Participants

A total of 1273 women were randomised in the 28 included studies. The total study population was 1207 postmenopausal women and 66 perimenopausal women. All studies but one (Barnhart 1999) included postmenopausal women only. One study included 100 menopausal women after surgery only (Gupta 2013). Overall, the age ranged from 36 to 80 years. Women from different ethnicities were included, but not all studies reported on this. The most reported group was Caucasian women. A total of 33% of the studies included healthy menopausal women (for example without serious comorbidity, not using medication). Two studies only included women with specific comorbidities: Sjögren syndrome (Forsblad‐d'Elia 2009) and fibromyalgia (Finckh 2005). Two studies included menopausal women with complaints of vaginal dryness, irritation or itching, or dyspareunia (Labrie 2008; Labrie 2009a). Two studies included women with hypoactive sexual desire disorder or low libido (Bloch 2013; Panjari 2009). One study included women with an increased cardiovascular risk (Silvestri 2005). The remaining studies did not report health status or comorbidities of the women who were included.

Study design

Nineteen parallel‐designed RCTs and nine crossover‐designed trials (Finckh 2005; Forsblad‐d'Elia 2009; Hirshman 2003; Hirshman 2004; Labrie 1997; Merritt 2012; Mortola 1990; Silvestri 2005; Stangl 2010) were included in the review. Fifteen studies were included for (quantitative) meta‐analysis.

Interventions

The following comparisons were made in the included studies: DHEA versus placebo, head‐to‐head (DHEA versus other active treatment) comparisons, and DHEA versus no treatment. There were a total of:

• 23 included trials comparing DHEA versus placebo (Amin 2005; Barnhart 1999; Bloch 2013; Casson 1998; Dayal 2005; El‐Alfy 2010; Finckh 2005; Forsblad‐d'Elia 2009; Gomez‐Santos 2011; Hirshman 2003; Hirshman 2004; Kratz 2000; Kritz‐Silverstein 2008; Labrie 1997; Labrie 2008; Labrie 2009a; Lasco 2001; Merritt 2012; Mortola 1990; Nouveau 2008; Panjari 2009; Stanczyk 2009; Stangl 2010);

• six included trials comparing DHEA versus HT (Dayal 2005; Genazzani 2011; Gupta 2013; Pluchino 2008; Silvestri 2005; Stomati 1999). HT included estrogen therapy (ET); conjugated equine estrogen (Dayal 2005; Gupta 2013); transdermal estradiol alone (Stomati 1999); transdermal estradiol plus oral progesterone (ET + P) (Pluchino 2008) or oral estradiol plus dihydrogesterone (ET + P) (Genazzani 2011);

• three trials compared DHEA versus tibolone (a synthetic steroid hormone with estrogen and androgens as metabolites) (Genazzani 2011; Gupta 2013; Silvestri 2005);

• one trial compared DHEA versus no treatment (Gupta 2013).

Outcomes

Primary outcome: effectiveness (quality of life, general wellbeing)

Fourteen studies reported quality of life or wellbeing. The results from 11 of these trials could be included in a meta‐analysis (Barnhart 1999; Bloch 2013; Dayal 2005; Finckh 2005; Hirshman 2003; Hirshman 2004; Kritz‐Silverstein 2008; Labrie 1997; Labrie 2009a; Merritt 2012; Panjari 2009; Stangl 2010); nine trials reported end scores for quality of life scales after treatment (Bloch 2013; Dayal 2005; Finckh 2005; Hirshman 2003; Hirshman 2004; Kritz‐Silverstein 2008; Merritt 2012; Panjari 2009; Stangl 2010) whereas two studies reported change scores (difference between baseline and end scores) (Barnhart 1999; Labrie 2009a). Three trials (Amin 2005; Kratz 2000; Labrie 1997) investigated quality of life or wellbeing but required additional data for inclusion in the meta‐analysis.

Different questionnaires were used to assess quality of life including: Beck Depression Inventory (BDI) (Hirshman 2003; Hirshman 2004; Kritz‐Silverstein 2008; Merritt 2012; Stangl 2010), Symptom Checklist‐90 (SCL‐90) (Merritt 2012; Stangl 2010), Women's Health Questionnaire (WHQ) (Dayal 2005), Psychological General Well‐Being Index (PGWBI) (Finckh 2005; Labrie 2009a; Panjari 2009), Hamilton Depression Rating Score (HAM‐D) (Barnhart 1999), SmithKline Beecham Quality of Life Self Report Questionnaire (SKQOL) (Barnhart 1999), Life Satisfaction Inventory‐Z (LSI‐Z) (Kritz‐Silverstein 2008), Short Form Health Survey (SF‐36) (Kritz‐Silverstein 2008), Satisfaction With Life Scale (SWLS) (Kritz‐Silverstein 2008), Mental Health Inventory (MHI) (Bloch 2013) and Menopause‐specific Quality of Life Questionnaire (MENQOL) (Labrie 2009a; Panjari 2009). If studies used multiple questionnaires to assess quality of life we included scores from questionnaires that were validated and used in most publications (BDI, PGWBI and SCL‐90). Where these questionnaires were not used, we included the scales we judged as most representative for quality of life.

Primary outcome: adverse effects

A total of 23 studies reported on side effects. Five trials could be included for meta‐analysis; four trials that used oral administration (Bloch 2013; Finckh 2005; Lasco 2001; Panjari 2009) and one trial that used intravaginal administration (Nouveau 2008). Other studies could not be included for meta‐analysis for the following reasons: 10 studies reported that there were no side effects (Casson 1998; Genazzani 2011; Gomez‐Santos 2011; Hirshman 2003; Hirshman 2004; Pluchino 2008; Silvestri 2005; Stanczyk 2009; Stomati 1999); one study reported there were no significant side effects (Labrie 2008), and for one study the author confirmed that there were no significant side effects found (Labrie 2009a); two studies only reported adverse effects for women discontinuing the study (Barnhart 1999; El‐Alfy 2010), it was unclear whether there were any (non‐significant) side effects reported for these studies; one study, which had male and female participants, did not separate side effects by sex and could therefore not be included (Kritz‐Silverstein 2008); only one trial reported adverse effects for DHEA versus HT and tibolone (Gupta 2013); results from most crossover trials (Forsblad‐d'Elia 2009; Labrie 1997; Mortola 1990) could not be included due to lack of data (no combined statistical tests were done for adverse effects). Attempts were made to contact all authors to request additional data.

Adverse effects that were reported for DHEA treatment but could not be included for meta‐analysis due to lack of information, as described above, were as follows:
acne (Labrie 1997), hirsutism (Forsblad‐d'Elia 2009; Labrie 1997; Mortola 1990), contact dermatitis (El‐Alfy 2010), parasthesia and numbness of upper extremity (Barnhart 1999), depressiveness (Forsblad‐d'Elia 2009), calf cramps (Forsblad‐d'Elia 2009), increase in nightly dreams (Forsblad‐d'Elia 2009) and dizziness (Forsblad‐d'Elia 2009).

Secondary outcome: menopausal symptoms

Five studies investigated the effect of the intervention on menopausal symptoms (Barnhart 1999; Dayal 2005; Genazzani 2011; Gupta 2013; Stomati 1999). Five studies reported continuous outcomes using scales to assess menopausal symptoms, three studies reported end scores after treatment and could be combined for meta‐analysis. Scales used were: Green Climacteric Scale (Genazzani 2011), Kupperman Index (Stomati 1999) and an element of the Woman's Health Questionnaire (Dayal 2005). Two studies could not be included for meta‐analysis because one study reported change scores (Barnhart 1999) and the other study reported dichotomous outcomes (Gupta 2013).

Secondary outcome: sexual function

A total of eight studies investigated sexual function as an outcome (Bloch 2013; Dayal 2005; Finckh 2005; Genazzani 2011; Kritz‐Silverstein 2008; Labrie 2009a; Mortola 1990; Panjari 2009). Six studies reported end scores after treatment whereas one study reported change scores (Barnhart 1999) and could therefore not be included in this meta‐analysis. Different parameters were used to assess sexual function including: the McCoy Female Sexuality Score (MFSQ), Female Sexual Function Index (FSFI), Derogatis Interview for Sexual Functioning (DISF), Sabbatsberg Sexual Self‐rating Scoring (SSS) and Abbreviated Sexual Function Questionnaire (ASFQ). The most used questionnaires were the MFSQ and FSFI. One study stated that no changes in sexual drive had been reported but did not use a tool to quantify sexual drive (Mortola 1990). Most studies utilised questionnaires with a total score to represent sexual function as a whole as opposed to splitting up the various elements such as libido, dyspareunia etc. Where studies did provide scores for the separate elements of sexual function, libido was chosen as the function to represent sexual function. There were not enough data from the studies to undertake meaningful subanalyses of the various elements of sexual function.

Sequence generation

Of all 28 included trials, 19 were randomised with a parallel design and 9 were randomised with a crossover design. Seventeen trials described their methods of sequence generation, which were usually computer‐generated, randomised block schemes or random number tables (Barnhart 1999; Bloch 2013; Casson 1998; Dayal 2005; Finckh 2005; Forsblad‐d'Elia 2009; Genazzani 2011; Hirshman 2003; Hirshman 2004; Lasco 2001; Merritt 2012; Nouveau 2008; Panjari 2009; Pluchino 2008; Silvestri 2005; Stanczyk 2009; Stangl 2010). Eleven trials simply reported the trial as randomised without a description of the randomisation method (Amin 2005; El‐Alfy 2010; Gomez‐Santos 2011; Gupta 2013; Kratz 2000; Kritz‐Silverstein 2008; Labrie 1997; Labrie 2008; Labrie 2009a; Mortola 1990; Stomati 1999).

Allocation concealment

Nine trials described their allocation concealment as a sequentially numbered list kept by the pharmacy, a non‐investigator or using sealed envelopes (Barnhart 1999; Bloch 2013; Dayal 2005; Finckh 2005; Forsblad‐d'Elia 2009; Lasco 2001; Panjari 2009; Silvestri 2005; Stanczyk 2009). Nineteen trials did not describe allocation concealment. Attempts were made to contact these authors to clarify allocation concealment techniques, unfortunately we were unable to retrieve the information needed.

Blinding of participants and personnel

Fourteen studies (Barnhart 1999; Bloch 2013; Dayal 2005; Finckh 2005; Forsblad‐d'Elia 2009; Genazzani 2011; Hirshman 2003; Hirshman 2004; Lasco 2001; Merritt 2012; Nouveau 2008; Panjari 2009; Stanczyk 2009; Stangl 2010) described blinding methods that we graded as low risk. Blinding in these trials was done by using treatments looking identical in appearance (for example identical looking DHEA and placebo capsules or creams). In some trials personnel were described as blinded as well, but the majority of the studies did not report on blinding personnel. Five studies (Genazzani 2011; Gupta 2013; Labrie 1997; Pluchino 2008; Stomati 1999) were graded as high risk for bias as they were not blinded. These trials compared different active treatments that were applied differently (for example transdermal patches versus oral capsules) or had different appearances. Nine studies (Amin 2005; El‐Alfy 2010; Gomez‐Santos 2011; Kratz 2000; Kritz‐Silverstein 2008; Labrie 2008; Labrie 2009a; Mortola 1990; Silvestri 2005) did not describe their blinding process and were therefore graded as at unclear risk of bias.

Blinding of outcome assessments

If blinding was unlikely to have been broken for participants it was unlikely that the outcomes were influenced as the outcomes were generally done by self‐assessment, for all outcomes. If a trial was not blinded it was likely that this may have influenced the outcomes. Therefore trials that did not use blinding were graded as high risk of bias for blinding. We judged the remaining trials as unclear as it was unclear whether blinding was done or was likely to have been broken.

1.1 Quality of life (QoL): DHEA versus control (placebo or no treatment) (end scores)

DHEA did not significantly improve quality of life or wellbeing end scores compared to placebo (SMD ‐0.04, 95% CI ‐0.20 to 0.13, P = 0.67, 9 studies, 402 women (247 from parallel and 155 from crossover trials), I² = 67%) (Analysis 1.1). There were no studies that compared DHEA to no treatment for quality of life or wellbeing.

1.2 Quality of life (QoL): DHEA versus control (placebo or no treatment) (end scores) sensitivity analysis

As the I² value was greater than 50%, we repeated the analysis but this time we restricted eligibility to studies without a high risk of bias (SMD 0.16, 95% CI ‐0.03 to 0.34, P = 0.10, 8 studies, 287 women (132 from parallel and 155 from crossover trials), I² = 0%, moderate quality evidence) (Analysis 1.2; Figure 5). No large differences in treatment effect were found after conducting the sensitivity analysis. The heterogeneity observed in 1.1 was found to be caused by one study (Kritz‐Silverstein 2008). This study showed imbalanced baseline scores and did not report information on the randomisation technique, allocation, blinding and attrition and was therefore graded as high risk of bias. We judged the end scores of this study to be unreliable as they did not represent the effect of the treatment due to the imbalanced baseline scores for quality of life. Therefore, we chose to present the result of the sensitivity analysis as the main result.

Figure 5

---

---

Forest plot of comparison: 1 DHEA versus control (placebo or no treatment), outcome: 1.2 Sensitivity analysis QoL and wellbeing (end scores).

The findings from two studies that could not be included for meta‐analysis due to lack of data (no scores available) agreed with the results above (Kratz 2000; Labrie 1997). One other study that lacked data for inclusion and only had an abstract available (Amin 2005) reported a slight decrease of quality of life in the placebo group compared to DHEA treatment.

1.3 and 1.4 Quality of life (QoL): DHEA versus control ‐ subgrouped based on low libido or hypoactive sexual desire disorder (HSDD) and treatment duration

See 'overall subgroups'.

1.5 Quality of life (QoL): DHEA versus control (placebo or no treatment) (change scores)

Two studies reported change scores (difference between baseline and end scores) (Barnhart 1999; Labrie 2009a). These could not be pooled due to a high level of heterogeneity (I² = 94%). Sensitivity analysis did not change this. Heterogeneity may be explained by route of administration. One study administered DHEA orally (Labrie 2009a) whereas the other study used intravaginal application (see Analysis 2.1).

1.6 Adverse effects: DHEA versus control (placebo or no treatment)

Nine studies reported there were no side effects found at all (Casson 1998; Genazzani 2011; Gomez‐Santos 2011; Hirshman 2003; Hirshman 2004; Pluchino 2008; Silvestri 2005; Stanczyk 2009; Stomati 1999). For one study comparing DHEA with placebo (Labrie 2008) the author confirmed there were no significant side effects found, but it was unclear whether there were any minor adverse effects. There were no studies that compared DHEA to no treatment.

The results for androgenic side effects could only be pooled for acne and androgenic side effects (Bloch 2013; Finckh 2005; Lasco 2001; Nouveau 2008; Panjari 2009). The results showed evidence of an association with androgenic side effects (mainly acne) in 15% of the women on DHEA versus < 3% of women on placebo (OR 3.77, 95% CI 1.36 to 10.40, P = 0.01, 5 studies, 376 women, I² = 10%, moderate quality evidence). One study (Panjari 2009) reported on hirsutism as an androgenic side effect but was included for acne already. Therefore, this study could not be pooled for hirsutism as well due to risk of double counting participants.

1.6.2. Acne: DHEA versus control (placebo or no treatment)

There were no significant differences found in the frequency of acne between DHEA and placebo (OR 2.26, 95% CI 0.56 to 9.02, P = 0.25, 4 studies, 158 women, I² = 0%, moderate quality evidence) (Analysis 1.6; Figure 6). Two studies (Forsblad‐d'Elia 2009; Labrie 1997) reported on acne as a side effect of DHEA treatment. For one study it was unclear whether this adverse effect was investigated for placebo as well (Labrie 1997) and the other study was a crossover trial which had no data available for paired tests on adverse effects (Forsblad‐d'Elia 2009).

Figure 6

---

---

Forest plot of comparison: 1 DHEA versus control (placebo or no treatment), outcome: 1.6 Side effects.

1.7 Adverse effects: DHEA versus control ‐ acne subgrouped based on study duration

See 'subgroups overall'.

1.8 Menopausal symptoms: DHEA versus control (placebo or no treatment) (continuous)

Insufficient data were available to conduct a meta‐analysis for DHEA versus placebo (Analysis 1.8). Two studies reported menopausal symptoms as a continuous outcome, one study (including 17 postmenopausal women) reported end scores (Dayal 2005) whereas the other study (including 66 perimenopausal women) reported change scores (Barnhart 1999). Both studies showed no decrease in menopausal symptoms for DHEA.

1.9 Menopausal symptoms: DHEA versus control (placebo or no treatment) (dichotomous)

Only one study (Gupta 2013) compared DHEA versus no treatment in 50 menopausal women after surgery, therefore a meta‐analysis could not be conducted. This trial showed a decrease in menopausal symptoms, for tiredness, night sweats and loss of libido. Other menopausal symptoms reported (hot flushes, insomnia, depression, vaginal dryness and pruritis vulvae) did not differ for the DHEA treatment and no treatment groups. This study was not blinded and this may have influenced the outcome.

1.10 Sexual function: DHEA versus control (placebo or no treatment)

DHEA may have improved sexual function (SMD 0.19, 95% CI ‐0.01 to 0.40, P = 0.06, 6 studies, 376 women (354 women from parallel trials and 22 women from crossover trials), I² = 10%) (Figure 7). However, one trial including 115 postmenopausal women (Kritz‐Silverstein 2008) showed imbalanced baseline scores and did not report information on the randomisation technique, allocation, blinding and attrition and was therefore graded as at high risk of bias. Therefore, we chose to conduct a sensitivity analysis.

Figure 7

---

---

Forest plot of comparison: 1 DHEA versus control (placebo or no treatment), outcome: 1.11 Sexual function (end scores).

1.11 Sexual function: DHEA versus control (placebo or no treatment) sensitivity analysis

After conducting a post hoc sensitivity analysis excluding one study with a high risk of bias, DHEA was shown to be significantly more effective on sexual function compared to placebo (SMD 0.31, 95% CI 0.07 to 0.55, P = 0.01, 5 studies, 261 women (239 women from parallel trials and 22 women from crossover trials), I² = 0%, moderate quality of evidence) (Figure 8). We judged the end scores of the excluded study to be unreliable as it did not represent the effect of the treatment found in this study due to imbalanced baseline scores for sexual function. Therefore, we choose to present the results of the sensitivity analysis as the main results. The reported improvement in sexual function was minimal although significant and may not be clinically significant as, for example, an SMD of 0.31 would translate to a change of only 2.3 points on the Sabbatsberg‐Sexual Self rating (SSS) scale, which ranges from 0 to 84. No studies investigated DHEA versus no treatment.

Figure 8

---

---

Forest plot of comparison: 1 DHEA versus control (placebo or no treatment), outcome: 1.11 Sexual function (end scores) (sensitivity analysis).

2.1 Quality of life (QoL): oral DHEA versus control subgrouped by route of administration (e.g. intravaginal, transdermal patches)

There were insufficient data available to conduct a meta‐analysis for effectiveness based on route of administration (Analysis 2.1). Only one trial (including 107 postmenopausal women) that investigated a different route of administration than oral administration (intravaginal application) reported QoL change scores (Labrie 2009a); and only one trial studied oral DHEA and reported change scores for QoL (Barnhart 1999). The I² value was greater than 50% (94%) therefore these studies could not be pooled for meta‐analysis. Intravaginal administration showed a beneficial effect for DHEA on QoL, whereas the other study did not show any differences in treatment effect.

2.2 Adverse effects: oral DHEA versus control subgrouped by route of administration (e.g. intravaginal DHEA, transdermal patches)

No evidence was found that acne was associated more with oral application of DHEA (OR 2.16, 95% CI 0.47 to 9.96, P = 0,90, 3 studies, 136 women, I² = 5%, very low quality of evidence) than with skin application of DHEA (OR 2.74, 95% CI 0.10 to 74.87, P = 0,90, 1 study, 22 women, very low quality evidence), with the test for subgroup differences showing Chi² = 0.02 (df = 1, P = 0.90, I² = 0%) (Analysis 2.2). The studies that investigated intravaginal administration did not report on acne whereas acne (or other skin effects) was reported in all trials investigating skin application (El‐Alfy 2010; Labrie 1997; Nouveau 2008) and acne was reported in five more trials investigating oral administration (Bloch 2013; Finckh 2005; Forsblad‐d'Elia 2009; Lasco 2001; Panjari 2009). Data for different side effects could not be matched for an overall analysis because it was unclear if these side effects were reported in the same or different participants. This information was needed to create a proper analysis.

2.3 Menopausal symptoms: oral DHEA versus control subgrouped by route of administration (e.g. intravaginal DHEA, transdermal patches)

There were no trials using other routes of administration that investigated menopausal symptoms as an outcome.

2.4 Sexual function: oral DHEA versus control subgrouped by route of administration (e.g. intravaginal DHEA, transdermal patches)

No significant differences were found between studies looking at oral administration (SMD 0.11, 95% CI ‐0.13 to 0.35, P = 0.36, 5 studies, 340 women, I² = 0) and a study using intravaginal DHEA (SMD 0.42, 95% CI 0.03 to 0.81, 1 study, 218 women) with the test for subgroup differences showing Chi² = 1.77 (df = 1, P = 0.18, I² = 43.4%) (Analysis 2.4). One trial including 115 postmenopausal women (Kritz‐Silverstein 2008) reported imbalanced baseline scores and was therefore graded as high risk of bias. Therefore, we repeated the analysis but this time without studies with a high risk of bias. The results from the sensitivity analysis also revealed no evidence for differences between oral administration and intravaginal application of DHEA on sexual function (SMD 0.31, 95% CI 0.07 to 0.55, P = 0.88, 5 studies, 261 women (of which 107 received intravaginal DHEA or placebo capsules, 239 women were from parallel trials and 22 women from crossover trials), I² = 0%).

3. DHEA versus hormone therapy (HT) (estrogen and progesterone in combination and separately, androgen therapy or tibolone)

3.1 Quality of life (QoL): DHEA versus hormone therapy (HT) (estrogen and progesterone in combination and separately, androgen therapy or tibolone)

Only one study reported results of DHEA versus estrogen (Dayal 2005). Therefore a meta‐analysis could not be conducted. This study found no significant difference between treatment effects of DHEA compared to HT on QoL.

There were no trials that compared DHEA to a different form of HT for QoL or wellbeing.

3.2 Adverse effects: DHEA versus hormone therapy (HT) (dichotomous)

See 3.4.

3.3 Menopausal symptoms: DHEA versus hormone therapy (HT)

It was unclear whether DHEA differed from HT in effects on menopausal symptoms. Two studies were included for the meta‐analysis of estrogen therapy (ET) (Genazzani 2011; Stomati 1999). The I² value was greater than 50% (74%) therefore these studies could not be pooled for meta‐analysis (Analysis 3.3). One study showed a small beneficial effect on menopausal symptoms for HT (Stomati 1999) whereas the other showed a more beneficial effect for DHEA (Genazzani 2011).

One study that investigated dichotomous outcomes (type and frequency of menopausal symptoms) and could therefore not be included in the meta‐analysis showed no significant differences in frequency of various menopausal symptoms between treatments (Gupta 2013). One trial showed no difference in frequency of menopausal symptoms between treatments (Gupta 2013). One study (Genazzani 2011) including 24 postmenopausal women investigated continuous outcomes and showed a slightly beneficial effect for DHEA.

3.4 Sexual function: DHEA versus HT

DHEA did not significantly improve sexual function compared to HT. Two trials (Dayal 2005; Genazzani 2011) compared the effects of DHEA and HT on sexual function. There was no significant difference found between these treatments in these two trials (MD 1.26, 95% CI ‐0.21 to 2.73, P = 0.09, 2 studies, 41 women, I² = 0%) (Analysis 3.4).

Only one trial investigated DHEA versus tibolone (Genazzani 2011). This trial included 24 postmenopausal women and reported a beneficial effect for DHEA compared to tibolone for improvement of sexual function.

4. Adverse effects: DHEA versus HT (dichotomous)

Only one study (Gupta 2013) that included menopausal women after surgery (n = 25 received DHEA, n = 25 received conjugated equine estrogen (CEE) as ET, n = 25 received tibolone) investigated this outcome and therefore a meta‐analysis could not be conducted. This study showed an increase in reported acne and hair loss with DHEA compared to ET.

The same study (Gupta 2013) investigated DHEA versus tibolone, but other studies had investigated this comparison for adverse effects only and a meta‐analysis could not be conducted. This study showed little evidence on adverse effects associated with any of the treatments. The study reported more acne and hair loss with DHEA compared to tibolone treatment.

5. DHEA versus any other medical treatment (e.g. antidepressants or clonidine)

No trials were found for this comparison on any of following outcomes: QoL or wellbeing, adverse effects, menopausal symptoms or sexual function.

6. DHEA versus any other non‐medical treatment (e.g. non‐medical therapies such as acupuncture or complementary therapies)

No trials were found for this comparison on any of following outcomes: QoL or wellbeing, adverse effects, menopausal symptoms or sexual function.