46, 47, 53, 92

Surgical Versus Natural Menopause

Two good-quality29, 56 and one fair-quality cohort studies63 reported data for women with surgical menopause. The Ohio Midlife Women's Study found that hysterectomy was not a predictor of anxiety or depression.29 In contrast, a community-based study of women in Manitoba, Canada reported that women with hysterectomies had increased odds of depression.63 In the Melbourne Women's Midlife cohort, surgical menopause was not a significant predictor of urinary incontinence.56 Other symptoms were not addressed in these studies.

Sixteen cross-sectional studies reported symptoms among women with surgical and natural menopause.30, 38, 45, 48, 56, 63–66 70, 72, 73, 80, 86, 89, 90 Four studies showed no significant differences in vasomotor symptoms,48, 70, 80, 90 and three studies showed slightly higher rates of vasomotor symptoms among women with surgical menopause (seven percent to eight percent higher).64, 66, 73 Vasomotor symptoms were similar among Japanese and Canadian women with surgical and natural menopause in another study.70 Three studies evaluated vaginal dryness; two showed similar rates,48, 72 and one reported a nine percent higher prevalence among women with surgical compared to natural menopause.80 One cross-sectional study found no differences in rates of sexual dysfunction by type of menopause,48 and another showed higher rates of urinary incontinence (36 percent vs. 22 percent) among women with surgical menopause.86

The prevalence of anxiety and depression was evaluated in four cross-sectional studies,66, 70, 80, 89 and two showed higher prevalence rates among women with surgical menopause.66, 80 In a cross-sectional study from the Massachusetts Women's Health Study evaluating depression, women who had undergone surgical menopause in the previous three months were twice as likely as the rest of the cohort to have elevated depression scores.24


The interaction of baseline depression with menopausal stage and/or transition and the presence or absence of symptoms was not evaluated in any cohort study. A cross-sectional study reported that prior anxiety or depression did not predict menopausal symptoms.48 One other cross-sectional study evaluating menopausal symptoms adjusted for depression in a multivariate model but did not report the contribution of depression to symptoms.92


Three good-quality studies from the Melbourne Women's Midlife cohort examined the association of smoking status with mood symptoms and urinary incontinence.54–56 In this cohort, current smoking was associated with negative mood,54 but not positive mood,55 and was not evaluated as an interaction term with menopausal transition or stage. Smoking was not a significant predictor of urinary incontinence.56 Two good-quality studies adjusted for smoking in multivariate models, but did not report any data for this variable.46, 47

Five cross-sectional studies provide data on smoking and menopausal symptoms.35, 37, 48, 62, 88 Two studies showed no association between smoking and menopausal symptoms,37, 62 one reported that smoking was associated with psychological distress but did not evaluate smoking as an interaction term,35 and another showed that smoking at anytime in the past predicted greater vasomotor symptoms during menopause.48 Another study reported no association between hot flashes and smoking, but identified higher rates of vasomotor symptoms among thin smokers than among thin non-smokers.88 Other studies evaluating menopausal symptoms adjusted for smoking status in multivariate models but did not report the contribution of smoking status to symptoms39, 74, 92


  • Included studies do not provide adequate details to characterize the onset, severity, and duration of specific symptoms. Frequency is described by prevalence data in Key Question 1.
  • Race and ethnicity: The influence of race and ethnicity on menopausal symptoms has not been extensively studied. Prevalence rates of vasomotor and mood symptoms vary among race and ethnic groups in the large SWAN cohort.
  • Age at onset of menopausal transition: Available studies are inconclusive.
  • Body mass index: Available studies are inconclusive.
  • Surgical versus natural menopause: Studies present mixed results regarding the impact of surgical menopause on vasomotor symptoms, vaginal dryness, and mood. Adjustment for confounders is necessary because women undergoing hysterectomy differ from women with natural menopause in ways that may also influence their menopause related symptoms.
  • Depression: Only one cross sectional study was included and reported that prior anxiety or depression did not predict menopausal symptoms.
  • Smoking: Available studies are inconclusive.

Key Question 3. Benefits and Adverse Effects of Therapies

What is the evidence for the benefits and harms of commonly used interventions for relief of menopause-related symptoms?


Three recent good-quality systematic reviews and meta-analyses of randomized controlled trials of estrogen, alone (unopposed) or combined with progestin or progesterone (opposed), describe benefits and harms of various forms, doses, and regimens for treating menopausal symptoms. These include a Cochrane review of trials of oral estrogen for hot flashes,16, 102, 103 a meta-analysis of trials of oral conjugated equine estrogen (CEE) and oral and transdermal estradiol for hot flashes,15, 104 and a review of commonly used types of estrogen for several menopausal symptoms.15 Trials were conducted predominantly in the U.S. or western Europe and recruited participants from primary care or gynecology practices. Most trials focused on healthy menopausal women in their early 50s with baseline symptoms that varied by study. All trials were double blind, of 12 weeks duration or more, and rated good or fair quality.

Vasomotor symptoms. The Cochrane meta-analysis indicated a significant reduction in frequency of weekly hot flashes for oral estrogen compared to placebo with a pooled weighted mean difference of -17.9 hot flashes per week (95% confidence interval [CI], -22.9 to -13.0; nine trials), equivalent to a 75.3 percent reduction in frequency (95% CI, 64.3 percent to 82.3 percent).16 Severity of symptoms was also significantly reduced compared to placebo (odds ratio [OR] 0.13; 95% CI, 0.07 to 0.23; 13 trials). Differences between types of estrogens were not determined, although trials of estradiol and CEE predominated. The review also found that the reduction in weekly hot flash frequency was similar for opposed and unopposed estrogen regimens compared to placebo (opposed: -19.9 per week; 95% CI, -26.7 to -12.6; three trials; unopposed: -14.7; 95% CI, -20.1 to -8.7; six trials). Symptom severity seemed to be better treated by opposed (OR 0.10; 95% CI, 0.06 to 0.19; 10 trials) than by unopposed estrogen (OR 0.35; 95% CI, 0.22 to 0.56; four trials), although heterogeneity of trials could also explain this difference.

Hot flash frequency, severity, or both improved in nine of 10 trials of oral estradiol and in all 11 trials of transdermal estradiol compared to placebo in a review of these agents.15 The pooled weighted mean differences in hot flashes were -16.8 per week (95% CI, -23.4 to -10.2; five trials) for oral estradiol, and -22.4 per week (95% CI, -35.9 to -10.4; six trials) for transdermal estradiol. Results were similar when opposed and unopposed regimens were considered separately.15 All eight trials of oral CEE reported statistically significant improvements in hot flash frequency, severity, or both compared with placebo. One trial of CEE that provided data comparable to the estradiol meta-analyses reported a mean reduction of -19.1 hot flashes per week (95 percent CI, -33.0 to -5.1). Four trials comparing estrogen agents head-to-head (CEE compared with oral or transdermal estradiol) indicated improved number and severity of hot flashes for all treatment groups with no significant differences between them (pooled weighted mean difference in hot flashes -0.3; 95 percent CI, -3.4 to 2.7; three trials).

In a review of 32 trials of estrogen for treatment of vasomotor symptoms, all but five were less than one year in duration and only three trials enrolled more than 500 participants, limiting evaluation of adverse effects.15 Trials reported multiple specific adverse effects including atypical bleeding and endometrial hypertrophy, nausea and vomiting, breast tenderness, headache, weight change, dizziness, venous thromboembolic events, cardiovascular events, rash and pruritus, cholecystitis, liver effects, and other adverse events. These outcomes were reported unevenly across studies and could not be combined in summary statistics. Trials included in the Cochrane review indicated similar adverse effects with breast tenderness and uterine bleeding most commonly reported.16

Urogenital symptoms/sexual dysfunction. In a systematic review of good and fair-quality trials of a variety of estrogen agents, urogenital symptoms improved in eight of nine trials, and sexual function (including pain during intercourse and responses to questionnaires) in eight of 11 trials.15 Vaginal forms of estrogen were evaluated in five trials. In a placebo controlled trial, vaginal dryness and pain during intercourse improved with use of the estradiol intravaginal ring compared to placebo.105 A trial of the estradiol intravaginal ring and oral estradiol reported improved vaginal dryness, involuntary loss of urine, and pain during intercourse for both groups.106 Head-to-head trials comparing the intravaginal estradiol ring with CEE vaginal cream among women with signs and symptoms of urogenital atrophy indicated that the two agents were comparable for relief of vaginal dryness and pain during intercourse, resolution of atrophic signs, improvement in vaginal mucosal maturation indices, and reduction in vaginal pH.107, 108 Similar findings were reported in a trial of the estradiol vaginal tablet and CEE cream.109

Four placebo controlled trials of transdermal estradiol indicated improvements in some, but not all, urinary and sexual symptoms. Two trials reported improvements in responses to the McCoy Sex Scale Questionnaire compared to placebo,9, 110 although, another trial indicated no differences for symptoms of vaginal discomfort, loss of libido, and incontinence.111 Another trial of transdermal estradiol indicated improvement in vaginal dryness, but not pain during intercourse, frequent urination, dysuria, stress incontinence, and nocturia, compared to placebo.112

A trial of oral CEE reported significantly improved vaginal dryness and urinary frequency, but no significant improvement on six other items related to sexual function on a General Health Questionnaire compared to placebo.113 A trial comparing oral CEE and transdermal estradiol indicated that the majority of women reported either no change or improvements in vaginal dryness and itching, pain during intercourse, and urinary pain and burning in all treatment groups with no major differences between groups.114 All treatment groups demonstrated improved vaginal cytology, measured by the maturation index, with the biggest improvement in the higher dose estradiol group (0.1 mg/day).

Sleep disturbance. Improved measures of sleep were specifically reported in three trials included in the systematic review.15 Transdermal estradiol provided significant improvements in sleep quality, sleep onset, and decreased nocturnal restlessness and awakening compared to placebo.115 In a subanalysis of the Women's Health Initiative (WHI) that considered women ages 50 to 54 years old who reported moderate to severe vasomotor symptoms at baseline, there was a positive effect on sleep disturbance with CEE compared to placebo.116 In contrast, a trial of CEE in women with hot flashes and nighttime awakening at baseline indicated improvement in menopausal symptoms and measures of psychological well-being, but not in parameters of sleep quality such as total sleep time, sleep onset time, number of awakenings, and REM sleep duration compared to placebo.117

Mood symptoms. Thirteen trials of estrogen reporting mood and depression outcomes met eligibility criteria (Table 5, Appendix 6-2), including three trials rated good-quality,118–120 four fair,121–124 and six poor.125–130 Trials compared placebo with CEE,119, 121, 124, 126, 128 oral estradiol,120, 122, 125 transdermal estradiol,118, 126 piperazin oestrone sulphate,129, 130 and estropiate.127 Head-to-head comparisons included estradiol and estradiol valerate,123 CEE and clonidine,128 CEE and raloxifene,119 and estradiol and raloxifene.120

Table 5. Trials of estrogen for depression.

Table 5

Trials of estrogen for depression.

Two of the four good and fair-quality placebo controlled trials reporting between group differences found improved depressive symptoms among women using transdermal estradiol118 and oral estradiol with norethindrone acetate.122 No differences were found in trials comparing placebo with CEE/medroxyprogesterone acetate (MPA)121 or placebo with CEE.119 There were no differences in symptoms when comparing estradiol with estradiol valerate123 and estradiol with raloxifene.120 Another trial indicated improved anxiety with raloxifene vs. CEE, but no differences in depressed mood.119

Reduced quality of life. In a systematic review,15 four trials of transdermal estradiol,112, 131–133 two trials of oral estradiol,122, 134 and one trial of esterified estrogens135 indicated improved health related quality of life and well-being compared to placebo, as measured by various instruments. In a subanalysis of the WHI that considered women ages 50 to 54 years old who reported moderate to severe vasomotor symptoms at baseline, there was no effect on health-related quality of life measures on the RAND 36-item Health Survey, despite significant improvements in vasomotor symptoms.116 A head-to-head comparison of CEE vs. transdermal estradiol utilizing the Menopause Specific Quality of Life Questionnaire indicated improvement in all areas with no significant differences between groups.136


A total of 244 abstracts were identified and five randomized controlled trials met inclusion criteria (Table 6),137–141 including one rated good-quality,137 one fair,138 and three poor.139–141 Studies included transdermal progesterone,137, 138 intramuscular progesterone,139 medroxyprogesterone acetate (MPA) tablet,141 and MPA injection.140

Table 6. Trials of progestins.

Table 6

Trials of progestins.

Two trials of transdermal progesterone report conflicting results. A good-quality trial reported no differences between progesterone (32 mg/day) and placebo groups for vasomotor or somatic symptoms, mood, or sexual feelings as measured on previously validated instruments.137 A fair-quality trial of a lower dose (20 mg/day) reported significant reductions in vasomotor symptoms, but not depression scores, with progesterone compared to placebo (83 percent vs. 19 percent; p<0.001).138 A poor-quality trial of intramuscular progesterone described small differences between groups, however, the clinical significance of these are unclear.139 Two poor-quality trials of MPA did not provide adequate statistical comparisons.140, 141

None of the trials reported adverse effects.


Testosterone. A total of 1,095 abstracts were identified from a search for trials of androgens that included testosterone and dehydroepiandrosterone (DHEA). Of these, 10 trials of testosterone met inclusion criteria including one rated good-quality,142 four fair,144–146, 148 and five poor143, 147–149, 151 (Table 7, Appendix 6-3). Major limitations of studies include few subjects, undefined inclusion and exclusion criteria, use of unclear outcome measures, and using open label rather than double blind methodology. Two additional studies were reported in abstract form and are included only in the evidence table (Appendix 6-3).152, 153

Table 7. Trials of Androgens.

Table 7

Trials of Androgens.

Trials used methyltestosterone,142, 143, 145–149 testosterone undecanoate,150, 151 and transdermal testosterone.144 All trials combined testosterone with estrogen and compared it head-to-head with estrogen, and two studies compared it with placebo.144, 149 Outcomes included sexual dysfunction,142, 144–146, 150, 151 hot flashes,143–149 mood,144, 145, 147, 149, 150 sleep disturbances,143, 145, 147–149 vaginal dryness,143, 146–149 and quality of life.144, 145, 150

Only one good-quality and two fair-quality trials reported statistical comparisons between testosterone/estrogen and estrogen or placebo groups.142, 143, 144 One trial indicated no differences between testosterone/estrogen and estrogen alone for hot flash severity, vaginal dryness, or sleep problems.143 Sexual interest and responsiveness were improved with testosterone/estrogen compared to estrogen alone based on scores on the Sexual Interest Questionnaire in another trial.142 Sexual symptoms were also improved in a trial comparing testosterone/estrogen with placebo.144

Adverse effects were reported in all 10 trials. Acne142, 143, 146, 148 and hirsutism146, 148 occurred significantly more often among women using methyltestosterone/estrogen than with estrogen alone. Adverse effects were not always separated by treatment group, making it difficult to determine which drug was responsible for the reported effect.144, 149–151

Dehydroepiandrosterone (DHEA). One fair-quality154 and one poor-quality155 randomized controlled trial met inclusion criteria (Table 7, Appendix 6-4). Studies were small, lacked clear exclusion criteria,155 and had high losses to follow up.154

Trials used dehydroepiandrosterone (DHEA)154 and dehydroepiandrosterone sulfate (DHEAS)155 alone154, 155 or combined with transdermal estradiol.155 No significant differences between DHEA and placebo were determined on any measures of symptoms including hot flashes, vaginal dryness, insomnia, depression, urinary difficulties, libido, concentration, dizziness, forgetfulness, irritability, aches, anxiety, headaches, and fatigue in the one trial reporting between group differences.154

Adverse effects were reported for DHEA only and included paresthesia of an upper extremity.154


A total of 162 abstracts were reviewed and 20 randomized controlled trials met inclusion criteria (Table 8, Appendix 6-5) including three rated good-quality,156–158 four fair,159–162 and 13 poor.163–175 Four additional studies were reported in abstract form and are included only in the evidence table (Appendix 6-5).176–179

Table 8. Trials of tibolone.

Table 8

Trials of tibolone.

Trials compared tibolone with estrogen alone,169 estrogen combined with progestin or progesterone,158–160, 162, 163, 166, 170, 171, 173–175or placebo.156, 157, 161, 164, 165, 167, 168, 172 Outcomes included hot flashes, sexual dysfunction, mood, uterine bleeding, somatic complaints, vaginal dryness, sleep disturbance, cognitive effects, urinary complaints, and quality of life.

Three good or fair-quality trials reported between group differences for tibolone and placebo.156, 157, 161 Results indicated improved hot flashes,157 sleep,157 and somatic complaints156 with tibolone, but no differences between groups in Greene Climacteric Score,156 quality of life,157, 161 or mood, energy, pain, social isolation, or urinary symptoms.157

Four good or fair-quality trials reported between group differences for tibolone and estrogen alone or estrogen combined with progestin or progesterone.158–160 162 Two trials indicated improved hot flashes with estrogen compared to tibolone.159, 160 Sexual interest, drive, and/or performance were improved with tibolone compared to CEE/MPA in another trial.160 No differences between groups were determined for several other menopause related symptoms.158–160, 162

Uterine bleeding was increased with tibolone in four trials,157, 167, 168, 180 decreased in two,159, 167 and was not significantly changed in four others.158, 162, 163, 174 Uterine bleeding was dose related in two studies.168, 180 Adverse effects that occurred significantly more often among women using tibolone than placebo were reported in 16 of 20 trials and included body pain,156, 160, 163, 164, 166, 170, 174, 175 weight gain,156, 158, 160, 164, 166, 169, 175 and headache.156, 160, 163, 166, 167, 170 Four studies did not report adverse effects.162, 171–173

Antidepressant drugs

A total of 184 abstracts were reviewed and eight randomized controlled trials met inclusion criteria including one trial rated good-quality,181 three fair,182, 183, 188 and four poor184–187 (Table 9, Appendix 6-6). Trials used paroxetine,181 venlafaxine,183 moclobemide,188 and veralipride.182, 184–187 Most studies were limited by small sample sizes, short durations of follow-up, lack of follow-up data, inadequate descriptions of inclusion criteria, and, in one trial, lack of blinding.187

Table 9. Trials of antidepressant drugs.

Table 9

Trials of antidepressant drugs.

Selective serotonin reuptake inhibitors (SSRIs). A good-quality trial of paroxetine reported statistically significant reductions in mean hot flash frequency and hot flash composite score at two doses compared with placebo.181 No differences were detected in sleep, depression, anxiety, sexual interest, disability, or side effects.181 A fair-quality trial of venlafaxine reported improvement on a single item measuring how significantly hot flashes interfered with daily activities compared to placebo, but no differences in hot flash frequency or severity as measured on daily diaries.183 The venlafaxine group also had improved mood and vitality compared with placebo.183 Other trials of SSRIs enrolled women with breast cancer and are described below.

Moclobemide. A fair-quality trial evaluating two doses of moclobemide showed reduced hot flash composite scores with moclobemide at both doses, although comparisons with placebo were not reported.188 This trial was further limited by including only 30 participants and lasting only five weeks.

Veralipride. Veralipride, an antidopaminergic drug, was evaluated in five trials; one rated fair-quality182 and four poor-quality.184–187 Two poor-quality trials comparing veralipride to placebo reported reduced rates of hot flashes with veralipride.184, 185 Of the two studies comparing veralipride to estrogen, a poor-quality trial found that veralipride had less benefit,187 and a fair-quality trial found it to be equivalent.182

In four of the five studies, the side effects of mastodynia and/or galactorrhea occurred more frequently in the veralipride group.

Other Drugs

A total of 51 abstracts were identified and 15 randomized controlled trials met inclusion criteria (Appendix 6-7). Trials used clonidine,128, 189–197 gabapentin,198 methyldopa,199–201 and Bellergal Retard.202 Major limitations of trials include few subjects, lack of clear inclusion and exclusion criteria, high attrition or loss to follow up, no washout period in crossover trials, lack of data for pre-crossover comparisons, and short treatment duration.

Clonidine. Clonidine was compared to placebo in eight trials,189–195, 197 and compared to other active drugs in trials with lisuride, sodium valproate, transdermal estradiol,195 and CEE and medrogestone128, 196 (Table 10). Two trials were rated fair-quality,189, 192 and eight were poor.128, 190, 191, 193–197 Outcomes included hot flashes, insomnia, and mood symptoms.

Table 10. Trials of Clonidine.

Table 10

Trials of Clonidine.

Of the trials for which pre-crossover statistics were provided or could be calculated, there were no statistically significant differences between clonidine and placebo in the improvement of hot flash symptoms.189, 191, 192 In the non-crossover placebo controlled trials, one trial reported that clonidine was more effective than placebo in reducing hot flash frequency and intensity,195 and the other found no differences between groups in number of hot flashes, although women on clonidine were significantly more likely to report a decrease in subjective frequency, severity, and duration of hot flashes.197 Of the 3 remaining crossover studies reporting only summary statistics,190, 193, 194 one trial demonstrated that clonidine was more effective than placebo in reducing the number of hot flashes and improving subjective hot flash measures.193 Two head-to-head trials found that estrogen, not clonidine, effectively decreased the number of hot flashes and improved measures of depression and anxiety.128, 196 Two trials found no differences between clonidine and placebo in measures of psychological symptoms,191, 192 and one found no difference in frequency of insomnia.192

Adverse effects were reported in six of 10 trials; dry mouth occurred more frequently in women taking clonidine than placebo,189, 191, 193 and blood pressure was not affected by clonidine.189, 190, 192–194, 197

Methyldopa. Three crossover trials, two rated fair-quality200, 201 and one poor,199 compared methyldopa with placebo (Table 11). All trials provided pre-crossover statistics regarding improvement in hot flash frequency and none found significant differences between groups. A fair-quality study reporting post-crossover data separately found that methyldopa was more effective than placebo in reducing the number of hot flashes and improving Visual Analog Scores after crossover.201

Table 11. Trials of methyldopa, Bellergal, and gabapentin.

Table 11

Trials of methyldopa, Bellergal, and gabapentin.

All trials noted substantially more adverse effects with methyldopa than placebo. Fatigue or drowsiness, dizziness, and dry mouth occurred more often among women using methyldopa than placebo. One patient on methyldopa developed a lupus-like rash.200 No significant changes in blood pressure were noted,199, 201 however, one study reported orthostatic hypotension in one patient.200

Gabapentin. In a good-quality study comparing gabapentin to placebo, hot flash frequency and a composite menopause symptom score were significantly improved with gabapentin (Table 11).198 No significant differences in mood, quality of life, Patient Global Impression of Change scores, and sleep quality were detected.198

Gabapentin use reduced levels of albumin, total protein, total bilirubin, blood urea nitrogen, and platelets compared to placebo. Somnolence, dizziness, rash, and peripheral edema were reported for gabapentin but not placebo. Fifty percent of the gabapentin group and 28 percent of the placebo group reported at least one adverse event.

Bellergal. One poor-quality trial compared Bellergal Retard, a combination of 0.6 mg ergotamine, 40 mg phenobarbital, and 0.2 mg levorotatory alkaloids, with placebo (Table 11).20 There were no statistically significant differences between Bellergal and placebo in measures of vasomotor symptoms, insomnia, nervousness, hyperirritability, headache, paresthesia, loss of libido, and dizziness.

Adverse effects were similar between groups and included dry mouth, dizziness, and sleepiness.


A total of 195 abstracts were identified and 21 randomized controlled trials met inclusion criteria (Table 12, Appendix 6-8). Trials used dietary soy isoflavones,203–213 soy isoflavone extracts,214–218 other forms of phytoestrogens,219–223 and phytoestrogens combined with other agents.224 Eight additional studies were reported in abstract form and are included only in the evidence table (Appendix 6-8).225–232

Table 12. Trials of phytoestrogens and isoflavones.

Table 12

Trials of phytoestrogens and isoflavones.

Soy isoflavones—dietary. Dietary forms of soy isoflavones were compared to placebo in 10 trials (Table 12) that included soy powder,203, 204, 211, 213 cereal,205 drink,206, 212 diet,208, 211 flour,209 and tablets.207, 210 Of these, seven were rated fair-quality,203, 204, 206, 207, 209–212 and three poor.205, 208, 213 Outcomes included hot flashes, sleep, mood or depression, vaginal dryness, sexual symptoms, scores on symptoms scales, and single trials reporting palpitations, headaches, and general health.

Of the eight trials providing between group comparisons, one fair-quality and one poor-quality trial reported improved hot flashes with soy compared to placebo203, 204, 213 and six reported no differences.205, 206, 209–212 A fair-quality trial of soy powder reported improved hot flash frequency with soy compared to placebo (44 percent vs. 31 percent, p<0.01), but no differences between groups on the Kupperman Index.203, 204 A poor-quality trial of soy protein tablets indicated improved severity of hot flashes and improved hypoestrogenic symptoms score with soy, but no differences in number of hot flushes, night sweats, sleep disturbance, or general health score.213 In most trials, hot flashes improved in both soy and placebo groups,203, 204, 206, 209–212 although in two trials rated poor-quality, the placebo group reported improved symptoms while the soy group did not.205, 208

Soy isoflavones—extract. Soy isoflavone extracts were compared to placebo in five trials (Table 12), including four rated fair-quality,214–216, 218 and one poor.217 Outcomes included hot flashes, cognitive tests, scores on the Greene Climacteric Index, and mood.

Of the three trials providing between group comparisons and hot flash outcomes, all reported improved hot flash frequency216, 217 or severity218 with soy compared to placebo, and one trial reported no differences in frequency of night sweats.218 The Greene Climacteric Scale score was improved with soy compared to placebo in one trial,217 but not the other.215 Performance on cognitive tests, including tests of memory, improved with soy compared to placebo in the two trials evaluating this outcome.214, 215 Mood was not improved with soy in one trial.215

A trial reporting adverse effects of isoflavones indicated an increased rate of endometrial hyperplasia after five years of using 150 mg/day of isoflavone soy supplement.233

Other phytoestrogens. Phytoestrogens were compared to placebo in five trials (Table 12), including two rated fair-quality221, 223 and three poor.219, 220, 222 Trials included diet,219 topical cream,220, 222 and genistein tablets.221, 223 One trial included a comparison group using estradiol and noresthisterone.221 Outcomes included hot flashes in all trials, vaginal dryness, scores on menopausal symptom scales, mood, sexual symptoms, and energy level.

Three of five trials providing between group comparisons and hot flash outcomes reported improved hot flash severity219 or score221, 223 with phytoestrogen compared to placebo. Two trials reported no differences compared to placebo in frequency and severity of hot flashes222 or Kupperman Index score,223 and another indicated that the hot flash score was better with estradiol and noresthisterone than genistein.221 Vaginal dryness was improved with phytoestrogen diet in one trial,219 but there were no significant differences in overall symptom scores.219 A trial of wild yam cream found no differences in mood, libido, or energy level.222

Combinations. A poor-quality trial of soy based isoflavones combined with C. racemosa (black cohosh) reported improved hot flush symptoms with treatment compared to placebo, as reported on a questionnaire.224

Complementary and Alternative Medicine

A total of 1,237 abstracts were identified and 28 randomized controlled trials met inclusion criteria (Table 13, Appendix 6-8). Trials used acupuncture,234–236 Chinese herbs,237–242 red clover,243–248 black cohosh,249 combinations,250 other supplements,251–259 manual therapies,260 and energy therapies.261 Four trials were reported in abstract form and are included only in the evidence table (Appendix 6-8).227, 231, 262, 263

Table 13. Trials of complementary and alternative therapies.

Table 13

Trials of complementary and alternative therapies.

Acupuncture. Four small trials compared a series of acupuncture treatments specific to menopausal symptoms with alternate nonspecific acupuncture procedures (Table 13).234–236, 264 Of these, one trial was rated fair-quality264 and three poor.234–236 The fair-quality trial found no differences between groups in any measures,236 while a poor-quality study reported improved mood in the treatment group, but no differences in menopausal symptoms or well-being.235 A poor-quality trial compared acupuncture treatment specific to menopausal symptoms to both nonspecific acupuncture and estrogen and found a more pronounced improvement in hot flashes with estrogen than acupuncture.264

Chinese herbs. Three fair-quality trials240–242 and three poor-quality trials237–239 of Chinese herbs met criteria for this review. Five trials compared herbs against placebo238–242 and two against CEE/MPA237, 242 (Table 13). Treatments included specific formulations of herbs,237, 238 ginseng,239, 241 dong quai,240 and an herbal remedy (pueraria lobata) combined with isoflavone.242 Outcomes included hot flashes and other vasomotor symptoms, mood and depression, well-being and quality of life, memory, somatic symptoms, sexual symptoms, and general health. Studies were small, had short durations, and included highly selected populations.

A fair-quality trial comparing ginseng with placebo reported significant improvements for depression, well-being, and health scores on the Psychological General Well Being Index, but no differences in hot flashes or scores on other instruments.241 Results of two other fair-quality placebo controlled trials indicated no differences between groups for menopausal symptoms,240, 242 well-being,240, 242 or rate of learning.242

Red clover. Red clover isoflavone tablets (Promensil and Rimostil) were compared against placebo in six trials (Table 13) including one rated good-quality,247 four fair,243–245, 248 and one poor.246 Outcomes included hot flashes and other vasomotor symptoms including measures from the Greene Climacteric Scale. One of five trials reporting between group comparisons indicated improved hot flashes with treatment compared to placebo,248 and no trials reported differences in scores on the Greene Climacteric Scale or symptom diary.243, 244, 246–248

A good-quality meta-analysis of 25 trials of phytoestrogens for treatment of menopausal symptoms included an analysis of red clover using five of the six studies in this review.244–248 Results of the meta-analysis indicated no significant differences in hot flash frequency between treatment and placebo groups (weighted mean difference -0.60; 95 percent CI, -1.71 to 0.51).18

Black cohosh. One fair-quality trial compared C. racemosa preparation (black cohosh) with estrogen and placebo (Table 13).249 Between group comparisons indicated improved hot flushes with estrogen vs. placebo only.

Combinations. A small poor-quality trial of a botanical formula (burdock root, licorice root, motherwort, dong quai, and wild yam) compared with placebo indicated no differences between groups in the number and severity of hot flashes (Table 13).250 Results indicated a significant decrease in the total number of other types of symptoms after three months of therapy for the botanical formula group compared to placebo (p<0.03).250

Other supplements. Nine trials evaluated effects of other forms of supplements including melatonin,251 vitamin E,252 kavakava,253 evening primrose oil with vitamin E,254 guar gum,255 vaginal moisturizer,256, 257 phospohlipid liposome injections,258 and S-adenosyl-L-methionine259 (Table 13). Of these, two trials were rated fair-quality253, 258 and seven poor.251, 252, 254–257, 259 Outcomes included hot flashes, mood or depression, vaginal dryness, and scores on symptom scales.

Between group differences for placebo controlled trials indicated improved symptoms with treatment compared to placebo for mood or depression (melatonin, S-adenosyl-L-methionine),251, 259 anxiety (kavakava, phospholipids liposomes),253, 258 and scores on symptom scales (phospholipid liposomes).258 No differences between groups were found for hot flashes (guar gum),255 mood or depression (kavakava),253 and scores on symptom scales (kavakava).253 In a poor-quality trial of evening primrose oil, hot flashes were significantly improved with placebo compared to treatment.254 A poor-quality trial comparing vaginal moisturizer against dienoestrol vaginal cream reported improved vaginal dryness with dienoestrol.256

Manual therapies. A fair-quality trial of low force osteopathic manipulation of the pelvis, spine, and cranium was compared against sham low force touch in similar areas (Table 13).260 After 10 weeks of therapy, results indicated improved hot flashes and nights sweats, urinary frequency, depression, and insomnia with treatment compared with placebo.260

Energy therapies. A poor-quality trial of women undergoing nine sessions of reflexology provided over 19 weeks resulted in no differences in severity of hot flashes and night sweats or measures of health and well-being compared to women receiving standard foot massages (Table 13).261

Behavioral Interventions

A total of 436 abstracts were identified and eight randomized controlled trials met inclusion criteria (Table 14, Appendix 6-8). Trials consisted of exercise,265–267 relaxation, 268–270 low frequency sound wave audiotape,271 and education interventions.272 Of these, three trials were rated fair-quality,265, 267, 272 and five poor.266, 268–271 One additional trial was reported in abstract form and is included only in the evidence table (Appendix 6-8).226

Table 14. Trials of behavioral interventions.

Table 14

Trials of behavioral interventions.

Exercise. Two fair-quality studies evaluated the effects of aerobic exercise using comparison groups that underwent stretching activities265 or received usual care267 (Table 14). Outcomes included hot flashes and other vasomotor symptoms, mood and depression, cognitive function, sleep, well-being, and quality of life. Results of one trial indicated no differences in outcomes in the exercise group compared to the stretching group.265 A subset of women categorized as recently entering menopause showed improved memory in the aerobic vs. stretching group. 265 In a trial of exercise compared to usual care, exercisers showed significantly improved quality of life on the Nottingham Health Profile.267

Other Interventions. A fair-quality trial compared health education with usual care and found no differences between groups for measures of mood, health, vaginal dryness, or sexual relationships.272 Knowledge of menopause increased in the intervention group, while the control group was more likely to attribute symptoms to menopause.272 Other trials rated poor-quality due to limitations in methods and/or small numbers of subjects evaluated biofeedback, relaxation, paced respiration, and use of audiotapes (Table 14).


Results of trials of therapies providing between group comparisons are summarized in Table 15. Although benefits and adverse effects of therapies were equally important in this review, most trials did not report adverse effects or reported them incompletely. Studies of estrogen provided the most information. It is unclear how applicable the adverse effects reported from other larger estrogen trials, such as the Women's Health Initiative (WHI), are for younger, symptomatic women taking estrogen for brief periods.104 For women in the WHI who reported having menopausal symptoms, strokes and coronary heat disease events were not significantly elevated.273, 274 Other important adverse outcomes, such as breast cancer, thrombosis, and gall bladder disease, have not been reported by age and symptom status in this trial.

Table 15. Summary of benefits of therapies.

Table 15

Summary of benefits of therapies.

Main findings from trials of therapies include:

  • Estrogen, in either opposed or unopposed regimens, is the most consistently effective therapy for vasomotor symptoms, and demonstrates benefit in most trials evaluating urogenital symptoms. Some, but not all, trials evaluating sleep, mood and depression, sexual function, and quality of life outcomes also report benefit with estrogen compared to placebo.
  • Breast tenderness and uterine bleeding are the most commonly reported adverse outcomes in estrogen trials; others include nausea and vomiting, headache, weight change, dizziness, venous thromboembolic events, cardiovascular events, rash and pruritus, cholecystitis, and liver effects.
  • Trials of progestin indicate mixed results for treatment of vasomotor symptoms.
  • Few trials of testosterone are available; one trial indicated no differences between testosterone/estrogen and estrogen alone for hot flash severity, vaginal dryness, or sleep problems. Sexual symptoms were improved with testosterone/estrogen compared to estrogen alone or placebo in two other trials.
  • For women using testosterone combined with estrogen, acne and hirsutism occur significantly more often than for women using estrogen alone.
  • Based on only a few fair or good-quality trials, tibolone demonstrated benefit for vasomotor symptoms, sleep, and somatic complaints compared to placebo, and was similar to estrogen for some, but not all, symptoms.
  • Uterine bleeding, body pain, weight gain, and headache were more common in tibolone vs. placebo groups.
  • Several agents demonstrate benefits in managing vasomotor symptoms in some, but not all trials, or in only a few available trials, including paroxetine, veralipride, gabapentin, soy isoflavones, and other phytoestrogens.
  • Trials of soy isoflavones and other complementary and alternative medicine therapies report benefits in improving nonvasomotor symptoms, although results vary widely, methods are lacking, and studies are typically small and not generalizable.
  • Placebo effects in trials are large reflecting underlying fluctuations of symptoms.

Key Question 4. Therapies for Women with Specific Characteristics

What are the important considerations in managing menopause-related symptoms in women with clinical characteristics or circumstances that may complicate decision-making?

Bilateral Oophorectomy

A large number of studies reported data on women with bilateral oophorectomies, but did not stratify results by this characteristic.142, 144–148, 150, 163, 165, 169, 173, 181, 188, 192, 193, 195, 197, 217, 242, 256 Trials of estrogen that exclusively enrolled women with oophorectomies to take unopposed estrogen reported similar improvements in vasomotor symptoms as trials of women without oophorectomies taking opposed estrogen.16

Premature Ovarian Failure

Only one trial of soy isoflavones considered premature ovarian failure, but results were not reported specifically for women with this condition.217

Breast Cancer

A total of 200 abstracts were identified and 15 randomized controlled trials met inclusion criteria (Table 16, Appendix 6-9).275–289 Of these, 3 trials were rated good-quality,278, 280, 285 10 fair,275–277, 279, 281–284, 287, 288 and 2 poor.286, 289 Major limitations of trials include small size, lack of sufficient detail regarding recruitment and randomization, non blinding, and lack of intention-to-treat analyses. Two studies recruited women who either had a history of breast cancer or a perceived increased risk of breast cancer,282, 283 and all other trials included women who had prior diagnoses of breast cancer but no residual disease and who were no longer receiving chemotherapy or radiation. Most studies included women taking tamoxifen, although tamoxifen-users represented varying proportions of the study populations (31 percent to 100 percent). Results for tamoxifen-users were presented separately in only one trial,276 and in three trials, all women were taking tamoxifen.284–286 Study designs did not allow ascertainment of whether treatments were addressing a side effect of tamoxifen therapy284–286 or vasomotor symptoms that occur naturally in women with a history of breast cancer. All studies recruited women from breast cancer clinics.

Table 16. Trials in women with breast cancer.

Table 16

Trials in women with breast cancer.

Trials used conventional and complementary/alternative medical treatments including venlafaxine,282 fluoxetine,283 clonidine,284, 285 megestrol acetate,280 various preparations of soy/isoflavone products,277–279, 288 black cohosh,276, 286 magnets,289 vitamin E,275 and a polycarbonphil-based vaginal moisturizer.281 All compared treatments with placebo or usual care. Outcomes included hot flashes, sleep disturbances, mood, somatic complaints, sexual dysfunction, vaginal dryness, quality of life, and overall menopausal symptoms. No trials evaluated cognitive symptoms, urinary symptoms, or uterine bleeding.

Clonidine, venlafaxine, and megestrol acetate were associated with significantly improved measures of hot flashes.280, 282, 284, 285 A good-quality trial of oral clonidine indicated a 38 percent decrease in hot flashes for clonidine (8 to 5 hot flashes/day) vs. 24 percent for placebo (7.4 to 5.7 hot flashes/day).285 In a fair-quality trial of transdermal clonidine, hot flashes decreased 44 percent with clonidine (6.1 to 3.4 hot flashes/day) compared to 27 percent for placebo (7 to 5.1 hot flashes/day).284 Megestrol acetate reduced hot flashes by 74 percent compared to 27 percent for placebo in a good-quality trial.280 A fair-quality trial of venlafaxine reported a mean decrease in hot flashes of 30 percent to 58 percent for varying doses of venlafaxine vs. 19 percent for placebo.282 Vitamin E, black cohosh, isoflavones, magnets, and fluoxetine alone did not reduce hot flashes.

Results for other outcomes were mixed. Sleep was improved in a fair-quality trial evaluating fluoxetine,283 but not in a fair-quality trial using black cohosh.276 None of the four fair-quality trials assessing mood (fluoxetine, venlafaxine, phytoestrogen and black cohosh) found them to be effective.276, 279, 282, 283 Of the trials evaluating somatic complaints (headaches, palpitations, excessive sweating, nausea, fatigue), the only benefit reported was from a single fair-quality trial using black cohosh showing improvement of excessive sweating.276 Behavioral counseling with tailored therapies was effective at improving sexual functioning in one trial,287 but a fair-quality trial of venlafaxine found no difference in libido.282 The one trial evaluating polycarbophil-based vaginal preparation found it no more effective than placebo in reducing vaginal dryness, but effective at improving dyspareunia scores.281 Of the four trials evaluating quality of life, clonidine improved measures,285while fluoxetine, magnets, and behavioral counseling did not.282, 287, 289 Three trials assessed overall menopausal symptoms, of which only behavioral counseling was effective.287

Adverse effects were reported in 12 of 15 trials.275–285, 289 Gastrointestinal adverse effects occurred more often among women using phytoestrogen products,278, 279 particularly a soy beverage preparation.278 Two other trials reported no gastrointestinal adverse effects with soy preparations,277, 288 although one trial excluded all women with intolerance to soy prior to randomization. 288 Fluoxetine was associated with worse appetite, nausea, constipation and dry mouth than placebo.283 The clonidine patch was associated with dry mouth, constipation, itchiness under the patch, and drowsiness.284 Clonidine tablets were associated with a trend toward more difficulty sleeping than placebo.285 Participants in the trial of magnets had difficulty with itching, redness, and perspiration at the site where magnets were affixed.289 Polycarbophil-based vaginal preparation was associated with an undesirable “wetness sensation.”281

Concurrent Use of Selective Estrogen Receptor Modulators (SERMs) and Other Agents

Patients using tamoxifen were enrolled in most of the trials of therapies in women with breast cancer.275–279, 282–287 One fair-quality study of black cohosh reported hot flash symptoms separately for tamoxifen users vs. non-users, but found no significant differences between groups.276 Other studies of treatment of menopausal symptoms did not describe how concurrent use of SERMs affects therapy.

Lifestyle and Behavioral Factors

Few trials considered lifestyle and behavioral factors such as smoking, alcohol use, diet, or others,161, 162, 169, 170 and no trials reported results according to these factors.

Recent Discontinuation of Menopausal Hormone Therapy

Most trials required discontinuation of hormone therapy prior to enrollment, however, no results were specifically reported for women with recent discontinuation.

Very Low or Very High Body Mass Index (BMI)

Some trials specified a range of acceptable BMIs in their eligibility criteria thereby disallowing enrollment of women with very high or low BMI. No studies reported results according to BMI.


  • Evidence is not available to determine if the effectiveness of therapy or adverse effects differ for women with bilateral oophorectomy, premature ovarian failure, concurrent use of SERMs or other potentially interacting agents, lifestyle and behavioral factors, recent discontinuation of menopausal hormone therapy, or very low or very high BMI.
  • For women with breast cancer, results of 15 randomized controlled trials indicate that clonidine, venlafaxine, and megestrol acetate are associated with significantly improved measures of hot flashes, and vitamin E, black cohosh, isoflavones, magnets, and fluoxetine are not. Result for nonvasomotor outcomes are mixed.

Key Question 5. Future Research on Therapies

What are the future research directions for treatment of menopause-related symptoms and conditions?

Although many trials of several types of interventions have been published, results are consistent and conclusive only for estrogen agents for the treatment of vasomotor and urogenital symptoms. Despite this evidence, many questions remain about the benefits and adverse effects of estrogen for treating menopause related symptoms long-term. For nonestrogen therapies, larger, more rigorous, and more comprehensive trials are needed in order to determine benefits and adverse effects. Funding for nonindustry sponsored trials would enable trials of non drug therapies, including head-to-head trials comparing several types of drug and non drug interventions. Future research should include:

  • Determination of optimally effective doses, combination regimens, durations of use, and timing of therapy.
  • Evaluation of approaches to identify optimal candidates for specific therapies (e.g., identification of thrombophilias).
  • Head-to-head and placebo comparisons of estrogen alone and combined with other types of therapies including non drug interventions.
  • Trials demonstrating how to discontinue estrogen when symptoms subside, including the effectiveness of tapering doses and/or replacing with other therapies including non drug interventions.
  • Better reporting of adverse effects in trials and use of standardized categories of adverse effects so data can be combined across trials.
  • Improved analysis of results including analysis by hysterectomy and oophorectomy status, stage of menopause, age, concurrent conditions and medications, and other factors.
  • More comprehensive trials to determine the role of regular exercise, sleep management, optimal nutrition, healthy relationships, social support, and relaxation; effects of mind-body techniques such as biofeedback and breathing; effects of a whole system approach with Chinese medicine.
  • Additional, well-designed and controlled trials of phytoestrogens, botanicals, and bio-identical hormones, especially estriol, estradiol, and progesterone. Further study of antidepressants for vasomotor symptoms would be justified based on evidence of currently available trials.
  • Enrollment of women with specific characteristics who have not previously been evaluated such as nonwhite women, women with premature ovarian failure, those using SERMs and other agents influencing symptoms concurrently, women with very high or low BMI, and those with lifestyle and behavioral factors influencing symptoms. Trials should report data specific to these groups in order to interpret their influence on therapy.
  • Use of standard definitions, measures, and outcomes so results can be compared across trials.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...