Hormone therapy for sexual function in perimenopausal and postmenopausal women

Abstract

Background

The perimenopausal and postmenopausal periods are associated with many symptoms, including sexual complaints. This review is an update of a review first published in 2013.

Objectives

We aimed to assess the effect of hormone therapy on sexual function in perimenopausal and postmenopausal women.

Search methods

On 19 December 2022 we searched the Gynaecology and Fertility Group Specialised Register, CENTRAL, MEDLINE, Embase, PsycINFO, CINAHL, LILACS, ISI Web of Science, two trials registries, and OpenGrey, together with reference checking and contact with experts in the field for any additional studies.

Selection criteria

We included randomized controlled trials that compared hormone therapy to either placebo or no intervention (control) using any validated assessment tool to evaluate sexual function. We considered hormone therapy: estrogen alone; estrogen in combination with progestogens; synthetic steroids, for example, tibolone; selective estrogen receptor modulators (SERMs), for example, raloxifene, bazedoxifene; and SERMs in combination with estrogen.

Data collection and analysis

We used standard methodological procedures recommended by Cochrane. We analyzed data using mean differences (MDs) and standardized mean differences (SMDs). The primary outcome was the sexual function score. Secondary outcomes were the domains of sexual response: desire; arousal; lubrication; orgasm; satisfaction; and pain. We assessed the certainty of the evidence using the GRADE approach.

Main results

We included 36 studies (23,299 women; 12,225 intervention group; 11,074 control group), of which 35 evaluated postmenopausal women; only one study evaluated perimenopausal women. The 'symptomatic or early postmenopausal women' subgroup included 10 studies, which included women experiencing menopausal symptoms (symptoms such as hot flushes, night sweats, sleep disturbance, vaginal atrophy, and dyspareunia) or early postmenopausal women (within five years after menopause). The 'unselected postmenopausal women’ subgroup included 26 studies, which included women regardless of menopausal symptoms and women whose last menstrual period was more than five years earlier. No study included only women with sexual dysfunction and only seven studies evaluated sexual function as a primary outcome. We deemed 20 studies at high risk of bias, two studies at low risk, and the other 14 studies at unclear risk of bias. Nineteen studies received commercial funding.

Estrogen alone versus control probably slightly improves the sexual function composite score in symptomatic or early postmenopausal women (SMD 0.50, 95% confidence interval (CI) (0.04 to 0.96; I² = 88%; 3 studies, 699 women; moderate‐quality evidence), and probably makes little or no difference to the sexual function composite score in unselected postmenopausal women (SMD 0.64, 95% CI −0.12 to 1.41; I² = 94%; 6 studies, 608 women; moderate‐quality evidence). The pooled result suggests that estrogen alone versus placebo or no intervention probably slightly improves sexual function composite score (SMD 0.60, 95% CI 0.16 to 1.04; I² = 92%; 9 studies, 1307 women, moderate‐quality evidence).

We are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on the sexual function composite score in unselected postmenopausal women (MD 0.08 95% CI −1.52 to 1.68; 1 study, 104 women; very low‐quality evidence).

We are uncertain of the effect of synthetic steroids versus control on the sexual function composite score in symptomatic or early postmenopausal women (SMD 1.32, 95% CI 1.18 to 1.47; 1 study, 883 women; very low‐quality evidence) and of their effect in unselected postmenopausal women (SMD 0.46, 95% CI 0.07 to 0.85; 1 study, 105 women; very low‐quality evidence).

We are uncertain of the effect of SERMs versus control on the sexual function composite score in symptomatic or early postmenopausal women (MD −1.00, 95% CI −2.00 to ‐0.00; 1 study, 215 women; very low‐quality evidence) and of their effect in unselected postmenopausal women (MD 2.24, 95% 1.37 to 3.11 2 studies, 1525 women, I² = 1%, low‐quality evidence).

We are uncertain of the effect of SERMs combined with estrogen versus control on the sexual function composite score in symptomatic or early postmenopausal women (SMD 0.22, 95% CI 0.00 to 0.43; 1 study, 542 women; very low‐quality evidence) and of their effect in unselected postmenopausal women (SMD 2.79, 95% CI 2.41 to 3.18; 1 study, 272 women; very low‐quality evidence).

The observed heterogeneity in many analyses may be caused by variations in the interventions and doses used, and by different tools used for assessment.

Authors' conclusions

Hormone therapy treatment with estrogen alone probably slightly improves the sexual function composite score in women with menopausal symptoms or in early postmenopause (within five years of amenorrhoea), and in unselected postmenopausal women, especially in the lubrication, pain, and satisfaction domains. We are uncertain whether estrogen combined with progestogens improves the sexual function composite score in unselected postmenopausal women. Evidence regarding other hormone therapies (synthetic steroids and SERMs) is of very low quality and we are uncertain of their effect on sexual function. The current evidence does not suggest the beneficial effects of synthetic steroids (for example tibolone) or SERMs alone or combined with estrogen on sexual function. More studies that evaluate the effect of estrogen combined with progestogens, synthetic steroids, SERMs, and SERMs combined with estrogen would improve the quality of the evidence for the effect of these treatments on sexual function in perimenopausal and postmenopausal women.

Plain language summary

Does hormone therapy improve sexual function in women going through or after menopause?

Key messages

• Estrogen (the hormone associated with sexual and reproductive development in women) alone probably improves sexual function scores compared to placebo.

• We are unsure of the effect of estrogen plus progestogens (another female hormone), synthetic steroids (such as tibolone), selective estrogen receptor modulators (that affect how estrogen works) or selective estrogen receptor modulators plus estrogen on sexual function compared to placebo or no treatment.

• Different hormone treatments and doses, and questionnaires used for assessment, may have caused the variation seen in results.

What is menopause and its effects on women?

Menopause is when women's periods stop, usually at around the age of 45 to 55 years. During menopause the ovaries gradually stop producing estrogen, the hormone that regulates periods. The reduction in estrogen can cause unwanted symptoms before periods stop (perimenopausal), during menopause and after menopause (postmenopausal). Symptoms include mood changes, hot flushes, and night sweats. Sexual complaints such as painful intercourse, lack of interest in sex, and problems related to arousal or orgasm are common after menopause and can affect women's self‐esteem, self‐confidence, and sexual health.

What is hormone therapy, and how might it help women’s sexual function?

Hormone therapy consists of various hormones or combinations of hormones that can help reduce menopausal symptoms. It can be given as skin patches, sprays or gels, tablets, or implants, and is used to treat a wide range of perimenopausal and postmenopausal symptoms. Hormone therapy might improve symptoms affecting sexual function such as dryness, itching, and painful intercourse by increasing lubrication, blood flow, and sensation in vaginal tissues.

What did we want to find out?

We wanted to find out if hormone therapy improves women's sexual functioning and whether effects are different in different stages of menopause. We were also interested in which types of hormone therapy were most effective: estrogen alone, estrogen combined with other hormones, or synthetic (manufactured) steroids or hormones.

What did we do?

We searched for studies that looked at the different types of hormone therapy compared to placebo (a dummy drug) or no treatment and its effect on sexual function in perimenopausal or postmenopausal women. We searched for studies that used estrogen alone; estrogen in combination with progestogens; synthetic steroids, (such as tibolone); selective estrogen receptor modulators (SERMs, that affect how estrogen works by blocking or activating different parts of the body, such as raloxifene or bazedoxifene); and selective estrogen receptor modulators combined with estrogen.

We were most interested in the effect of hormone therapy on the global sexual function score, which measures the effect of hormone therapy on all the areas of sexual function combined: desire, arousal, lubrication, orgasm, satisfaction and pain. We also wanted to know the effect of hormone therapy on the individual areas that make up the global sexual function score. Scores had to be evaluated using a recognised and validated questionnaire.

We divided women by length of time since their last period:

• within 5 years of their last period with or without menopausal symptoms; and

• more than 5 years since their last menstrual period, regardless of menopausal symptoms.

We compared and summarized the results of the studies and assessed our confidence in the evidence based on factors such as study methods and sizes.

What did we find?

We found 36 studies including 23,299 women. All but one study involved women after menopause; the other included women during menopause. Some but not all women had bothersome symptoms such as hot flushes, night sweats, and vaginal dryness.

• For women within 5 years of their last period, treatment with estrogen alone probably slightly improves sexual function based on the sexual function composite score compared to placebo.

• For women whose last period was more than 5 years earlier, estrogen alone probably makes little or no difference to sexual function based on sexual function scores compared to a placebo.

• For both groups of women, we are unsure of the effect of estrogen plus progestogens, synthetic steroids, selective estrogen receptor modulators alone, or selective estrogen receptor modulators plus estrogen on sexual function compared to placebo or no treatment.

What are the limitations of the evidence?

Our confidence in the evidence is moderate to very low due to variation in the results. This was probably because studies used different drugs and doses, and different questionnaires for assessment.

How up‐to‐date is the evidence?

The evidence is current to December 2022.

Summary of findings

Summary of findings 1. Hormonal therapy compared with placebo or no intervention for sexual function in symptomatic or early postmenopausal women.

Hormonal therapy compared with placebo or no intervention for sexual function in symptomatic or early postmenopausal women
Patient or population: symptomatic or early postmenopausal womena Settings: outpatient health centres in different regions of the world Interventionsb: estrogen alone, estrogen + progestogens, synthetic steroids (tibolone), SERMs, SERMs + estrogen Comparison: placebo or no intervention
Outcomes	Ilustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk (placebo or no intervention)	Corresponding risk (hormonal therapy)
Estrogen alone Sexual function score measured by MENQoL sexual subscore (higher score = better sexual functioning); GCS sexual subscore (higher score = worse sexual functioning and for analysis, the scoring direction was reversed); and MSQ (higher score = better sexual functioning)	The mean score ranged from 0.71 to 46	SMD 0.50, 95% higher (0.04 to 0.96 higher)	Not estimable	699 (3 studies)	⊕⊕⊕⊝ Moderatec	Probably slightly improves sexual function
Estrogen + progestogens	‐	‐	‐	‐	‐	Not reported by any study in this comparison
Synthetic steroids Sexual function score measured by WHQ sexual subscore (higher score = better sexual functioning)	The mean score was 0.16	SMD 1.32, 95% higher (1.18 to 1.47 higher)	Not estimable	883 (1 study)	⊕⊝⊝⊝ Verylowd,e	We are uncertain whether synthetic steroids improve sexual function
SERMs Composite sexual function score measured by AESEX scale (higher score = worse sexual functioning and for analysis, the scoring direction was reversed)	The mean score was 0.34	MD −1.00, 95% lower (−2.00 lower to −0.00)	Not estimable	215 (1 study)	⊕⊝⊝⊝ Very lowc,d	We are uncertain whether SERMs improve sexual function
SERMs + estrogen Composite sexual function score measured by ASEX scale (higher score = worse sexual functioning and for analysis, the scoring direction was reversed)	The mean score was 2.0	SMD 0.22, 95% higher (0.00 to 0.43 higher)	Not estimable	542 (1 study)	⊕⊝⊝⊝ Very lowc,d	We are uncertain whether SERMs + estrogen improve sexual function
*The basis for the assumed risk is the mean score. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group ‐ expressed as change in standardized mean difference. ASEX: Arizona Sexual Experiences scale; CI: confidence interval; GCS: Greene Climacteric Scale, sexual subscore (symptom 21); MD: mean difference; MENQoL: Menopause‐specific Quality of Life Questionnaire, sexual subscore; MSQ: McCoy Sex Scale Questionnaire; SERMs: selective estrogen receptor modulators; SMD: standardized mean difference; WHQ: Women's Health Questionnaire, sexual subscore
GRADE Working Group grades of evidence High quality: further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: we are very uncertain about the estimate.

^aSymptomatic or early postmenopausal women: this subgroup included women with any menopause‐related symptoms (not only sexual dysfunction) or postmenopausal women less than 5 years from the last menstrual period. In this group, only one study included perimenopausal women, defined as women who had their last menstrual period within 12 months of inclusion.

^bInterventions

• Oral hormone therapy: tibolone 2.5 mg/day; bazedoxifene 20 mg/day; bazedoxifene 20 mg conjugated estrogen 0.45 mg /day; bazedoxifene 20 mg, conjugated estrogen 0.625 mg /day.
• Transdermal hormone therapy: estradiol 0.014 mg/d, patch delivering micro‐dose; estradiol 50 ug, twice a week.
• Local hormone therapy: estradiol 50 ug/day, vaginal route.

^cDowngraded by one level for serious risk of bias: most domains of risk of bias were assessed as 'unclear'.
^dDowngraded by two levels for serious imprecision or inconsistency: wide confidence interval and only one study in this analysis, small sample size, and low event rate.
^eDowngraded by two levels for very serious risk of bias; inadequate randomization, allocation concealment, and/or no evidence of blinding or other sources of bias.

Summary of findings 2. Hormonal therapy compared with placebo or no intervention for sexual function in unselected postmenopausal women.

Hormonal therapy compared with placebo or no intervention for sexual function in unselected postmenopausalwomen
Population: unselected postmenopausal womena Settings: outpatient health centres in different regions of the world Interventionsb: estrogen alone, estrogen + progestogens, synthetic steroids (tibolone), SERMs, SERMs + estrogen Comparison: placebo or no intervention
Outcomes	Ilustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk (placebo or no intervention)	Corresponding risk (hormonal therapy)
Estrogen alone Sexual function score measured by FSFI (higher score = better sexual functioning); WHQ sexual subscore (higher score = better sexual functioning); MENQoL sexual subscore (higher score = better sexual functioning)	The mean score ranged from 2.3 to 25.14	SMD 0.64, 95% higher (−0.12 lower to 1.41 higher)	Not estimable	608 (6 studies)	⊕⊕⊕⊝ Moderatec	Probably little or no difference in sexual function
Estrogen + progestogens Sexual function score measured by WHQ sexual subscore (higher score = better sexual functioning)	The mean score was 12.72	MD 0.08, 95% higher (−1.52 to 1.68 higher)	Not estimable	104 (1 study)	⊕⊝⊝⊝ Very lowc,d	We are uncertain whether estrogen + progestogens improve sexual function
Synthetic steroids Sexual function score measured by FSFI; WHQ sexual subscore (higher score = better sexual functioning)	The mean score ranged from 0.16 to 14.4	SMD 0.46 higher (0.07 to 0.85 higher)	Not estimable	105 (1 study)	⊕⊝⊝⊝ Very lowc,d	We are uncertain whether synthetic steroids improve sexual function
SERMs Sexual function score measured by FSFI (higher score = better sexual functioning)	The mean score ranged from 5.7 to 6.69	MD 2.24, 95% higher (1.3715 to 3.11 higher)	Not estimable	1525 (2 studies)	⊕⊕⊝⊝ Lowd	May slightly improve sexual function
SERMs + estrogen Sexual function score measured by MENQoL sexual subscore (higher score = better sexual functioning)	The mean score was 1.35	SMD 2.79, 95% higher (2.41 to 3.18 higher)	Not estimable	272 (1 study)	⊕⊝⊝⊝ Very lowc,d	We are uncertain whether SERMs + estrogen improve sexual function
*The basis for the assumed risk is the mean score ranged. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group ‐ standardized mean difference. CI: confidence interval; FSFI: Female Sexual Function Index; MD: mean difference; MENQoL: Menopause‐specific Quality of Life Questionnaire, sexual subscore; SERMs: selective estrogen receptor modulators; SMD: standardized mean difference; WHQ: Women's Health Questionnaire, sexual subscore
GRADE Working Group grades of evidence High quality: further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate or other sources of bias. Very low quality: we are very uncertain about the estimate.

^aUnselected postmenopausal women: this subgroup included women regardless of menopausal symptoms and allowed inclusion of women whose final menstrual period was more than five years earlier.
^bInterventions

• Oral hormone therapy: equine estrogen 0.625 mg/day; estradiol valerate 2 mg/day; tibolone 2.5 mg/day; ospemifene 60 mg/day; bazedoxifene 20 mg conjugated estrogen 0.45 mg/day; bazedoxifene 20 mg conjugated estrogen 0.625 mg/day.
• Local hormone therapy: conjugated estrogen 0.625 mg/day, vaginal route; estradiol ovule 0.03 mg/day, vaginal route; estradiol tablet 10 mg/day, vaginal route; estradiol 10.3 tablet ug/day, vaginal route; estradiol 0.625 mg/day.

^cDowngraded by one level for serious risk of bias: most domains of risk of bias were assessed as 'unclear'.
^dDowngraded by two levels for serious imprecision and inconsistency: wide confidence interval and only one study in this analysis, small sample size, and low event rate.
^eDowngraded by two levels for very serious risk of bias; inadequate randomization, allocation concealment, and/or no evidence of blinding or other sources of bias.

Background

Sexual function is a very important aspect of women’s health. In theory, hormone therapy may benefit aspects of a woman's sexual health. Medical counselling about hormone therapy should balance the benefits and risks, and take into account the aggregate value of hormone therapy (NAMS 2017; Santen 2010). Thus, information regarding the influence of hormone therapy on sexual function is important when considering the use of these drugs (Wierman 2010). This review was developed to help healthcare professionals make decisions about the use of hormone therapy to alleviate symptoms related to sexual dysfunction during perimenopause and postmenopause.

Description of the condition

Menopause is defined as occurring after 12 consecutive months of spontaneous amenorrhoea with no obvious pathologic cause (NAMS 2017); induced menopause is defined as the permanent cessation of menstruation after either surgical removal of the ovaries or iatrogenic ablation of ovarian function, for example, chemotherapy or pelvic radiation therapy (NAMS 2017). With the onset of menopause, women can experience problems related to hormonal changes (Avis 2015; Nappi 2009a; Whiteley 2013), such as vasomotor symptoms (hot flushes and night sweats; 75%), sleep disruption (40% to 60%), mood changes, cognitive decline, headaches (12%), vaginal dryness (83%), pain during intercourse (42%), soreness (27%), itching (26%), and burning (14%) (Monteleone 2018). Another important problem is the increase in sexual disorders, which increase from 27.2% in women aged 18 to 44 years to 44.6% in women between 45 and 64 years (Simon 2018). Hypoestrogenism associated with menopause promotes changes in the vaginal wall such as vaginal epithelium thinning and reduction in vessel density in the lamina propria layer (da Silva Lara 2009a; da Silva Lara 2009b), leading to poor lubrication and dyspareunia (Worsley 2017).

The most frequent sexual complaint in women after menopause is hypoactive sexual desire disorder (HSDD) (McCabe 2016a; Worsley 2017), defined as a persistent absence of sexual drive, sexual fantasies, and reduced desire for sexual activities for at least six months, followed by interpersonal difficulties or personal distress, or both (Basson 2000). Hormone therapy is a generic term that refers to any hormone therapy used in postmenopausal women for the alleviation of menopause‐related symptoms including vasomotor symptoms and genitourinary syndrome of menopause. Hormone therapy also has a positive effect on bone health (Flores 2021), and has also been studied in various settings. Clinical benefit has been observed when hormone therapy is used to relieve hot flushes and night sweats (Formoso 2012; MacLennan 2004), and for the relief of vaginal atrophy symptoms (Lethaby 2016). There is evidence that women using estrogen therapy are more likely to report sexual activity compared to controls using a placebo (Gass 2019). The most commonly used types of hormone therapy are estrogen alone, estrogen in combination with progestogens, synthetic steroids (for example tibolone), and selective estrogen receptor modulators (SERMs, for example, raloxifene). Other substances, such as androgens and phytoestrogen, might be used to relieve some menopausal‐related symptoms; however, we will not evaluate the effect of these drugs in this review because they are not widely accepted as hormone therapy.

Description of the intervention

Hormone therapy is a generic term that refers to any type of hormone therapy used during menopause for the alleviation of menopause‐related symptoms (Santen 2010). Clinical benefit has been observed when hormone therapy is used for the relief of hot flushes and night sweats (Formoso 2012; MacLennan 2004), and for the relief of vaginal atrophy symptoms (Lethaby 2016). Hormone therapy may be indicated to alleviate vasomotor symptoms and dyspareunia (Lethaby 2016; NAMS 2017). A meta‐analysis showed that hormone therapy treatment with estrogen alone or in combination with progestogens was associated with a small to moderate improvement in sexual problems, particularly in pain, when used in women with menopausal symptoms or even in asymptomatic women in early postmenopause (within five years of amenorrhoea), but not in postmenopausal women (Nastri 2013). The most commonly used types of hormone therapy are estrogen alone, estrogen in combination with progestogen, synthetic steroids such as tibolone, and selective estrogen receptor modulators (SERMs), for example, raloxifene. We examined the effect of these types of hormone therapy on sexual function in perimenopausal and postmenopausal women.

How the intervention might work

Sexual desire seems to be regulated by hormones and neurotransmitters of excitatory pathways such as dopamine, norepinephrine, melanocortin, and oxytocin (Clayton 2010). In animal models, estrogen facilitates dopamine release, and dopamine appears to enhance sexual desire, subjective arousal, and the interest to engage in sex following sexual stimulus (Hull 1999). This mechanism may be compromised by hypoestrogenism after menopause. Women with hot flushes and night sweats may find that their subsequent poor sleep patterns lead to tiredness and loss of energy. In addition, vaginal symptoms may affect women’s self‐esteem, self‐confidence, and sexual health (Portman 2014b). For perimenopausal and postmenopausal women, hormone therapy might improve hot flushes and night sweats as well as the symptoms of vulvovaginal atrophy increasing lubrication, blood flow, and sensation in vaginal tissues (NAMS 2017), and consequently, improve sexual functioning by directly impacting the dimensions of desire, arousal, lubrication, orgasm, satisfaction and sexual pain described in the literature (Basson 2000).

Why it is important to do this review

The World Health Organization (WHO) classifies sexual function as a cornerstone of quality of life irrespective of ageing (Few 1997). In fact, a cohort study with middle‐aged women in the USA showed that almost half of them considered sexual activity to be of moderate to extreme importance for a good quality of life, and thus, doctors should address this issue mainly in gynaecological outpatient care (Cain 2003). Despite its importance, female sexual function is not easy to investigate because it depends on several factors such as health, well‐being, cultural habits, socioeconomic status, relationship issues, and the presence of a partner, who may also have some health issues (Davis 2009). Female sexual dysfunction might be evaluated in different domains: sexual desire; arousal; orgasm; satisfaction; and pain (Basson 2000, McCabe 2016b). Although sexual function declines throughout the menopause transition (NAMS 2017; Rosen 2012), it is unclear whether this is caused by the lowering of estrogen levels, ageing, or both (da Silva Lara 2009a; Nappi 2009a).

The true effect of hormone therapy on sexual function is difficult to determine based on the literature published so far (McCabe 2016b); partially because of the different drugs and doses, different tools to evaluate the sexual function, and the particular population studied in each study. All these aspects highlight the importance of carrying out this review, which seeks to provide new evidence regarding the prescription of hormone therapy for the relief of sexual disorders during menopause. This review includes meta‐analysis data obtained using Review Manager software (Review Manager 2020), and includes the latest evidence on the effect of hormone therapy (HT) on sexual function in perimenopausal and postmenopausal women.

Objectives

We aimed to assess the effect of hormone therapy on sexual function in perimenopausal and postmenopausal women.

Methods

Criteria for considering studies for this review

Types of studies

We included published and unpublished randomized controlled trials (RCTs). We excluded non‐randomized studies (for example studies with evidence of inadequate sequence generation such as alternate days, and patient numbers) as they are associated with a high risk of bias. We included cross‐over studies but we planned to include in the meta‐analysis only data from the first phase, as the cross‐over is not a valid design in this context. The withdrawal of hormone therapy may cause symptoms, being a potential source of bias.

Types of participants

We defined perimenopausal women as women with menopausal symptoms who had their last menstrual period within the 12 months before inclusion in the study. We defined postmenopausal women as women with induced menopause (bilateral oophorectomy, as a result of chemotherapy or pelvic radiation therapy) or natural menopause (12 consecutive months of spontaneous amenorrhoea with no obvious pathologic cause or with at least six months of amenorrhoea and elevated follicle‐stimulating hormone (FSH)). We grouped the studies by participant characteristics with regard to menopausal symptoms. The symptomatic (women who experienced vasomotor or other menopausal symptoms) or early menopausal (up to five years postmenopause) group, and the unselected postmenopausal group, which included women regardless of menopausal symptoms and allowed inclusion of women whose last menstrual period was more than five years earlier.

Types of interventions

Prescription of hormone therapy compared with the prescription of placebo or no intervention. We considered hormone therapy, estrogen alone; estrogen in combination with progestogens; synthetic steroids (for example tibolone); SERMs (for example raloxifene, bazedoxifene); and SERMs in combination with estrogen. We did not include other drugs possibly used in the relief of menopausal symptoms in this review. In particular, we did not review studies on testosterone as this is the topic of another Cochrane Review (Somboonporn 2005). We only included studies in which the interval between the onset of the intervention and the assessment of outcomes was one month or more. In studies with multiple assessments, we considered only assessments between three and six months after starting the intervention.

Types of outcome measures

Primary outcomes

Sexual function

The primary outcome was the evaluation of the sexual function composite score, measured by validated tools.

1. Sexual function composite score

Secondary outcomes

Domains of sexual function

The secondary outcomes were the evaluation of the multidimensional domains of sexual function: desire, arousal, lubrication, orgasm, satisfaction, and pain (Basson 2000; McCabe 2016b), scored by validated tools.

1. Desire

2. Arousal

3. Lubrication

4. Orgasm

5. Satisfaction

6. Pain

We did not include adverse events related to hormone therapy as outcomes because we only included studies that reported sexual function; therefore, evidence for these topics would be a subset of the overall body of evidence, which could not be representative of the complete current evidence. In order to help weigh the benefits and harms of the treatments, we took into account the findings of other systematic reviews that addressed adverse events (see Discussion).

Search methods for identification of studies

We searched for RCTs following a search strategy developed in consultation with the Cochrane Gynaecology and Fertility Information Specialist. There was no language restriction in these searches.

Electronic searches

We searched for RCTs using the following electronic databases:

1. Cochrane Gynaecology and Fertility Specialised Register, ProCite platform, searched 19 December 2022 (Appendix 1);

2. Cochrane Central Register of Controlled Trials (CENTRAL; 2022, Issue 12) via the Cochrane Register of Studies Online (CRSO), Web platform, searched 19 December 2022 (Appendix 2);

3. MEDLINE, Ovid platform, searched from 1946 to 19 December 2022 (Appendix 3), this search was combined with the Cochrane highly sensitive search strategy for identifying randomized trials (Lefebvre 2022);

4. Embase, Ovid platform, searched from 1980 to 19 December 2022 (Appendix 4), this search was combined with trial filters developed by Scottish Intercollegiate Guidelines Network (SIGN) www.sign.ac.uk/what-we-do/methodology/search-filters/);

5. PsycINFO, Ovid platform, searched form 1806 to 19 December 2022 (Appendix 5);

6. CINAHL (Cumulative Index to Nursing and Allied Health Literature), Ebsco platform, searched from 1961 to 19 December 2022 (Appendix 6);

7. LILACS (Literatura Latino‐Americana e do Caribe em Ciências da Saúde), Web platform, searched 19 December 2022 (Appendix 7);

8. We searched for study protocols and ongoing trials using the following websites: ClinicalTrials.gov (Appendix 8); ISRCTN registry (Appendix 9); and the World Health Organization International Clinical Trials Registry Platform search portal (Appendix 10), Web platforms, searched 21 May 2018, more recent trial registrations were found in the CENTRAL 2022 output;

9. We searched for conference abstracts in the Web of Science, Web platform, and searched 19 December 2022 (Appendix 11).

Searching other resources

We handsearched the reference list of relevant articles retrieved by the search and contacted experts in the field for any additional studies. We also handsearched relevant journals and conference abstracts.

Data collection and analysis

Selection of studies

In this update, two review authors (MFT and JBPF) independently performed an initial screen of the titles and abstracts retrieved by the search; disagreements were resolved by discussion with another review author (LASL). We retrieved the full texts of all potentially eligible studies. Two review authors (MFT and DCR) independently examined these full‐text articles for compliance with the inclusion criteria and selected studies eligible for inclusion in the review. We resolved disagreements about study eligibility by discussion with another review author (WMP). We corresponded with study authors, as required, to clarify study eligibility. The selection process is documented in a PRISMA flow chart (Moher 2009; Figure 1).

1.

Study flow diagram

Data extraction and management

In this update, two review authors (MFT and DCR) independently extracted data from eligible studies using a data extraction form designed and pilot‐tested by the review authors. We resolved any disagreements by discussion or by consulting a third review author (LASL). Data extracted included study characteristics and outcome data. Where studies had multiple publications, we used the main trial report as the reference and derived additional details from secondary papers. We corresponded with study investigators for further data on methods or results as required.

Assessment of risk of bias in included studies

Two review authors (MFT and LASL) independently assessed the included studies for risk of bias using Cochrane RoB 1 (Higgins 2011), to assess: selection (random sequence generation and allocation concealment); performance (blinding of participants and personnel); detection (blinding of outcome assessors); attrition (incomplete outcome data); reporting (selective reporting); and other bias. We assigned judgements as recommended in section 8.5 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved disagreements through discussion. We described all judgements fully and presented the conclusions in the risk of bias table. We incorporated the consideration of bias risk into the interpretation of review findings by means of sensitivity analyses (see Sensitivity analysis).

When evaluating selective reporting bias, we took care to search for within‐trial selective reporting, such as studies failing to report obvious outcomes, or reporting them in insufficient detail to allow inclusion. We sought published protocols and compared the outcomes between the protocol and the final published study.

Measures of treatment effect

For continuous data measured with the same scale, we calculated the mean difference (MD) using the inverse variance method and for continuous data measured with different scales, we calculated the standardized mean difference (SMD; Suissa 1991; Walter 2001), using the inverse variance method. We used a fixed‐effect model unless we considered statistical heterogeneity to be substantially high, that is, with I² statistic (Higgins 2003), values of above 50%, then we used a random‐effects model. Wherever studies presented two groups with the same intervention (for example two different dosages), we combined them to create a single pair‐wise comparison to prevent loss of information related to the results (Higgins 2022). We reversed the direction of the effect of individual studies, if required, to ensure consistency across studies. We treated ordinal data as continuous data. We presented 95% confidence intervals (CI) for all outcomes. Where data to calculate SMDs were not available, we utilized the most detailed numerical data available that might facilitate a similar analysis of included studies (for example, P values). When studies reported sufficient detail to calculate MDs but gave no information on the associated standard deviation (SD), we assumed the outcome to have an SD equal to the highest SD from other studies using the same assessment scale. We compared the magnitude and direction of effects reported by studies with how they were presented in the review, taking into account legitimate differences. We included both data reported as final mean scores in each group and mean change scores from baseline in each group. Where studies reported both values, we preferentially included mean change scores from the baseline.

We graded the effect size according to the SMD or MD, based on the following criteria: SMD or MD less than 0.41 was equivalent to a small effect, SMD or MD from 0.40 to 0.70 was a moderate effect, SMD or MD more than 0.70 was a large effect (Norman 2003). We interpreted the effect size according to the recommendations of Cochrane Gynaecology and Fertility. We interpreted the categories based on the following criteria: SMD more than 0.7 was equivalent to important results; SMD from 0.2 to 0.7 means less important results; and SMD less than 0.2 means non‐important results, based on established criteria for effect size (Norman 2003).

Unit of analysis issues

The primary analysis was per randomized woman.

Dealing with missing data

We analyzed the data on an intention‐to‐treat basis as far as possible and attempted to obtain missing data from the original studies. If studies reported sufficient detail to calculate MDs but did not give any information on the associated SDs, we assumed the outcomes to have SDs equal to the highest SD from other studies within the same analysis.

Assessment of heterogeneity

We considered whether the clinical (participants and type of hormonal therapy) and methodological (study design, time cut‐off points in the measurement of sexual function) characteristics of the included studies were sufficiently similar for meta‐analysis to provide a clinically meaningful summary. We assessed statistical heterogeneity by the I² statistic (Higgins 2003). An I² statistic value greater than 50% was taken to indicate substantial heterogeneity (Deeks 2022). Where we found substantial heterogeneity, we explored additional sources of heterogeneity in the cases of the diversity of measurement tools used to assess sexual function and different doses and routes of administration of hormonal therapy.

Assessment of reporting biases

In view of the difficulty in detecting and correcting for publication bias and other reporting biases, the review authors aimed to minimize their potential impact by ensuring a comprehensive search for eligible studies and by being alert for duplication of data. If there were 10 or more studies in an analysis, we planned to use a funnel plot to explore the possibility of small study effects (a tendency for estimates of the intervention effect to be more beneficial in smaller studies).

Data synthesis

If the studies were sufficiently similar, we combined the data using a fixed‐effect or random‐effects model in the following comparisons.

1. Estrogen alone versus placebo or no intervention, stratified by the characteristics of the included participants:

   1. studies that only included symptomatic women (symptoms such as hot flushes, night sweats, sleep disturbance, vaginal atrophy, and dyspareunia) or early postmenopausal women (within the five years after menopause), to verify if hormone therapy could improve the sexual function of women complaining of sexual dysfunction with the onset of menopause;

   2. studies that included unselected postmenopausal women (regardless of menopausal symptoms, with or without sexual dysfunction) to verify whether hormone therapy can improve the sexual function of these women.

2. Estrogen combined with progestogens versus placebo or no intervention:

   1. studies that only included women with menopausal symptoms or women with less than five years since the onset of menopause, or both (early postmenopausal);

   2. studies that included unselected postmenopausal women.

3. Synthetic steroids (tibolone) versus placebo or no intervention:

   1. studies that only included women with menopausal symptoms or women with less than five years since the onset of menopause, or both (early postmenopausal);

   2. studies that included unselected postmenopausal women.

4. SERMs (raloxifene and bazedoxifene) versus placebo or no intervention:

   1. studies that only included women with menopausal symptoms or women with less than five years since the onset of menopause, or both (early postmenopausal);

   2. studies that included unselected postmenopausal women.

5. SERMs (bazedoxifene) combined with estrogen versus placebo or no intervention:

   1. studies that only included women with menopausal symptoms or women with less than five years since the onset of menopause, or both (early postmenopausal);

   2. studies that included unselected postmenopausal women.

An increase in a particular outcome, which may be beneficial (for example composite score) or detrimental (for example dyspareunia), was displayed graphically in the meta‐analysis to the right of the centre‐line and a decrease in an outcome to the left of the centre‐line. We merged women with menopausal symptoms with early postmenopausal women (STRAW stage +1; STRAW 2008), because menopausal‐related symptoms are common in early postmenopausal women with almost half of such women reporting experiencing vasomotor symptoms (Gold 2000). We did not observe any important heterogeneity when we compared the subgroups used for stratification. We planned to carry out pooled analysis when there was no significant heterogeneity, that is, when we observed an I² statistic of less than 50% in the subgroup difference test.

Subgroup analysis and investigation of heterogeneity

We planned the stratification based on the characteristics of the included participants described above (studies that only included women with menopausal symptoms or less than five years since the onset of menopause, or both; and studies that included unselected postmenopausal women) and according to clinically reasonable time points after the intervention (a group between 12 and 16 weeks; and a group between 22 and 26 weeks, depending on the available studies). When we detected substantial heterogeneity, we explored possible explanations by again checking the data and by sensitivity analyses. If it persisted, we adopted a random‐effects model in order to have a more conservative estimate of the effect. We took any statistical heterogeneity into account when interpreting the results, especially if there was a variation in the direction of effect.

Sensitivity analysis

We conducted sensitivity analyses for the primary outcome (sexual function composite score) to determine whether the conclusion would be different if we restricted eligibility to studies without a high risk of bias for our main comparison, hormone therapy compared with placebo or no intervention for sexual function composite score in symptomatic or recently postmenopausal women; and unselected postmenopausal women.

Summary of findings and assessment of the certainty of the evidence

We prepared summary of findings tables using GRADEpro GDT and Cochrane methods for hormonal therapy compared with no intervention or placebo for the main review outcome, sexual function composite score (Schünemann 2021a). The summary of findings tables show the evaluation of the overall quality of the body of evidence for sexual function composite score for the review comparisons (estrogen alone; estrogen in combination with progestogens; synthetic steroids; SERMs; and SERMs in combination with estrogen) for symptomatic or recently postmenopausal women (Table 1) and unselected postmenopausal women (Table 2).

We assessed the quality of the evidence using GRADE criteria: risk of bias, consistency of effect, indirectness, imprecision, and publication bias (Schünemann 2021b). Two review authors working independently made judgements about evidence quality (high, moderate, low, or very low), with disagreements resolved by discussion. We justified, documented, and incorporated our judgements into the reporting of results for each outcome. We interpreted the effect according to guidance from Cochrane Gynaecology and Fertility.

Results

Description of studies

Results of the search

We performed the last search for this updated review on 19 December 2022. We included nine new studies in this update (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Caruso 2020; Constantine 2015; Diem 2018; Mitchell 2018; Seyyedi 2016; Taylor 2017).

At the update, we assessed 80 studies at the full‐text stage and excluded 59 studies. See Figure 1 for an overview of the search results. We tried to contact all study authors to retrieve unpublished details. See details of included studies in Characteristics of included studies and reasons for excluding studies in the Characteristics of excluded studies.

Seven studies are awaiting classification because we are awaiting a response from the authors (Hirschberg 2020; NCT01845649; NCT02460302; NCT02834312; NCT03363997; Pinkerton 2017; Tanmahasamut 2020), and four studies are ongoing because they are still at the protocol stage or have not yet reported the results on the effect of hormone therapy on sexual function (NCT02668796; NCT02859285; NCT02670785; NCT03116022).

Included studies

We included 36 studies: 34 parallel‐group and two cross‐over studies (Fonseca 2007; Laan 2001). Sixteen studies were single‐centre studies. Four were conducted in the USA (Constantine 2015; Dayal 2005; Mitchell 2018; Taylor 2017), three in Brazil (Fonseca 2007; Mameri Filho 2005; Morais‐Socorro 2012), two in Italy (Caruso 2020; Gambacciani 2011), one in Iran (Seyyedi 2016), one in Poland (Czarnecka 2000), one in Romania (Hudita 2003), one in Taiwan (Yang 2004), one in Thailand (Bumphenkiatikul 2020), one in Turkey (Osmanagaoglu 2006) and one in the Netherlands (Laan 2001). Fourteen studies were multi‐centre, performed in only one country: nine in the USA (Archer 2019;Bushmakin 2014;Diem 2018; Huang 2008; Maki 2007; SMART‐3; Speroff 2003; Strickler 2000; WHI 1999), two in Estonia (Hemminki 2009; Veerus 2006), two in Sweden (Nathorst‐Böös 1993; Wiklund 1993), and one in Denmark (Nielsen 2006). Seven studies were multi‐centre studies conducted in more than one country: two in Canada and USA (Modugno 2003; Simon 2007); one in Brazil and USA (Gast 2009); one in Hong Kong, Malaysia, Philippines, Singapore, and Thailand (Haines 2009); one in Australia, New Zealand, and UK (WISDOM 2008); and one study was conducted in 31 countries (Liberate 2011).

Twenty studies were suitable for meta‐analysis, of these, 12 were included in the previous review (Dayal 2005; Haines 2009; Laan 2001; Liberate 2011; Maki 2007; Morais‐Socorro 2012; Nathorst‐Böös 1993; Osmanagaoglu 2006; SMART‐3; Speroff 2003; Wiklund 1993; Yang 2004), and eight were newly included studies for meta‐analysis (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Caruso 2020; Constantine 2015; Diem 2018; Mitchell 2018; Seyyedi 2016). Sixteen studies evaluated sexual function, but the reported data were not suitable for meta‐analysis (Czarnecka 2000; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Huang 2008; Hudita 2003; Mameri Filho 2005; Modugno 2003; Nielsen 2006; Simon 2007; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008). Twenty‐five studies received funding from the manufacturer of the study drug .

Participants

The studies included 12,225 women in the treatment groups and 11,074 in the control groups: the total score and all domains of sexual function were assessed in 2554 women in the treatment groups and 2296 women in the control groups. We divided the available trial subgroups according to the participants’ characteristics (symptomatic or early postmenopausal, or unselected postmenopausal women; Types of participants).

Ten studies included symptomatic or early postmenopausal women (Bumphenkiatikul 2020; Bushmakin 2014; Gambacciani 2011; Haines 2009; Liberate 2011; Mameri Filho 2005; Morais‐Socorro 2012; SMART‐3; Speroff 2003; Wiklund 1993), and 26 studies included unselected postmenopausal women (Archer 2019; Caruso 2020; Constantine 2015; Czarnecka 2000; Dayal 2005; Diem 2018; Fonseca 2007; Gast 2009; Hemminki 2009; Huang 2008; Hudita 2003; Laan 2001; Maki 2007; Mitchell 2018; Modugno 2003; Nathorst‐Böös 1993; Nielsen 2006; Osmanagaoglu 2006; Seyyedi 2016; Simon 2007; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008; Yang 2004). No study included only women with sexual dysfunction, and only seven studies evaluated sexual function as a primary outcome (Constantine 2015; Fonseca 2007; Gast 2009; Mameri Filho 2005; Nathorst‐Böös 1993; Osmanagaoglu 2006; Taylor 2017). One study included only women in perimenopause (Morais‐Socorro 2012); two studies included only women within 36 months of menopause (Maki 2007; Taylor 2017); one included only women up to five years after the last menstrual period (Nielsen 2006); six studies included only women with vasomotor or other menopausal symptoms (Archer 2019; Gambacciani 2011; Haines 2009; Nathorst‐Böös 1993; Speroff 2003; Wiklund 1993); one study included women with hot flushes and sexual dysfunction (Mameri Filho 2005).

Interventions

Out of the 36 included studies:

• 18 studies compared estrogen alone versus control (Caruso 2020; Czarnecka 2000; Dayal 2005; Diem 2018; Fonseca 2007; Haines 2009; Huang 2008; Mameri Filho 2005; Mitchell 2018; Nathorst‐Böös 1993; Nielsen 2006; Seyyedi 2016; Simon 2007; Speroff 2003; Taylor 2017; WHI 1999; Wiklund 1993; WISDOM 2008);

• nine studies compared estrogen combined with progestogens versus control (Gambacciani 2011; Gast 2009; Hemminki 2009; Maki 2007; Osmanagaoglu 2006; Strickler 2000; Veerus 2006; WHI 1999; Yang 2004);

• five studies compared synthetic steroids (tibolone) versus control (Hudita 2003; Laan 2001; Liberate 2011; Morais‐Socorro 2012; Osmanagaoglu 2006);

• five studies compared SERMs (raloxifene, bazedoxifene, ospemifene) versus control (Archer 2019; Bushmakin 2014; Modugno 2003; SMART‐3; Strickler 2000);

• two studies compared SERMs combined with estrogen versus control (Bushmakin 2014; SMART‐3);

• five studies investigated the effect of more than one hormone versus control (Bushmakin 2014; Osmanagaoglu 2006; SMART‐3; Strickler 2000; WHI 1999).

In 19 studies the follow‐up period was three months (Archer 2019; Bumphenkiatikul 2020; Constantine 2015; Czarnecka 2000; Dayal 2005; Diem 2018; Gambacciani 2011; Gast 2009; Haines 2009; Laan 2001; Mameri Filho 2005; Mitchell 2018; Morais‐Socorro 2012; Nathorst‐Böös 1993; Seyyedi 2016; Simon 2007; Speroff 2003; Taylor 2017; Wiklund 1993); in one study the follow‐up period was 5.5 months (Yang 2004); in three studies the follow‐up was 12 weeks (Bushmakin 2014; Hemminki 2009; SMART‐3); in two studies the follow‐up period was four months (Caruso 2020; Maki 2007); in four studies the follow‐up period was six months (Fonseca 2007; Hudita 2003; Osmanagaoglu 2006; Veerus 2006); in one study the follow‐up period was 26 weeks (Liberate 2011); in three studies the follow‐up period was 12 months (Strickler 2000; WHI 1999; WISDOM 2008); in two studies the follow‐up period was 24 months (Huang 2008; Nielsen 2006); and in one study the follow‐up period was 36 months (Modugno 2003).

In 33 studies women in the control group used a matching placebo and in three studies women in the control group did not receive any intervention (Czarnecka 2000; Osmanagaoglu 2006; Veerus 2006). We merged groups where studies had more than one group that used the same drug, in different doses (Bushmakin 2014; Caruso 2020; SMART‐3; Speroff 2003; Strickler 2000). Sixteen studies evaluated sexual function but the reported data were not suitable for meta‐analysis (Czarnecka 2000; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Huang 2008; Hudita 2003; Mameri Filho 2005; Modugno 2003; Nielsen 2006; Simon 2007; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008). Of those studies, eight compared estrogen combined with progestogens against placebo (Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Maki 2007; Veerus 2006; WHI 1999; WISDOM 2008), and two studies compared synthetic steroids (tibolone) with placebo (Liberate 2011; Osmanagaoglu 2006)

Most interventions were delivered orally (Archer 2019; Bushmakin 2014; Constantine 2015; Dayal 2005; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Hudita 2003; Laan 2001; Liberate 2011; Maki 2007; Mameri Filho 2005; Modugno 2003; Morais‐Socorro 2012; Osmanagaoglu 2006; SMART‐3; Strickler 2000; Veerus 2006; Yang 2004), however, some were administered transdermally (Czarnecka 2000; Haines 2009; Huang 2008; Nathorst‐Böös 1993; Simon 2007; Wiklund 1993), vaginally (Bumphenkiatikul 2020; Caruso 2020; Diem 2018; Gast 2009; Mitchell 2018; Seyyedi 2016; Speroff 2003), orally and transdermally in separate groups (Taylor 2017), and nasally (Nielsen 2006).

Outcomes

Studies used the following tools for assessing sexual function:

• Arizona Sexual Experiences (ASEX) scale (McGahuey 2000 (SMART‐3))

• Menopause‐specific Quality of Life (MENQOL; Hilditch 1996 (Bushmakin 2014; Diem 2018; Haines 2009; Seyyedi 2016; SMART‐3))

• Golombok‐Rust Inventory of Sexual Satisfaction (GRISS; Rust 1986 (Fonseca 2007))

• McCoy Female Sexuality Questionnaire (MFSQ; McCoy 2000 (Gast 2009))

• Greene Climacteric Scale (GCS; Greene 2008 (Laan 2001; Maki 2007; Morais‐Socorro 2012; Speroff 2003; Yang 2004))

• McCoy Sex Scale Questionnaire (MSQ; McCoy 1985 (Hudita 2003; Mameri Filho 2005; Modugno 2003; Nathorst‐Böös 1993; Wiklund 1993))

• Brief Index of Sexual Functioning‐Women (BISF‐W; Taylor 1994 (Gast 2009))

• Medical Outcomes Study (MOS; McHorney 1994 (Huang 2008))

• Modified Mini‐Mental State Examination (3MS; Teng 1987 (SMART‐3; WHI 1999))

• Female Sexual Function Index (FSFI; Rosen 2000 (Archer 2019; Bumphenkiatikul 2020; Caruso 2020; Constantine 2015; Mitchell 2018; Osmanagaoglu 2006; Taylor 2017))

• Subjective Symptoms Assessment Profile (SSA‐P; Dimenäs 1990 (Czarnecka 2000))

• Utian Quality of Life Women's Health Questionnaire (WHQ; Hunter 2000 (Dayal 2005; Gambacciani 2011; Hemminki 2009; Liberate 2011; Mameri Filho 2005; Nathorst‐Böös 1993; Nielsen 2006; Simon 2007; Strickler 2000; Veerus 2006; WISDOM 2008))

See the Characteristics of included studies for details.

Where studies reported composite scores, they evaluated the following domains.

• All domains studied ‐ desire, arousal, lubrication, orgasm, satisfaction, and pain (Archer 2019; Constantine 2015; Mitchell 2018; Osmanagaoglu 2006; Taylor 2017)

• Arousal and orgasm (Czarnecka 2000; SMART‐3)

• Arousal and pain (Haines 2009; Liberate 2011; Nielsen 2006; Strickler 2000; WISDOM 2008)

Of the 16 studies included in the qualitative analysis, 20 studies presented complete data for all outcomes (Archer 2019; Bushmakin 2014; Bumphenkiatikul 2020; Caruso 2020; Constantine 2015; Dayal 2005; Diem 2018; Haines 2009; Laan 2001; Liberate 2011; Maki 2007; Mitchell 2018; Morais‐Socorro 2012; Nathorst‐Böös 1993; Osmanagaoglu 2006; Seyyedi 2016; SMART‐3; Speroff 2003; Wiklund 1993; Yang 2004). From three studies we could only include the final scores (Bumphenkiatikul 2020; Laan 2001; Morais‐Socorro 2012). We used the calculator in Review Manager 2020 to estimate missing SDs for seven studies (Bushmakin 2014; Caruso 2020; Liberate 2011; Maki 2007; Nathorst‐Böös 1993; SMART‐3; Speroff 2003).

Excluded studies

We excluded 59 studies at this update. Of these:

• 36 studies did not evaluate sexual function (2010‐024603‐26; 2014‐000179‐18; Abraham 2014; Archer 2011; Bosak 2022; Bride 2014; Cano 2012; Daneshmand 2015; Fabian 2019; Galesanu 2010; Garcia de Arriba 2022; Gaspard 2021; Griesser 2012; Harman 2014; IRCT138809032766N1; ISRCTN46661996; Kagan 2014; Kingsberg 2017; Kroll 2018; L'Hermite 2022; Liu 2012; Lobo 2017; NCT01585558; NCT01779947; NCT02253173; NCT02638337; Noguchi 2018; Paraiso 2020; Portman 2013; Portman 2014a; Pradeep 2016; Santofimia 2022; Silverman 2012; Simon 2012; Sun 2016; Tansupswatdikul 2012);

• six studies were not randomized or were pseudo‐randomized (Bennink 2016; Islam 2021; Ju 2021; NCT01822288; NCT03018106; Tungmunsakulchai 2015);

• three studies compared sexual function with another intervention (Cruz 2018; Fernandes 2014; Bianchi‐Ferraro 2022);

• four studies did not evaluate sexual function by a validated questionnaire (Aptecar 2021; Bachmann 2011; Mittal 2020; Saeideh 2010);

• eight studies did not use hormone therapy (Abedi 2018; Abedi 2020; Arêas 2019; Dalil 2020; RBR‐94dx93/1; Raz 2014; Serebani 2020; Yeo 2020);

• one study did not have a control group (Eftekhar 2020);

• one study's patient population was ineligible (Marschalek 2022).

Studies awaiting classification

Seven studies (Hirschberg 2020; NCT01845649; NCT02460302; NCT02834312; NCT03363997; Pinkerton 2017; Tanmahasamut 2020 ) are awaiting classification because we suspect that they evaluated the effect of hormone therapy on sexual function, but we have not yet obtained these data, and we are awaiting information from the authors.

Ongoing studies

Four studies were still at the protocol stage and had not yet reported results on the effect of hormone therapy on sexual function (NCT02668796; NCT02670785; NCT02859285; NCT03116022).

Risk of bias in included studies

We deemed 20 of 36 studies included in this review to have a high risk of bias in at least one domain. We considered 14 of 36 studies not to be at high risk of bias in any domain but to have an unclear risk of bias in at least one domain. We considered two of 36 studies to have a low risk of bias over all domains. See Figure 2 and Figure 3 for the graph and summary of the review authors' judgements about the risk of bias for included studies. The risk of bias and imprecision of evidence for many comparisons were reasons for downgrading evidence quality. Many included studies were supported by the pharmaceutical industry, in most cases the manufacturer of the studied drug (see Characteristics of included studies).

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Allocation

Sequence generation

We judged 14 studies to have unclear risk of selection bias as they did not describe the method of randomization (Caruso 2020; Constantine 2015; Czarnecka 2000; Fonseca 2007; Gast 2009; Haines 2009; Hudita 2003; Laan 2001; Modugno 2003; Morais‐Socorro 2012; Nathorst‐Böös 1993; Osmanagaoglu 2006; Wiklund 1993; Yang 2004). The other 22 studies reported adequate methods of randomization, and we deemed them to be at low risk of bias (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Dayal 2005; Diem 2018; Gambacciani 2011; Hemminki 2009; Huang 2008; Liberate 2011; Maki 2007; Mameri Filho 2005; Mitchell 2018; Nielsen 2006; Seyyedi 2016; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008).

Allocation concealment

Twenty‐seven studies did not report the method used for allocation concealment, and thus we considered them to be at unclear risk of selection bias (Archer 2019; Bushmakin 2014; Caruso 2020; Constantine 2015; Czarnecka 2000; Dayal 2005; Fonseca 2007; Gambacciani 2011; Gast 2009; Haines 2009; Hudita 2003; Laan 2001; Liberate 2011; Maki 2007; Mameri Filho 2005; Mitchell 2018; Modugno 2003; Morais‐Socorro 2012; Nathorst‐Böös 1993; Osmanagaoglu 2006; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; WHI 1999; Wiklund 1993; Yang 2004). The other nine studies reported adequate allocation concealment methods, and we deemed them to be at low risk of selection bias (Bumphenkiatikul 2020; Diem 2018; Hemminki 2009; Huang 2008; Nielsen 2006; Seyyedi 2016; Taylor 2017; Veerus 2006; WISDOM 2008).

Blinding

Blinding of participants and personnel

Three studies did not blind the participants or the personnel administering the intervention, and we considered them to be at high risk of performance bias (Czarnecka 2000; Osmanagaoglu 2006; Veerus 2006). Six studies did not describe how they blinded participants, and we considered them to be at unclear risk of performance bias (Caruso 2020; Constantine 2015; Gambacciani 2011; Laan 2001; Modugno 2003; Nathorst‐Böös 1993), whilst we deemed 27 studies to be at low risk of performance bias (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Dayal 2005; Diem 2018; Fonseca 2007; Gast 2009; Haines 2009; Hemminki 2009; Huang 2008; Hudita 2003; Liberate 2011; Maki 2007; Mameri Filho 2005; Mitchell 2018; Morais‐Socorro 2012; Nielsen 2006; Seyyedi 2016; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; Taylor 2017; WHI 1999; Wiklund 1993; WISDOM 2008; Yang 2004).

Blinding of outcome assessment

Seven studies described attempts to blind the outcome assessment, and we deemed them to be at low risk of detection bias (Bumphenkiatikul 2020; Liberate 2011; Mitchell 2018; Simon 2007; Taylor 2017; Wiklund 1993; WISDOM 2008), whilst we considered 29 studies to be at unclear risk of detection bias (Archer 2019; Bushmakin 2014; Caruso 2020; Constantine 2015; Czarnecka 2000; Dayal 2005; Diem 2018; Fonseca 2007; Gambacciani 2011; Gast 2009; Haines 2009; Hemminki 2009; Huang 2008; Hudita 2003; Laan 2001; Maki 2007; Mameri Filho 2005; Modugno 2003; Morais‐Socorro 2012; Nathorst‐Böös 1993; Nielsen 2006; Osmanagaoglu 2006; Seyyedi 2016; SMART‐3; Speroff 2003; Strickler 2000; Veerus 2006; WHI 1999; Yang 2004).

Incomplete outcome data

Ten studies reported an important or unbalanced loss of participants, and we deemed them to be at high risk of attrition bias (Constantine 2015; Haines 2009; Huang 2008; Laan 2001; Liberate 2011; Modugno 2003; Nielsen 2006; Simon 2007; WHI 1999; WISDOM 2008). In six studies, losses to follow‐up were high or not clear, and we deemed them to be at unclear risk of bias (Bushmakin 2014; Gambacciani 2011; Hudita 2003; Nathorst‐Böös 1993; SMART‐3; Yang 2004); and in 19 studies, there was no suspicion of incomplete data, and we considered them to be at low risk of bias (Archer 2019; Bumphenkiatikul 2020; Caruso 2020; Czarnecka 2000; Dayal 2005; Diem 2018; Fonseca 2007; Gast 2009; Hemminki 2009; Maki 2007; Mameri Filho 2005; Mitchell 2018; Morais‐Socorro 2012; Osmanagaoglu 2006; Seyyedi 2016; Speroff 2003; Strickler 2000; Taylor 2017; Veerus 2006; Wiklund 1993).

Selective reporting

Six studies applied a validated questionnaire to evaluate the sexual function (Caruso 2020; Constantine 2015; Fonseca 2007; Gambacciani 2011; Gast 2009; Hudita 2003), but one study did not provide the protocol and showed selective reporting and other bias (possible biases in randomization or blinding; Caruso 2020), one did not provide baseline domain scores or demographic data after losses, and losses were unbalanced in the groups (Constantine 2015); one provided only results for significant differences (Fonseca 2007), and three did not provide numeric results (Gambacciani 2011; Gast 2009; Hudita 2003). One study assessed sexuality, sexual thoughts/interest, or quality of sexual life, not sexual function, by validated questionnaires, but provided only partial results (total score and some domains) of one questionnaire, which do not represent sexual function itself (Maki 2007). Thus, we deemed all seven of these studies to be at high risk of reporting bias. One study was at unclear risk (Seyyedi 2016), while the remaining 28 were considered at low risk (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Czarnecka 2000; Dayal 2005; Diem 2018; Haines 2009; Hemminki 2009; Huang 2008; Laan 2001; Liberate 2011; Mameri Filho 2005; Mitchell 2018; Modugno 2003; Morais‐Socorro 2012; Nathorst‐Böös 1993; Nielsen 2006; Osmanagaoglu 2006; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; Wiklund 1993; WISDOM 2008; Yang 2004).

Other potential sources of bias

We considered four studies to be at high risk of bias because they had unbalanced groups at baseline and we could use only final scores in the meta‐analysis, or due to lack of detailed information about randomization of the participants (Caruso 2020; Czarnecka 2000; Dayal 2005; Osmanagaoglu 2006). We considered four studies to be at unclear risk of bias due to problems in the study design, or absence of a declaration of conflicts of interest, or lack of sample size (Constantine 2015; Maki 2007; Wiklund 1993; Yang 2004). We considered the other 28 studies to be at low risk of other bias (Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Diem 2018; Fonseca 2007; Gambacciani 2011; Gast 2009; Haines 2009; Hemminki 2009; Huang 2008; Hudita 2003; Laan 2001; Liberate 2011; Mameri Filho 2005; Mitchell 2018; Modugno 2003; Morais‐Socorro 2012; Nathorst‐Böös 1993; Nielsen 2006; Seyyedi 2016; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008).

Effects of interventions

See: Table 1; Table 2

See Table 1 for the main comparison, hormone therapy compared with placebo or no intervention for sexual function in symptomatic or early postmenopausal women; see Table 2 for hormone therapy compared with placebo or no intervention for sexual function in unselected postmenopausal women.

Qualitative analysis

We did not include 16 studies in the meta‐analysis (Czarnecka 2000; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Huang 2008; Hudita 2003; Mameri Filho 2005; Modugno 2003; Nielsen 2006; Simon 2007; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008).

Six of 16 studies administered estrogen alone versus placebo (Huang 2008; Mameri Filho 2005; Nielsen 2006; Simon 2007; Taylor 2017; WHI 1999). Of these:

• Huang 2008 administered 0.014 mg of estrogen per day via transdermal patch for 24 months and reported no significant difference in sexual function when comparing the effect of the intervention with placebo in unselected postmenopausal women;

• Mameri Filho 2005 administered 0.625 mg of conjugated equine estrogen per day, orally, for six months and reported no significant difference in sexual function when comparing the effect of the intervention with placebo in unselected postmenopausal women;

• Nielsen 2006 administered 150 mcg or 300 mcg of estrogen daily for six months, intranasally, and reported a slight improvement in sexual function when comparing the effect of the intervention with placebo in symptomatic or early postmenopausal women;

• Simon 2007 administered 0.87 g or 1.7 g or 2.6 g of estrogen per day for 12 months, by transdermal gel and reported a significant improvement in sexual function when comparing the effect of the intervention with placebo in postmenopausal women;

• Taylor 2017 administered oral estrogen 0.45 mg per day, orally, for 36 months and reported a significant improvement in sexual function when comparing the effect of the intervention with placebo in postmenopausal women;

• WHI 1999 administered 0.625 mg of conjugated equine estrogen per day for 12 months, orally, and reported no significant difference in effect in the domain of sexual satisfaction when comparing the intervention with placebo in postmenopausal women.

Eight of 16 studies administered estrogen in combination with progesterone (Czarnecka 2000; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Veerus 2006; WHI 1999; WISDOM 2008). Of these:

• Gast 2009 administered conjugated estrogen 0.45 mg + 1.5 mg medroxyprogesterone acetate daily for six 28‐day cycles intravaginally and reported a significant improvement in sexual function in postmenopausal women;

• Czarnecka 2000 and Fonseca 2007 administered for three and six months, respectively, 17β‐estradiol 2 mg in combination with 1 mg of norethisterone acetate daily in postmenopausal women.

  • Czarnecka 2000 reported no significant differences between women who received the intervention and controls;

  • Fonseca 2007 reported an improvement in sexual function, specifically, in the decrease in dyspareunia and increase in sexual satisfaction;

• Hemminki 2009 administered oral conjugated equine estrogen 0.625 mg + medroxyprogesterone acetate 5.0 daily with assessment at 12 months, and reported no significant difference in effect on sexual functioning when compared with placebo in postmenopausal women;

• Gambacciani 2011, on the other hand, administered 1 mg of estradiol in combination with 2 mg drospirenone orally for three months and did not report significant differences in sexual function in symptomatic or early postmenopausal women;

• Veerus 2006 administered oral conjugated equine estrogen 0.625 mg + medroxyprogesterone acetate 5.0 daily with assessment at 12 months, and reported no significant difference in effect on sexual functioning when compared with placebo in postmenopausal women, according to the final data reported in the study protocol.

• WHI 1999 and WISDOM 2008 administered 0.625 mg of conjugated equine estrogen in combination with 2.5 or 0.625 mg of medroxyprogesterone acetate (MPA) per day orally, for more than three years

  • WHI 1999 reported a significant improvement in sexual function;

  • WISDOM 2008, however, reported no significant differences in sexual function in postmenopausal women.

Two of 16 studies administered SERMs (Strickler 2000; Modugno 2003). Strickler 2000 administered 60 mg or 150 mg of raloxifene for six months, and Modugno 2003 administered 60 mg or 120 mg of raloxifene per day orally for 36 months. Both studies found no significant difference in sexual function in postmenopausal women.

Finally, a further study on tibolone (1.25 or 2.5 mg/day orally for 24 weeks) did not report data in a form that could be used in meta‐analysis but reported significant improvement in sexual function in postmenopausal women (Hudita 2003).

Regarding the quality of these 16 studies, two had no suspected risk of bias (Hemminki 2009; Taylor 2017), while 14 had suspected risk of bias mainly associated with incomplete data (Huang 2008; Modugno 2003; Mameri Filho 2005Nielsen 2006; Simon 2007; Strickler 2000; WHI 1999; WISDOM 2008), blinding of participants and staff (Czarnecka 2000; Veerus 2006), and selective reporting (Fonseca 2007; Gambacciani 2011; Gast 2009; Hudita 2003). See Figure 3.

Quantitative analysis

1. Estrogen alone versus placebo or no intervention

Primary outcome

1.1 Sexual function composite score

In the subgroup of symptomatic or early postmenopausal women, estrogen alone versus placebo or no intervention probably slightly improves the sexual function composite score (SMD 0.50, 95% CI 0.04 to 0.96; I² = 88%; 3 studies, 699 women; moderate‐quality evidence; Analysis 1.1; Figure 4). This means that the sexual function composite score is compatible with a moderate effect of 0.50 standard units with the use of estrogen alone versus placebo or no intervention. The sensitivity analysis (restricting eligibility criteria only to studies judged not be at high risk of bias) was compatible with a moderate effect (SMD 0.68, 95% CI 0.20 to 1.16; I² = 86%; 2 studies, 548 women; moderate‐quality evidence) of estrogen alone versus placebo or no intervention in this subgroup of women. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 88%) could be due to:

1.1. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 1: Sexual function composite score

4.

Forest plot of comparison: 1 Estrogens alone versus placebo/no treatment, outcome: 1.1 Sexual function composite score.

• clinical heterogeneity related to variability in the intervention (Haines 2009, estradiol 0.014 mg, transdermal route; Speroff 2003, estradiol 50 ug/day and 100 ug/day or placebo, vaginal route; Wiklund 1993, estradiol 50 ug twice a week or placebo, transdermal route);

• methodological heterogeneity related to variability in the measurement tool (Haines 2009, MENQoL; Speroff 2003, Greene Climacteric Scale sub‐score; Wiklund 1993, McCoy sex scale total score), and high risk of bias in one of three studies; and

• statistical heterogeneity related to the variability of effect (Haines 2009, SMD 0.12, 95% CI −0.20 to 0.44; Speroff 2003, SMD 0.92, 95% CI 0.68 to 1.17; Wiklund 1993, SMD 0.43, 95% CI 0.17 to 0.70).

In the subgroup of unselected postmenopausal women, estrogen alone versus placebo or no intervention probably makes little or no difference to the sexual function composite score (SMD 0.64, 95% CI −0.12 to 1.41; I² = 94%; 6 studies, 608 women; moderate‐quality evidence; Analysis 1.1; Figure 4). This means that the sexual function composite score is compatible with a moderate effect of 0.64 standard units with the use of estrogen alone versus placebo or no intervention. The sensitivity analysis (restricting eligibility criteria only to studies judged not be at high risk of bias) was compatible with a small effect (SMD 0.35, 95% CI 0.02 to 0.68; I² = 63%; 4 studies, 434 women) of estrogen alone versus placebo or no intervention in this subgroup of women. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 94%) could be due to:

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Dayal 2005, equine estrogen 0.625 mg/day or placebo, oral route; Diem 2018, estradiol 10 mg tablet/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug/day or placebo, vaginal route; and Seyyedi 2016, estradiol 0.625 mg/day or placebo, vaginal route);

• methodological heterogeneity related to variability in the measurement tool (Bumphenkiatikul 2020, Female Sexual Function Index; Caruso 2020, Female Sexual Function Index; Dayal 2005, health‐related quality of life; Diem 2018, MENQoL; Mitchell 2018, Female Sexual Function Index; Seyyedi 2016, MENQoL), and high risk of bias in two of six studies; and

• statistical heterogeneity related to global variability of effect.

Considering the pooled analysis, estrogen alone versus placebo or no intervention probably slightly improves the sexual function composite score (SMD 0.60, 95% CI 0.16 to 1.04; I² = 92%; 9 studies, 1307 women; moderate‐quality evidence; Analysis 1.1; Figure 4). This means that the sexual function composite score is compatible with a moderate effect of 0.60 standard units with the use of estrogen alone. The sensitivity analysis (restricting eligibility criteria only to studies judged not be at high risk of bias) was compatible with a moderate effect (SMD 0.48, 95% CI 0.18 to 0.79; I² = 80%; 6 studies, 982 women) of estrogen alone versus placebo or no intervention in this subgroup of women. The investigation of additional sources of bias for the high heterogeneity observed in the pooled analysis (I² = 94%), is related to the sources of clinical, methodological, and statistical heterogeneity explored in the two previous subgroups (symptomatic or early postmenopausal women and unselected postmenopausal women).

Secondary outcomes

1.2 Desire

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In unselected postmenopausal women, we are uncertain of the effect of estrogen alone versus placebo or no intervention in desire score (MD 0.33, 95% CI −0.27 to 0.93; I² = 77%; 3 studies, 416 women; low‐quality evidence; Analysis 1.2). This is compatible with a small effect of 0.06 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 77%) could be due to:

1.2. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 2: Desire

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to the variability of effect (Bumphenkiatikul 2020, MD 0.22, 95% CI −1.39 to 1.83; Caruso 2020, MD 0.71, 95% CI 0.33 to 1.08; Mitchell 2018, MD 0.00, 95% CI −0.29 to 0.29).

1.3 Arousal

In symptomatic or early postmenopausal women, we are uncertain of the effect of estrogen alone versus placebo or no intervention on arousal (SMD 0.14, 95% CI −0.12 to 0.39; 1 study, 239 women; low‐quality evidence; Analysis 1.3). This is compatible with a small effect of 0.14 standard units with the use of estrogen alone.

1.3. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 3: Arousal

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen alone versus placebo or no intervention on arousal (SMD 0.14, 95% CI −0.29 to 0.58; I² = 77%; 3 studies, 384 women; low‐quality evidence; Analysis 1.3). This is compatible with a small effect of 0.14 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 77%) could be related to:

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to variability of effect (Bumphenkiatikul 2020, SMD 0.22, 95% CI −0.26 to 0.70; Caruso 2020, SMD 0.44, 95% CI 0.13 to 0.75; Mitchell 2018, SMD −0.22, 95% CI −0.54 to 0.10).

Considering the pooled analysis, we are uncertain of the effect of estrogen alone versus placebo or no intervention (SMD 0.14, 95% CI −0.14 to 0.42; I² = 65%; 4 studies, 623 women; low‐quality evidence; Analysis 1.3). This is compatible with a small effect of 0.14 standard units with the use of estrogen alone. The investigation of additional sources of bias for the high heterogeneity observed in the pooled analysis (I² = 65%), is related to the sources of clinical, and statistical heterogeneity explored in the subgroup of unselected postmenopausal women.

1.4 Lubrication

In the subgroup of symptomatic or early postmenopausal women, estrogen alone versus placebo or no intervention may slightly improve lubrication (SMD 0.47, 95% CI 0.21 to 0.73; 1 study, 239 women; low‐quality evidence; Analysis 1.4). This is compatible with a moderate effect of 0.47 standard units with the use of estrogen alone.

1.4. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 4: Lubrication

In the subgroup of unselected postmenopausal women, estrogen alone versus placebo or no intervention probably slightly improves lubrication (SMD 0.47, 95% CI −0.09 to 1.02; I² = 85%; 3 studies, 380 women; moderate‐quality evidence; Analysis 1.4). This is compatible with a moderate effect of 0.47 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 85%) could be due to:

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to variability of effect (Bumphenkiatikul 2020, SMD 0.31, 95% CI −0.18 to 0.79; Caruso 2020, SMD 0.96, 95% CI 0.63 to 1.28; Mitchell 2018, SMD 0.12, 95% CI −0.20 to 0.44).

Considering the pooled analysis, estrogen alone versus placebo or no intervention probably slightly improves lubrication (SMD 0.47, 95% CI 0.12 to 0.83; I² = 78%; 4 studies, 619 women; moderate‐quality evidence; Analysis 1.4). This is compatible with a small effect of 0.47 standard units with the use of estrogen alone. The investigation of additional sources of bias suggests that the high heterogeneity observed in the pooled analysis (I² = 78%) is related to the sources of clinical, and statistical heterogeneity explored in the subgroup of unselected postmenopausal women.

1.5 Orgasm

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen alone versus placebo or no intervention for orgasm (MD 0.14, 95% CI −0.72 to 0.99; I² = 90%; 3 studies, 383 women; low‐quality evidence; Analysis 1.5). This is compatible with a small effect of 0.14 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 90%) could be due to:

1.5. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 5: Orgasm

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to the variability of effect (Bumphenkiatikul 2020, MD 0.36, 95% CI −0.93 to 0.21; Caruso 2020, MD 0.92, 95% CI 0.54 to 1.30; Mitchell 2018, MD −0.20, 95% CI −0.69 to 0.29).

1.6 Satisfaction

In the subgroup of symptomatic or early postmenopausal women, estrogen alone versus placebo or no intervention may slightly improve satisfaction (SMD 0.29, 95% CI 0.08 to 0.51; I² = 24%; 2 studies, 446 women; moderate‐quality evidence; Analysis 1.6). This is compatible with a small effect of 0.29 standard units with the use of estrogen alone.

1.6. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 6: Satisfaction

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen alone versus placebo or no intervention on satisfaction (SMD 0.48, 95% CI −0.21 to 1.16; I² = 90%; 3 studies, 378 women; low‐quality evidence; Analysis 1.6). This is compatible with a moderate effect of 0.48 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 90%) could be due to:

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to the variability of effect (Bumphenkiatikul 2020, SMD −0.06, 95% CI −0.54 to 0.42; Caruso 2020, SMD 1.14, 95% CI 0.81 to 1.47; Mitchell 2018, SMD 0.31, 95% CI −0.02 to 0.63).

Considering the pooled analysis, estrogen alone versus placebo or no intervention probably slightly improves satisfaction (SMD 0.41, 95% CI 0.04 to 0.77; I² = 84%; 5 studies, 824 women; moderate‐quality evidence; Analysis 1.6). This is compatible with a small effect of 0.41 standard units with the use of estrogen alone. The investigation of additional sources of bias for the high heterogeneity observed in the pooled analysis (I² = 84%), is related to the sources of clinical, and statistical heterogeneity explored in the subgroup of unselected postmenopausal women.

1.7 Pain

In the subgroup of symptomatic or early postmenopausal women, estrogen alone versus placebo or no intervention probably slightly improves pain (SMD 0.35, 95% CI 0.14 to 0.56; I² = 29%; 2 studies, 548 women; moderate‐quality evidence; Analysis 1.7). This is compatible with a small effect of 0.35 standard units with the use of estrogen alone.

1.7. Analysis.

Comparison 1: Estrogen alone versus placebo/no treatment, Outcome 7: Pain

In the subgroup of unselected postmenopausal women, estrogen alone versus placebo or no intervention may improve pain (SMD 0.72, 95% CI −0.09 to 1.52; I² = 93%; 3 studies, 410 women; low‐quality evidence; Analysis 1.7). This is compatible with a large effect of 0.72 standard units with the use of estrogen alone. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 93%) could be due to:

• clinical heterogeneity related to variability in the intervention (Bumphenkiatikul 2020, conjugated estrogen 0.625 mg, daily for 3 weeks then twice‐weekly for 9 weeks or placebo, vaginal route; Caruso 2020, estradiol ovule 0.03 mg/day or placebo, vaginal route; Mitchell 2018, estradiol 10.3 ug /day or placebo, vaginal route);

• methodological heterogeneity related to high risk of bias in one of three studies; and

• statistical heterogeneity related to variability of effect (Bumphenkiatikul 2020, SMD 0.46, 95% CI −0.02 to 0.95; Caruso 2020, SMD 1.44, 95% CI 1.12 to 1.77; Mitchell 2018, SMD 0.23 95% CI −0.08 to 0.54).

Considering the pooled analysis, estrogen alone versus placebo or no intervention may slightly improve pain (SMD 0.57, 95% CI 0.14 to 1.0; I² = 90%; 5 studies, 958 women; low‐quality evidence; Analysis 1.7). This is compatible with a moderate effect of 0.57 standard units with the use of estrogen alone. The investigation of additional sources of bias for the high heterogeneity observed in the pooled analysis (I² = 90%), is related to the sources of clinical, and statistical heterogeneity explored in the subgroup of unselected postmenopausal women.

2. Estrogen combined with progestogens versus placebo or no intervention

Primary outcome

2.1 Sexual function composite score

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on the sexual function composite score (MD 0.08 95% CI −1.52 to 1.68; 1 study, 104 women; very low‐quality evidence; Analysis 2.1; Figure 5). This means that sexual function is compatible with a small effect of 0.08 standard units with the use of estrogen combined with progestogens. Due to the limited studies available, it was not possible to perform sensitivity analyses.

2.1. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 1: Sexual function composite score

5.

Forest plot of comparison: 2 Estrogens + progestogens versus placebo/no treatment, outcome: 2.1 Sexual function composite score.

Secondary outcomes

2.2 Desire at 16‐month follow‐up

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on desire (MD 0.70, 95% CI 0.21 to 1.19, 1 study, 158 women, very low‐quality evidence; Analysis 2.2). This is compatible with a moderate effect of 0.70 standard units with the use of estrogen combined with progestogens.

2.2. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 2: Desire at 16‐month follow‐up

2.3 Desire at 22‐ and 24‐month follow‐up

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on desire (SMD 0.19, 95% CI −0.12 to 0.51; I² = 0%; 2 studies, 155 women; low‐quality evidence; Analysis 2.3). This is compatible with a small effect of 0.19 standard units with the use of estrogen combined with progestogens.

2.3. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 3: Desire at 22‐ and 24‐month follow‐up

2.4 Arousal

No study reported results for the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on arousal (MD 0.32, 95% CI −0.11 to 0.75; 1 study, 104 women; very low‐quality evidence; Analysis 2.4). This is compatible with a small effect of 0.32 standard units with the use of estrogen combined with progestogens.

2.4. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 4: Arousal

2.5 Lubrication

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, estrogen combined with progestogens versus placebo or no intervention may slightly improve lubrication (MD 0.45, 95% CI 0.13 to 0.77; 1 study, 104 women; low‐quality evidence; Analysis 2.5). This is compatible with a moderate effect of 0.45 standard units with the use of estrogen combined with progestogens.

2.5. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 5: Lubrication

2.6 Orgasm

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on orgasm (MD 0.04, 95% CI −0.53 to 0.61; 1 study, 104 women; very low‐quality evidence; Analysis 2.6). This is compatible with a small effect of 0.04 standard units with the use of estrogen combined with progestogens.

2.6. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 6: Orgasm

2.7 Satisfaction

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on satisfaction (MD 0.13, 95% CI −0.28 to 0.54; 1 study, 104 women; very low‐quality evidence; Analysis 2.7). This is compatible with a small effect of 0.13 standard units with the use of estrogen combined with progestogens.

2.7. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 7: Satisfaction

2.8 Pain

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of estrogen combined with progestogens versus placebo or no intervention on pain (MD 0.93, 95% CI 0.02 to 1.84; 1 study, 104 women; very low‐quality evidence; Analysis 2.8). This is compatible with a large effect of 0.93 standard units with the use of estrogen combined with progestogens.

2.8. Analysis.

Comparison 2: Estrogen + progestogens versus placebo/no treatment, Outcome 8: Pain

3. Synthetic steroids versus placebo or no intervention

Primary outcome

3.1 Sexual function ‐ composite score

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on the sexual function composite score (SMD 1.32, 95% CI 1.18 to 1.47; 1 study, 883 women; very low‐quality evidence; Analysis 3.1; Figure 6). This means that the sexual function composite score is compatible with a large effect of 1.32 standard units with the use of synthetic steroids versus placebo or no intervention. Due to the limited studies available, it was not possible to perform sensitivity analyses.

3.1. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 1: Sexual function composite score

6.

Forest plot of comparison: 3 Synthetic steroids versus placebo/no treatment, outcome: 3.1 Sexual function composite score.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on the sexual function composite score (SMD 0.46, 95% CI 0.07 to 0.86; 1 study, 105 women; very low‐quality evidence; Analysis 3.1; Figure 6). This means that the sexual function composite score is compatible with a small effect of 0.46 standard units with the use of synthetic steroids versus placebo or no intervention. Sensitivity analysis was not possible due to the suspected risk of bias in individual studies.

There was considerable heterogeneity (I² = 94%) when comparing these two subgroups and so pooled analysis was not appropriate (Figure 6). The investigation of additional sources of bias for the high heterogeneity observed in the pooled analysis (I² = 94%), is related to the sources of clinical, and statistical heterogeneity explored in the subgroup of unselected postmenopausal women.

Secondary outcomes

3.2 Desire at 12‐month follow‐up

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on desire (SMD 1.19, 95% CI 0.62 to 1.75; 1 study, 57 women; very low‐quality evidence; Analysis 3.2). This is compatible with a large effect of 1.19 standard units with the use of synthetic steroids.

3.2. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 2: Desire at 12‐month follow‐up

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on desire (SMD 0.19, 95% CI −0.44 to 0.82; 1 study, 39 women; very low‐quality evidence; Analysis 3.2). This is compatible with a small effect of 0.19 standard units with the use of synthetic steroids.

3.3 Desire at 24‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on desire (MD 0.35, 95% CI 0.21 to 0.49; 1 study, 105 women; very low‐quality evidence; Analysis 3.3). This is compatible with a moderate effect of 0.35 standard units with the use of synthetic steroids.

3.3. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 3: Desire at 24‐month follow‐up

3.4 Arousal at 12‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on arousal (MD 0.19, 95% CI −0.43 to 0.81; 1 study, 39 women; very low‐quality evidence; Analysis 3.4). This is compatible with a small effect of 0.19 standard units with the use of synthetic steroids.

3.4. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 4: Arousal at 12‐month follow‐up

3.5 Arousal at 24‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on arousal (MD 0.87, 95% CI 0.47 to 1.27; 1 study, 105 women; very low‐quality evidence; Analysis 3.5). This is compatible with a large effect of 0.87 standard units with the use of synthetic steroids.

3.5. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 5: Arousal at 24‐month follow‐up

3.6 Lubrication at 12‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on lubrication (MD 0.56, 95% CI 0.04 to 1.08; 1 study, 39 women; very low‐quality evidence; Analysis 3.6). This is compatible with a moderate effect of 0.56 standard units with the use of synthetic steroids.

3.6. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 6: Lubrication at 12‐month follow‐up

3.7 Lubrication at 24‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on lubrication (MD 0.62, 95% CI 0.28 to 0.96; 1 study, 105 women; very low‐quality evidence; Analysis 3.7). This is compatible with a moderate effect of 0.62 standard units with the use of synthetic steroids.

3.7. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 7: Lubrication at 24‐month follow‐up

3.8 Orgasm at 12‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on orgasm (MD 0.25, 95% CI −0.98 to 1.48; 1 study, 39 women; very low‐quality evidence; Analysis 3.8). This is compatible with a small effect of 0.25 standard units with the use of synthetic steroids.

3.8. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 8: Orgasm at 12‐month follow‐up

3.9 Orgasm at 24‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on orgasm (MD 0.31, 95% CI −0.19 to 0.81; 1 study, 105 women; very low‐quality evidence; Analysis 3.9). This is compatible with a small effect of 0.31 standard units with the use of synthetic steroids.

3.9. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 9: Orgasm at 24‐month follow‐up

3.10 Satisfaction

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on satisfaction (MD 0.51, 95% CI 0.12 to 0.90; 1 study, 105 women; very low‐quality evidence; Analysis 3.10). This is compatible with a moderate effect of 0.51 standard units with the use of synthetic steroids.

3.10. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 10: Satisfaction

3.11 Pain at 12‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on pain (MD 0.12, 95% CI −0.18 to 0.42; 1 study, 39 women; very low‐quality evidence; Analysis 3.11). This is compatible with a small effect of 0.12 standard units with the use of synthetic steroids.

3.11. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 11: Pain at 12‐month follow‐up

3.12 Pain at 24‐month follow‐up

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of synthetic steroids versus placebo or no intervention on pain (MD 0.90, 95% CI 0.01 to 1.79; 1 study, 105 women; very low‐quality evidence; Analysis 3.12). This is compatible with a large effect of 0.90 standard units with the use of synthetic steroids.

3.12. Analysis.

Comparison 3: Synthetic steroids versus placebo/no treatment, Outcome 12: Pain at 24‐month follow‐up

4. Comparison of selective estrogen receptor modulators (SERMs) versus placebo or no intervention

Primary outcome

4.1 Sexual function composite score

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs versus placebo or no intervention on the sexual function composite score (MD −1.00, 95% CI −2.00 to −0.00; 1 study, 215 women; low‐quality evidence; Analysis 4.1; Figure 7). This means that the sexual function composite score is compatible with a large effect of 1.0 standard unit with the use of SERMs. Due to the limited studies available, it was not possible to perform sensitivity analyses.

4.1. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 1: Sexual function composite score

7.

Forest plot of comparison: 4 Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, outcome: 4.1 Sexual function composite score.

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention may slightly improve the sexual function composite score (MD 2.24, 95% CI 1.37 to 3.11; 2 studies, 1525 women; low‐quality evidence; Analysis 4.1; Figure 7). This means that the sexual function composite score is compatible with a small effect of 0.25 standard units with the use of SERMs versus placebo or no intervention. Due to the limited studies available, it was not possible to perform sensitivity analysis.

Considerable heterogeneity was observed (I² = 92%) when comparing these two subgroups and so pooled analysis was not appropriate (Figure 7).

Secondary outcomes

4.2 Desire

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention probably slightly improve desire (MD 0.25, 95% CI 0.14 to 0.36; I² = 31%; 2 studies, 1545 women; low‐quality evidence; Analysis 4.2). This is compatible with a small effect of 0.25 standard units with the use of SERMs.

4.2. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 2: Desire

4.3 Arousal

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs versus placebo or no intervention on arousal (MD −0.22, 95% CI −0.44 to −00.0; 1 study, 215 women; very low‐quality evidence; Analysis 4.3). This is compatible with a small effect of 0.22 standard units with the use of SERMs.

4.3. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 3: Arousal

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention probably slightly improve arousal (MD 0.28, 95% CI 0.11 to 0.46; I² = 0%; 2 studies, 1543 women; low‐quality evidence; Analysis 4.3). This is compatible with a small effect of 0.28 standard units with the use of SERMs.

4.4 Lubrication

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs versus placebo or no intervention on lubrication (MD −0.17, 95% CI −1.44 to 1.10, 1 study, 215 women, very low‐quality evidence; Analysis 4.4). This is compatible with a small effect of 0.17 standard units with the use of SERMs.

4.4. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 4: Lubrication

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention probably slightly improve lubrication (MD 0.62, 95% CI 0.42 to 0.82; I² = 64%; 2 studies, 1546 women; low‐quality evidence; Analysis 4.4). This is compatible with a small effect of 0.62 standard units with the use of SERMs. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 64%) could be due to:

• clinical heterogeneity related to variability in the participants (Archer 2019 included women with postmenopausal symptoms with an intact uterus; Constantine 2015 included women with symptoms of vaginal atrophy or a bothersome symptom of dyspareunia); and

• methodological heterogeneity related to variability in the risk of bias (Archer 2019 high risk of bias; Constantine 2015, moderate risk of bias).

4.5 Orgasm

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs versus placebo or no intervention on orgasm (MD −0.38, 95% CI −0.66 to −0.10; 1 study, 215 women; very low‐quality evidence; Analysis 4.5). This is compatible with a small effect of 0.38 standard units with the use of SERMs.

4.5. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 5: Orgasm

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention probably slightly improve orgasm (MD 0.35, 95% CI 0.15 to 0.55; I² = 55%; 2 studies, 1546 women; very low‐quality evidence; Analysis 4.5). This is compatible with a large effect of 0.35 standard units with the use of SERMs. Investigation of additional sources of bias suggests that the high heterogeneity observed in this subgroup (I² = 55%) could be due to:

• clinical heterogeneity related to variability in the participants (Archer 2019 included women with postmenopausal symptoms with an intact uterus; Constantine 2015 included women with symptoms of vaginal atrophy or a bothersome symptom of dyspareunia); and

• methodological heterogeneity related to variability in the risk of bias (Archer 2019, high risk of bias; Constantine 2015, moderate risk of bias).

4.6 Satisfaction

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs versus placebo or no intervention on satisfaction (MD −0.28, 95% CI −0.56 to −0.00; 1 study, 215 women; very low‐quality evidence; Analysis 4.6). This is compatible with a small effect of 0.28 standard units with the use of SERMs.

4.6. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 6: Satisfaction

In the subgroup of unselected postmenopausal women, SERMs versus placebo or no intervention probably slightly improve satisfaction (MD 0.23, 95% CI 0.07 to 0.39; I² = 0.0%; 2 studies, 1527 women; low‐quality evidence; Analysis 4.6). This is compatible with a small effect of 0.23 standard units with the use of SERMs.

4.7 Pain

No study reported this outcome in the subgroup of symptomatic or early postmenopausal women.

In the subgroup of unselected postmenopausal women, the effect of SERMs versus placebo or no intervention probably slightly improve pain (MD 0.53, 95% CI 0.33 to 0.74; I² = 0%, 2 studies, 1543 women; low‐quality evidence; Analysis 4.7). This is compatible with a moderate effect of 0.53 standard units with the use of SERMs.

4.7. Analysis.

Comparison 4: Selective estrogen receptor modulators (SERMs) versus placebo/no treatment, Outcome 7: Pain

5. Comparison of SERMs combined with estrogen versus placebo or no intervention

Primary outcome

5.1 Sexual function composite score

In the subgroup of symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on the sexual function composite score (SMD 0.22, 95% CI 0.00 to 0.43; 1 study, 542 women; very low‐quality evidence; Analysis 5.1, Figure 8). This means that the sexual function composite score is compatible with a small effect of 0.22 standard units with the use of SERMs combined with estrogen. Due to the limited studies available, it was not possible to perform sensitivity analyses.

5.1. Analysis.

Comparison 5: Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, Outcome 1: Sexual function composite score

8.

Forest plot of comparison: 5 Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, outcome: 5.1 Sexual function composite score.

In the subgroup of unselected postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on the sexual function composite score (SMD 2.79, 95% CI 2.41 to 3.18; 1 study, 272 women; very low‐quality evidence; Analysis 5.1, Figure 8). This means that the sexual function composite score is compatible with a large effect of 2.79 standard units with the use of SERMs combined with estrogen versus placebo or no intervention. Due to the limited studies available, it was not possible to perform sensitivity analyses.

We observed considerable heterogeneity (I² = 99.2%) when comparing these two subgroups and so pooled analysis was not appropriate (Figure 8).

Secondary outcomes

5.2 Desire

No study reported this outcome.

5.3 Arousal

In the subgroup symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on arousal (MD 0.07, 95% CI 0.03 to 0.10; 1 study, 542 women; very low‐quality evidence; Analysis 5.3). This is compatible with a small effect of 0.07 standard units with the use of SERMs combined with estrogen.

5.3. Analysis.

Comparison 5: Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, Outcome 3: Arousal

No study reported this outcome in unselected postmenopausal women.

5.4 Lubrication

In the subgroup symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on lubrication (MD −0.33, 95% CI −0.51 to −0.16; 1 study, 542 women; very low‐quality evidence; Analysis 5.4). This is compatible with a small effect of 0.33 standard units with the use of SERMs combined with estrogen.

5.4. Analysis.

Comparison 5: Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, Outcome 4: Lubrication

No study reported this outcome in unselected postmenopausal women.

5.5 Orgasm

In the subgroup symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on orgasm (MD −0.05, 95% CI −0.42 to 0.32; 1 study, 542 women; very low‐quality evidence; Analysis 5.5). This is compatible with a small effect of 0.05 standard units with the use of SERMs combined with estrogen.

5.5. Analysis.

Comparison 5: Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, Outcome 5: Orgasm

No study reported this outcome in unselected postmenopausal women.

5.6 Satisfaction

In the subgroup symptomatic or early postmenopausal women, we are uncertain of the effect of SERMs combined with estrogen versus placebo or no intervention on satisfaction (MD −0.02, 95% CI −0.27 to 0.22; 1 study, 542 women; very low‐quality evidence; Analysis 5.6). This is compatible with a small effect of 0.02 standard units with the use of SERMs combined with estrogen.

5.6. Analysis.

Comparison 5: Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment, Outcome 6: Satisfaction

No study reported this outcome in unselected postmenopausal women.

5.7 Pain

No study reported this outcome.

Comments

We could not use a funnel plot to assess publication bias because no analysis included 10 or more studies.

Discussion

Summary of main results

We observed from the included studies that estrogen alone (moderate‐quality evidence) probably promotes a slight improvement in the sexual function composite score in women with menopause‐related symptoms (e.g. hot flushes, night sweats, vaginal dryness) or during the early postmenopausal period, or both. We are uncertain whether synthetic steroids (very low‐quality evidence), SERMs (very low‐quality evidence), and SERMs combined with estrogen (very low‐quality evidence) improve sexual function in these women (Table 1).

In the subgroup of unselected postmenopausal women, estrogen alone probably makes little or no difference to the sexual function composite score (moderate‐quality evidence). Findings for estrogen combined with progestogens (very low‐quality evidence) and synthetic steroids (very low‐quality evidence) showed that it is uncertain whether these treatments improve sexual function. Findings for SERMs versus placebo (very low‐quality evidence) showed that this treatment probably improves sexual function. For SERMs combined with estrogen (very low‐quality evidence), it is uncertain whether this hormone therapy improves sexual function (Table 2).

When examining the effect of estrogen alone on the specific sexual domains (sexual desire, arousal, lubrication, orgasm, satisfaction, and pain), we observed that only lubrication, pain, and satisfaction may slightly improve in the subgroup of symptomatic or early postmenopausal women (low‐quality evidence). In the subgroup of unselected postmenopausal women (moderate‐quality evidence), estrogen alone probably slightly improves lubrication. Regarding the effect of estrogen combined with progestogens, this intervention may slightly improve only lubrication in unselected postmenopausal women. SERMs (low‐quality evidence) probably slightly improve desire, arousal, lubrication, orgasm, satisfaction, and pain in the subgroup of unselected women.

Women considering the use of hormone therapy for other reasons may consider the observed effect on sexual function in the decision‐making process. Sexual complaints progressively increase with advancing age. The prospective observational Melbourne Women's Midlife Health Project annually assessed the sexual function of Australian women aged 45 years to 55 years for an eight‐year period. In this population, there was an increase in the prevalence of sexual complaints from 42% to 88% in this period, correlated with decreasing levels of estradiol (Dennerstein 2003)

The progressive decrease in sex hormones during the climacteric period until menopause can negatively affect sexual function (Ali 2020). This condition affects the ability to achieve vaginal lubrication due to atrophic changes in the vaginal mucosa that reduce vaginal lubrication and, therefore, generate a greater occurrence of dyspareunia and consequently, cause higher rates of sexual dissatisfaction in the pre‐ and postmenopausal period (da Silva Lara 2009a; Lara 2009; Simon 2011). It is also known that estrogen modulates the synthesis, action, and expression of receptors in the central nervous system such as nitric oxide and dopamine (Jacobs 2011), which modulate sexual response. The hypoestrogenism that occurs with the onset of menopause increases the risk for sexual dysfunction (Santoro 2016), such as reduced sexual desire, orgasmic disorder, and dyspareunia (Avis 2017).

The Global Study of Sexual Attitudes and Behaviors was carried out in 29 countries in a sample of 13,882 women ranging from 53 years to 58 years old. The most prevalent sexual dysfunction in this sample was lack of sexual interest, followed by the inability to reach orgasm, and poor lubrication (Nicolosi 2004). The Study of Women’s Health Across the Nation (SWAN) showed that 33% of women avoid sex due to lack of interest. However, regardless of age and menopausal status, sexual interest remained extremely important for a third of middle‐aged women (Cain 2003), reinforcing that sexual issues should be addressed by physicians in this period.

In the menopausal transition, sexual dysfunction is multifactorial and may result from several aetiologies including psychosocial factors (Yazdanpanahi 2018), chronic illness (Tabatabaeichehr 2018), medication side effects, and sexual partner problems (Lonnèe‐Hoffmann 2014), and hypoestrogenism in the menopausal transition. The exact mechanism by which hypoestrogenism affects female sexual dysfunction, particularly sexual desire disorder is unknown.

The mechanism for sexual pain involves a decline in gonadal steroid hormone levels leading to genitourinary menopause syndrome (SGUM; dryness, burning, irritation, dysuria, sexual pain; Kim 2015) resulting in sexual discomfort and pain during penis‐vagina intercourse (Portman 2014b). The structural changes in the vaginal wall due to hypoestrogenism include a reduction in the number of epithelial layers, greater deposition of short‐chain collagen, which increases tissue resistance, and increased apoptosis leading to wall fibrosis making the vagina tighter, which favours dyspareunia (Lara 2009). This in turn can lead to hypoactive sexual desire disorder (Tabatabaeichehr 2018).

The use of estrogen can mitigate the effects of hypoestrogenism on sexual function in women with menopause‐related symptoms or during the early postmenopausal period, or both (Other published versions of this review). However, before this intervention, it is crucial first to establish the role of other factors such as psychosocial and interpersonal factors, medication use, and concomitant illnesses in the sexual function of menopausal women. In addition, it is important to evaluate the sexual response as a whole (Basson 2000). In addition, it is crucial to use only appropriate terminologies related to sexual health issues, as sexual response should be evaluated in all phases including sexual satisfaction and distress (Basson 2000).

Currently, short‐term hormone therapy is recommended to alleviate menopausal symptoms and SGUM, and has been shown to prevent bone loss and fracture (NAMS 2017). Local estrogen is effective to improve SGUM symptoms (Biehl 2019). A recent study showed that ultra‐low‐dose (0.03 mg) vaginal estriol in postmenopausal women was effective to improve sexual function in women who underwent surgical treatment for stress urinary incontinence (Caruso 2020), probably due to the beneficial effect of estriol on SGUM symptoms in these women. The available data do not support systemic estrogen therapy for the treatment of female sexual dysfunction (Santoro 2016). However, systemic hormone therapy and low‐dose vaginal estrogen provide improvement of sexual problems by increasing lubrication, blood flow, and sensation in vaginal tissues (NAMS 2017).

Our findings demonstrate that hormone therapy treatment with estrogen alone is associated with a small improvement in sexual function when used in women with menopausal symptoms or early postmenopause (within five years of amenorrhoea). In addition, hormone therapy is associated with a small improvement in the domains of sexual response (sexual arousal, orgasm, and satisfaction) when used in women with menopausal symptoms or early postmenopausal women, as well as in the pain domain in unselected women. We found that estrogen alone is more effective than estrogen combined with progestogen to improve sexual function. This is in agreement with a previous study that demonstrated that sexual motivation is positively related to the level of salivary estradiol and negatively related to progesterone in women (Roney 2013). These findings should be considered as an added value to the use of hormone therapy. However, the evidence is still low quality to consider hormone therapy as a formal indication to improve sexual complaints in perimenopausal and postmenopausal women.

Finally, hormone therapy is not currently recommended for the treatment of female sexual dysfunction. However, when sexual complaints appear with the onset of menopause, not being associated with other causes of sexual dysfunction, women may benefit from hormone therapy. When the sexual complaint is limited to pain, topical estrogen works well with minimal absorption to reduce this symptom (Lethaby 2016), as confirmed by our findings, which demonstrated that estrogen and SERMs may decrease sexual pain.

Overall completeness and applicability of evidence

The review included 36 studies. Eighteen received funding from industry (Archer 2019; Bushmakin 2014; Fonseca 2007; Gast 2009; Haines 2009; Hemminki 2009; Huang 2008; Hudita 2003; Laan 2001; Liberate 2011; Maki 2007; Simon 2007; SMART‐3; Speroff 2003; Strickler 2000; Veerus 2006; WHI 1999; WISDOM 2008). Sixteen did not have data suitable for meta‐analysis (Czarnecka 2000; Fonseca 2007; Gambacciani 2011; Gast 2009; Hemminki 2009; Huang 2008; Hudita 2003; Mameri Filho 2005; Modugno 2003; Nielsen 2006; Simon 2007; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008). Nine of them compared conjugated equine estrogen with placebo (Gast 2009; Hemminki 2009; Mameri Filho 2005; Modugno 2003; Strickler 2000; Taylor 2017; Veerus 2006; WHI 1999; WISDOM 2008); six compared estrogen alone with placebo (Huang 2008; Mameri Filho 2005; Nielsen 2006; Simon 2007; Strickler 2000; Taylor 2017); six compared estrogen in combination with medroxyprogesterone acetate with placebo (Gast 2009; Hemminki 2009; Modugno 2003; Veerus 2006; WHI 1999; WISDOM 2008); three compared estrogen in combination with progestogens with placebo (Czarnecka 2000; Fonseca 2007; Gambacciani 2011); one compared synthetic steroids (tibolone) with placebo (Hudita 2003); and one compared SERMs with placebo (Modugno 2003). In seven studies the use of hormone therapy did not improve sexual function (Hemminki 2009; Huang 2008; Mameri Filho 2005; Modugno 2003; Strickler 2000; Veerus 2006; WISDOM 2008).

We were able to draw some conclusions on the effect of hormone therapy on the sexual function of women. One relevant finding of this review was that while we included 36 studies, 35 included studies evaluated postmenopausal women with at least six months of amenorrhoea and elevated follicle‐stimulating hormone (FSH). Only one study included perimenopausal women, using tibolone 2.5 mg, and provided data for the meta‐analysis (Morais‐Socorro 2012). So the conclusions are mostly based on postmenopausal women. Other relevant findings are that the studies differed in the type of participants, drugs used, dosages, administration routes, and the tools used to assess sexual function, which explains the observed heterogeneity. In addition, 20 of the 36 studies included in this review were at high risk of bias, 14 were at moderate risk, and only two were graded as low risk of bias. Thus, it is relevant to take all this information into account when deciding to prescribe hormone therapy in order to improve the sexual function of the woman.

Quality of the evidence

Of the 36 studies included in this review, 20 are at high risk of bias for at least one risk of bias domain, 14 are at unclear risk of bias for at least one risk of bias domain, and two are low risk of bias in all risk of bias domains. This degree of bias was responsible for reducing the quality and imprecision of the evidence for many comparisons. Many included studies were supported by the pharmaceutical industry, in most cases the manufacturer of the studied drug (see Characteristics of included studies). The quality of the evidence for specific comparisons was as follows.

Symptomatic and early postmenopausal women ‐ sexual function composite score

See Table 1.

We considered the evidence on the effect of estrogen alone to be of moderate quality because we assessed most risk of bias domains as 'unclear'.

We considered the evidence on the effect of synthetic steroids (tibolone) to be of very low quality because the sample size was small with a low event rate or an effect estimate with a wide confidence interval or both, or because random allocation was inadequate or absent.

We considered the evidence on the effect of SERMs to be of very low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both.

We considered the evidence on the effect of SERMs combined with estrogen to be of very low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both.

Unselected postmenopausal women ‐ sexual function composite score

See Table 2.

We considered the evidence on the effect of estrogen alone to be of moderate quality because we assessed most risk of bias domains as 'unclear'.

We considered the evidence on the effect of estrogen combined with progestogens to be of very low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both.

We considered the evidence on the effect of synthetic steroids (tibolone) to be of very low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both, or because of inadequate randomization, allocation concealment or no evidence of blinding.

We considered the evidence on the effect of SERMs to be of low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both.

We considered the evidence on the effect of SERMs combined with estrogen to be of very low quality because we assessed most risk of bias domains as 'unclear' or because the sample size was small with a low event rate or an effect estimate with a wide confidence interval, or both.

Potential biases in the review process

We undertook an extensive electronic and manual search and carefully evaluated the study reports for eligibility. Some studies assessed sexual function as a secondary outcome, which could impair the identification of all relevant studies. During data extraction, we made many arbitrary decisions. Each decision was reasoned by the authors, and we tried to take the most conservative decision. We contacted the authors of the majority of the included studies.

Agreements and disagreements with other studies or reviews

Systematic reviews and scientific institutions' statements are controversial regarding the use of estrogen to improve sexual function. Some reviews conclude that some types of estrogen therapy could improve some domains of sexual function (Alexander 2004; McCabe 2016b; Nappi 2009b), and some conclude they do not (NAMS 2012; Santen 2010). We observed a small benefit of hormone therapy with estrogen alone on the composite scores for sexual function. Regarding the specific domains, we observed that estrogen probably slightly improves lubrication, pain in unselected women and in women with menopause‐related symptoms, and satisfaction in symptomatic or early postmenopausal women. In unselected postmenopausal women, estrogen combined with progestogens may slightly improve lubrication, but we are uncertain of the effect of this treatment on desire, pain, orgasm, satisfaction, and arousal. SERMs may slightly improve pain in unselected postmenopausal women. Regarding synthetic steroids, other reviews agree that evidence of the effect of tibolone on sexual function is scarce (Kenemans 2005; Santen 2010).

Authors' conclusions

Implications for practice.

We included nine new studies in this update. This review brought further information on the effect of estrogen and the use of synthetic steroids and selective estrogen receptor modulators (SERMS) on female sexual function. The previous review showed that estrogen alone was associated with a small to moderate improvement in sexual function, particularly in pain in women with menopausal symptoms, or in early postmenopause, but not in unselected women. Findings for this review showed that estrogen alone probably promotes a slight improvement in the sexual function composite score in women with menopause‐related symptoms (e.g. hot flushes, night sweats, vaginal dryness) and in unselected women (moderate‐quality evidence). The domains of lubrication, pain, and satisfaction were improved in women with menopause‐related symptoms. It is uncertain whether estrogen combined with progestogen, synthetic steroids, and SERMS combined with estrogen increase sexual function due to insufficient data, poor quality of the studies, and serious risk of bias (small sample size, low event rate, or effect estimate with a wide confidence interval, or both, inadequate randomization, or inadequate blinding). This new information on the effect of estrogen therapy and also on the use of synthetic steroids and SERMS on female sexual function is possibly due to the use of validated questionnaires in most studies. These instruments are composed of domains of sexual function and do not use inaccurate denominations (e.g. libido) that may impair the evaluation of sexual function.

Implications for research.

Questions still remain on the true effect of hormone therapy in sexual function in perimenopausal and postmenopausal women. Further randomized controlled trials (RCTs) should focus on estrogen in combination with progestogens, synthetic steroids, SERMs alone, or in combination with estrogen. Most studies were at high risk of bias. Future clinical trials require a follow‐up period of at least six months to assess the effect of hormone therapy on sexual function during the climacteric period and menopause. There is a need for well‐designed randomized controlled trials to evaluate the real effect of hormone therapy on women's sexual function. Furthermore, future trials should endeavour to report the effect of hormone therapy when used specifically in women with sexual complaints. Studies including perimenopausal women (who had their last menstrual period within 12 months) are limited and further evidence is needed to clarify the real effect of hormone therapy on sexual function in this group of women.

What's new

Date	Event	Description
24 August 2023	New citation required but conclusions have not changed	The addition of nine new studies has not led to a change in the conclusions of this review.
24 August 2023	New search has been performed	The date of the last search was 19 December 2022. The text has been revised. Nine new studies added at this update: Archer 2019; Bumphenkiatikul 2020; Bushmakin 2014; Caruso 2020; Constantine 2015; Diem 2018; Mitchell 2018; Seyyedi 2016; Taylor 2017.

History

Protocol first published: Issue 2, 2012
Review first published: Issue 6, 2013

Appendices

Appendix 1. Cochrane Gynaecology and Fertility (CGF) Group Specialised Register search strategy

ProCite platform

Searched 19 December 2022

Keywords CONTAINS "menopausal" or "*Menopause" or "perimenopausal" or "perimenopause" or "Postmenopausal" or "postmenopause" or "climacteric " or Title CONTAINS "menopausal" or "*Menopause" or "perimenopausal" or "perimenopause" or "Postmenopausal" or "postmenopause" or "climacteric "

AND

Keywords CONTAINS "Hormone Replacement Therapy" or "*Hormone Substitution" or "hormone therapy" or "hormone therapy estrogen" or "estraderm" or "estradiol" or "Estriol" or "estrofen" or "estrogen" or "estrogen‐progestogen" or "*Estrogens" or "Estrone" or "estroprogestin" or "ethinyl‐estradiol" or "ethinyl estradiol‐cyproterone acetate" or "ethinyl estradiol + drospirenone" or "oestrodiol" or "oestrogen" or "conjugated equine " or "conjugated estrogen" or "conjugated estrogens" or "CEE" or "HRT" or "HT " or "premarin" or "Prempro" or "SERM" or "selective estrogen receptor modulator" or "*Clomiphene" or "raloxifene" or "toremifen" or "toremifene" or "bazedoxifene acetate" or "Evista" or "tamoxifen" or "tibolone" or "ospemifene" or Title CONTAINS "SERM" or "selective estrogen receptor modulator" or "*Clomiphene" or "raloxifene" or "toremifen" or "toremifene" or "bazedoxifene acetate" or "Evista" or "tamoxifen" or "Hormone Replacement Therapy" or "*Hormone Substitution" or "hormone therapy" or "ospemifene" or "tibolone"

AND

Keywords CONTAINS "sex drive" or "sexual" or "*Sexuality" or "dyspareunia" or "hypoactive sexual desire" or "hypoactive sexual desire disorder" or "orgasm" or "vaginal acidification" or "vaginal atrophy" or "vaginal dryness" or "vaginal lubrication" or "vaginal matuation "or "vaginal symptoms" or"*Vaginitis" or "vaginosis" or "libido" or "Vulvar Atrophy" or "vulvo‐vaginal symptoms" or "Vulvovaginal atrophy" or "vulvodynia "or "sexual arousal" or "sexual activity" or "sexual arousal disorder" or "sexual complain" or "sexual desire" or "sexual desire disorder" or "sexual dysfunction" or "sexual function" or "Sexual functioning" or "sexual interest" or "sexual life quality" or "sexual problems" or "sexual response" or "sexual satisfaction" or "sexual score" or "sexual wellbeing"

527 records

Appendix 2. CENTRAL via the Cochrane Register of Studies Online (CRSO) search strategy

Web platform

Searched 19 December 2022
#1 MESH DESCRIPTOR Climacteric EXPLODE ALL TREES 7550
#2 (climacter* or menopaus* or perimenopaus* or post?menopaus*):TI,AB,KY 26859
#3 #1 OR #2 27399
#4 MESH DESCRIPTOR Hormone Replacement Therapy EXPLODE ALL TREES 2985
#5 (HT or HRT):TI,AB,KY 5946
#6 (estrogen* or oestrogen*):TI,AB,KY 14478
#7 (estradiol* or oestradiol*):TI,AB,KY 11638
#8 (estrone* or estriol* or CEE):TI,AB,KY 1581
#9 (Dienestrol or estriol or estetrol):TI,AB,KY 490
#10 (premarin or prempro):TI,AB,KY 217
#11 MESH DESCRIPTOR Estrogens EXPLODE ALL TREES 7975
#12 MESH DESCRIPTOR Selective Estrogen Receptor Modulators EXPLODE ALL TREES 3067
#13 (SERMS or SERM):TI,AB,KY 284
#14 (selective estrogen receptor modulator*):TI,AB,KY 767
#15 (clomiphene or clomid):TI,AB,KY 1808
#16 (clomifene or clomifert):TI,AB,KY 336
#17 tamoxifen:TI,AB,KY 4939
#18 (serophene or raloxifene):TI,AB,KY 965
#19 (ly?139481 or ly?156758):TI,AB,KY 1
#20 evista:TI,AB,KY 48
#21 (dimethylamine* or novaldex):TI,AB,KY 60
#22 zitazonium:TI,AB,KY 1
#23 MF?101:TI,AB,KY 8
#24 femarelle:TI,AB,KY 3
#25 DT56a:TI,AB,KY 4
#26 toremifene:TI,AB,KY 193
#27 (bazedoxifene or ospemifene):TI,AB,KY 262
#28 tibolone:TI,AB,KY 541
#29 #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21 OR #22 OR #23 OR #24 OR #25 OR #26 OR #27 OR #28 32355
#30 #3 AND #29 12139
#31 MESH DESCRIPTOR Sexual Dysfunction, Physiological EXPLODE ALL TREES 2366
#32 MESH DESCRIPTOR Sexual Dysfunctions, Psychological EXPLODE ALL TREES 2233
#33 sex*:TI,AB,KY 71835
#34 dyspareunia:TI,AB,KY 1215
#35 (?vagina* adj2 atroph*):TI,AB,KY 560
#36 (?vagina* adj3 dry*):TI,AB,KY 608
#37 (orgasm* or libido):TI,AB,KY 1829
#38 (frigid* or arousal*):TI,AB,KY 7596
#39 HSDD:TI,AB,KY 177
#40 (hypoactive adj2 desire*):TI,AB,KY 299
#41 MESH DESCRIPTOR Libido EXPLODE ALL TREES 321
#42 psychosexual*:TI,AB,KY 218
#43 #31 OR #32 OR #33 OR #34 OR #35 OR #36 OR #37 OR #38 OR #39 OR #40 OR #41 OR #42 80727
#44 #30 AND #43 1758

Appendix 3. MEDLINE search strategy

Ovid platform

Searched from 1946 to 19 January 2022

1 exp climacteric/ or exp menopause/ or exp menopause, premature/ or exp perimenopause/ or exp postmenopause/ (64824)
2 climacter$.tw. (5143)
3 menopaus$.tw. (53491)
4 perimenopaus$.tw. (4967)
5 post?menopaus$.tw. (60052)
6 or/1‐5 (118189)
7 exp hormone replacement therapy/ or exp estrogen replacement therapy/ (26245)
8 HRT.tw. (11260)
9 (hormon$ adj2 therap$).tw. (43608)
10 exp estrogens/ or exp epimestrol/ or exp estradiol/ or exp estrogenic steroids, alkylated/ or exp "estrogens, conjugated (usp)"/ or exp "estrogens, esterified (usp)"/ or exp estrone/ or exp ethinyl estradiol/ or exp mestranol/ or exp quinestrol/ (169730)
11 (estrogen$ or oestrogen$).tw. (171042)
12 estradiol$.tw. (85493)
13 oestradiol$.tw. (13587)
14 CEE.tw. (1391)
15 (estrone$ or estriol$).tw. (12481)
16 exp Dienestrol/ (187)
17 Dienestrol$.tw. (231)
18 exp estriol/ or exp estetrol/ or exp ethinyl estradiol‐norgestrel combination/ (6628)
19 (estriol or estetrol).tw. (4563)
20 premarin.tw. (527)
21 prempro.tw. (29)
22 exp selective estrogen receptor modulators/ or exp clomiphene/ or exp raloxifene/ or exp tamoxifen/ or exp toremifene/ (29414)
23 SERMS.tw. (1577)
24 selective estrogen receptor modulator$.tw. (3279)
25 Ospemifene.tw. (204)
26 (clomiphene or clomid).tw. (5422)
27 (clomifene or clomifert).tw. (147)
28 serophene.tw. (4)
29 raloxifene.tw. (3474)
30 keoxifene.tw. (49)
31 ly?139481.tw. (8)
32 ly?156758.tw. (32)
33 evista.tw. (67)
34 tamoxifen.tw. (24510)
35 dimethylamine$.tw. (2001)
36 novaldex.tw. (1)
37 tomaxithen.tw. (0)
38 soltamox.tw. (0)
39 zitazonium.tw. (6)
40 menerba.tw. (0)
41 MF?101.tw. (10)
42 femarelle.tw. (13)
43 DT56a.tw. (17)
44 LY?117018.tw. (82)
45 toremifene.tw. (699)
46 serm.tw. (1341)
47 bazedoxifene.tw. (438)
48 tibolone.tw. (1031)
49 or/7‐48 (350728)
50 6 and 49 (51133)
51 exp sexual dysfunction, physiological/ or exp dyspareunia/ (32912)
52 exp Sexual Dysfunctions, Psychological/ (29593)
53 sex$.tw. (858533)
54 dyspareunia.tw. (4456)
55 (vagina$ adj2 atroph$).tw. (771)
56 (vulva$ adj2 atroph$).tw. (80)
57 (vulvo$ adj2 atroph$).tw. (418)
58 (vagina$ adj2 dry$).tw. (1197)
59 orgasm.tw. (3339)
60 libido.tw. (4224)
61 orgasmic.tw. (1210)
62 frigid$.tw. (1003)
63 arousal.tw. (31381)
64 HSDD.tw. (315)
65 (hypoactive adj2 desire).tw. (657)
66 exp Libido/ (4994)
67 psychosexual$.tw. (2031)
68 or/51‐67 (909629)
69 randomized controlled trial.pt. (582929)
70 controlled clinical trial.pt. (95137)
71 randomized.ab. (576755)
72 placebo.tw. (237952)
73 clinical trials as topic.sh. (200654)
74 randomly.ab. (393184)
75 trial.ti. (271269)
76 (crossover or cross‐over or cross over).tw. (94919)
77 or/69‐76 (1510979)
78 exp animals/ not humans.sh. (5074784)
79 77 not 78 (1387646)
80 50 and 68 and 79 (1417)

Appendix 4. Embase search strategy

Ovid platform

Searched from 1980 to 19 January 2022

1 exp climacterium/ (8862)
2 exp menopause/ or exp early menopause/ or exp menopause related disorder/ or exp "menopause and climacterium"/ (149229)
3 exp postmenopause/ (73252)
4 climacter$.tw. (5276)
5 menopaus$.tw. (80877)
6 perimenopaus$.tw. (7260)
7 post?menopaus$.tw. (85922)
8 or/1‐7 (188612)
9 exp hormone substitution/ or exp hormonal therapy/ or exp substitution therapy/ or exp estrogen therapy/ (411894)
10 (hormon$ adj2 therap$).tw. (64008)
11 HRT.tw. (16283)
12 exp estrogen/ (272789)
13 exp estradiol/ or exp drospirenone plus estradiol/ or estradiol valerate/ or exp desogestrel plus estradiol/ or estradiol valerate plus medroxyprogesterone acetate/ or exp cyproterone acetate plus estradiol valerate/ or exp dydrogesterone plus estradiol/ or exp 17alpha estradiol/ or exp dienogest plus estradiol valerate/ (122369)
14 exp conjugated estrogen/ (9922)
15 (estrogen$ or oestrogen$).tw. (201745)
16 estradiol$.tw. (97338)
17 oestradiol$.tw. (12815)
18 CEE.tw. (1944)
19 (estrone$ or estriol$).tw. (11826)
20 exp dienestrol/ (543)
21 Dienestrol$.tw. (192)
22 exp estriol/ (6472)
23 exp estetrol/ (259)
24 exp ethinylestradiol/ (16833)
25 (estriol or estetrol).tw. (3977)
26 premarin.tw. (2934)
27 prempro.tw. (394)
28 exp selective estrogen receptor modulator/ (8538)
29 exp clomifene/ (4960)
30 exp raloxifene/ (12076)
31 exp tamoxifen/ (69084)
32 exp toremifene/ (2362)
33 (SERMS or SERM).tw. (3559)
34 selective estrogen receptor modulator$.tw. (4505)
35 Ospemifene.tw. (320)
36 (clomiphene or clomid).tw. (6808)
37 (clomifene or clomifert).tw. (245)
38 serophene.tw. (202)
39 raloxifene.tw. (4945)
40 keoxifene.tw. (57)
41 ly?139481.tw. (11)
42 ly?156758.tw. (31)
43 evista.tw. (1032)
44 (tamoxifen or toremifene).tw. (37443)
45 dimethylamine$.tw. (2138)
46 novaldex.tw. (118)
47 tomaxithen.tw. (0)
48 soltamox.tw. (19)
49 zitazonium.tw. (29)
50 menerba.tw. (5)
51 MF?101.tw. (20)
52 femarelle.tw. (23)
53 DT56a.tw. (24)
54 LY?117018.tw. (89)
55 bazedoxifene.tw. (729)
56 tibolone.tw. (1415)
57 or/9‐56 (793078)
58 exp sexual dysfunction/ (93080)
59 exp dyspareunia/ or exp female sexual dysfunction/ or exp pain/ (1548735)
60 dyspareunia.tw. (8557)
61 sexual$.tw. (328134)
62 (vagina$ adj2 atroph$).tw. (1408)
63 (vulva$ adj2 atroph$).tw. (124)
64 (vulvo$ adj2 atroph$).tw. (710)
65 (vagina$ adj2 dry$).tw. (2208)
66 orgasm.tw. (5844)
67 libido.tw. (6132)
68 orgasmic.tw. (2092)
69 exp libido disorder/ or exp libido/ (13718)
70 frigid$.tw. (818)
71 arousal.tw. (40139)
72 HSDD.tw. (702)
73 exp hypoactive sexual desire disorder/ or exp psychosexual disorder/ (23653)
74 psychosexual$.tw. (2923)
75 (hypoactive adj2 desire).tw. (1278)
76 or/58‐75 (1932729)
77 Clinical Trial/ (1039669)
78 Randomized Controlled Trial/ (736575)
79 exp randomization/ (95813)
80 Single Blind Procedure/ (48500)
81 Double Blind Procedure/ (198511)
82 Crossover Procedure/ (72218)
83 Placebo/ (375048)
84 Randomi?ed controlled trial$.tw. (302029)
85 Rct.tw. (49762)
86 random allocation.tw. (2405)
87 randomly allocated.tw. (42909)
88 allocated randomly.tw. (2831)
89 (allocated adj2 random).tw. (854)
90 Single blind$.tw. (29718)
91 Double blind$.tw. (229040)
92 ((treble or triple) adj blind$).tw. (1689)
93 placebo$.tw. (346574)
94 prospective study/ (814176)
95 or/77‐94 (2612946)
96 case study/ (90165)
97 case report.tw. (497427)
98 abstract report/ or letter/ (1214388)
99 or/96‐98 (1788047)
100 95 not 99 (2550726)
101 8 and 57 and 76 and 100 (4917)

Appendix 5. PsycINFO search strategy

Ovid platform

Searched from 1806 to 19 December 2022

1 exp menopause/ (4044)
2 climacter$.tw. (541)
3 menopaus$.tw. (5387)
4 perimenopaus$.tw. (767)
5 post?menopaus$.tw. (2816)
6 or/1‐5 (7492)
7 exp Hormone Therapy/ or exp Estrogens/ (8782)
8 (hormon$ adj2 therap$).tw. (2807)
9 HRT.tw. (675)
10 exp Estradiol/ (3320)
11 (estrogen$ or oestrogen$).tw. (9144)
12 (estradiol$ or oestradiol$).tw. (7022)
13 CEE.tw. (227)
14 (estrone$ or estriol$).tw. (256)
15 Dienestrol.tw. (0)
16 exp Estradiol/ (3320)
17 (estriol or estetrol).tw. (67)
18 premarin.tw. (43)
19 prempro.tw. (11)
20 selective estrogen receptor modulator$.tw. (187)
21 Ospemifene.tw. (9)
22 (clomiphene or clomid).tw. (57)
23 (clomifene or clomifert).tw. (0)
24 raloxifene.tw. (149)
25 tamoxifen.tw. (620)
26 toremifene.tw. (5)
27 (SERMS or SERM).tw. (113)
28 serophene.tw. (0)
29 bazedoxifene.tw. (9)
30 menerba.tw. (0)
31 MF?101.tw. (0)
32 tibolone.tw. (42)
33 or/7‐32 (16484)
34 exp sexual satisfaction/ or exp sexuality/ (19348)
35 sexual$.tw. (198107)
36 dyspareunia.tw. (634)
37 (vagina$ adj2 atroph$).tw. (57)
38 (vagina$ adj2 dry$).tw. (177)
39 orgasm$.tw. (3392)
40 libido.tw. (2895)
41 orgasmic.tw. (943)
42 frigid$.tw. (446)
43 arousal.tw. (34189)
44 HSDD.tw. (231)
45 (hypoactive adj2 desire).tw. (505)
46 exp Libido/ (885)
47 psychosexual$.tw. (3673)
48 or/34‐47 (231715)
49 6 and 33 and 48 (414)
50 random.tw. (67215)
51 control.tw. (495413)
52 double‐blind.tw. (24631)
53 clinical trials/ (12127)
54 placebo/ (6375)
55 exp Treatment/ (1169415)
56 or/50‐55 (1608917)
57 49 and 56 (268)

Appendix 6. CINAHL search strategy

Ebsco platform

Searched from 1961 to 12 January 2021

(menopaus*) AND (sexual) AND ((trial) OR (random*))

CINAHL (Dec‐12‐2012) = 51 records

‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Search strategy (May‐21‐2018)

(menopaus*) AND (sexual) AND ((trial) OR (random*))

CINAHL (May‐21‐2018) = 48 records

‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

Search strategy (Jan‐11‐2021)

(menopaus*) AND (sexual) AND ((trial) OR (random*))

CINAHL (May‐21‐2018) = 32 records

Appendix 7. LILACS search strategy

Web platform

Searched 11 January 2021

(menopaus$) AND (sexual) AND ((trial) OR (random$))

= 37 records

Appendix 8. ClinicalTrials.gov search strategy

Web platform

Searched 21 May 2018, more recent trial registrations were found in the CENTRAL 2020 output

(menopause) AND (sexual)

= 181 records

NOTE: The search was not carried out in 2021, because since 2021, the trials are included in CENTRAL

Appendix 9. Controlled‐Trials search strategy

Web platform

Searched 21 May 2018, more recent trial registrations were found in the CENTRAL 2020 output

(menopause) AND (sexual)

= 21 records

NOTE: The search was not carried out in 2021, because since 2021, the trials are included in CENTRAL

Appendix 10. WHO International Clinical Trials Registry Platform search strategy

Web platform

Searched 27 May 2018, more recent trial registrations were found in the CENTRAL 2020 output

(menopause) AND (sexual)

= 87 records

NOTE: The search was not carried out in 2021, because since 2021, the trials are included in CENTRAL

Appendix 11. Web of Science search strategy

Web platform

Searched 11 January 2021

(menopause) AND (sexual) AND ((trial) OR (random$))

= 457 records

Data and analyses

Comparison 1. Estrogen alone versus placebo/no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Sexual function composite score	9	1307	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.16, 1.04]
1.1.1 Symptomatic or early post‐menopausal	3	699	Std. Mean Difference (IV, Random, 95% CI)	0.50 [0.04, 0.96]
1.1.2 Unselected post‐menopausal women	6	608	Std. Mean Difference (IV, Random, 95% CI)	0.64 [‐0.12, 1.41]
1.2 Desire	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.3 Arousal	4	623	Std. Mean Difference (IV, Random, 95% CI)	0.14 [‐0.14, 0.42]
1.3.1 Symptomatic or early post‐menopausal	1	239	Std. Mean Difference (IV, Random, 95% CI)	0.14 [‐0.12, 0.39]
1.3.2 Unselected post‐menopausal women	3	384	Std. Mean Difference (IV, Random, 95% CI)	0.14 [‐0.29, 0.58]
1.4 Lubrication	4	619	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.12, 0.83]
1.4.1 Symptomatic or early post‐menopausal	1	239	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.21, 0.73]
1.4.2 Unselected post‐menopausal women	3	380	Std. Mean Difference (IV, Random, 95% CI)	0.47 [‐0.09, 1.02]
1.5 Orgasm	3		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.6 Satisfaction	5	824	Std. Mean Difference (IV, Random, 95% CI)	0.41 [0.04, 0.77]
1.6.1 Symptomatic or early post‐menopausal	2	446	Std. Mean Difference (IV, Random, 95% CI)	0.29 [0.08, 0.51]
1.6.2 Unselected post‐menopausal women	3	378	Std. Mean Difference (IV, Random, 95% CI)	0.48 [‐0.21, 1.16]
1.7 Pain	5	958	Std. Mean Difference (IV, Random, 95% CI)	0.57 [0.14, 1.00]
1.7.1 Symptomatic or early post‐menopausal	2	548	Std. Mean Difference (IV, Random, 95% CI)	0.35 [0.14, 0.56]
1.7.2 Unselected post‐menopausal women	3	410	Std. Mean Difference (IV, Random, 95% CI)	0.72 [‐0.09, 1.52]

Comparison 2. Estrogen + progestogens versus placebo/no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Sexual function composite score	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.1.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.08 [‐1.52, 1.68]
2.2 Desire at 16‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.2.1 Symptomatic or early post‐menopausal	1	158	Mean Difference (IV, Fixed, 95% CI)	0.70 [0.21, 1.19]
2.3 Desire at 22‐ and 24‐month follow‐up	2	155	Std. Mean Difference (IV, Fixed, 95% CI)	0.19 [‐0.12, 0.51]
2.3.1 Unselected post‐menopausal women	2	155	Std. Mean Difference (IV, Fixed, 95% CI)	0.19 [‐0.12, 0.51]
2.4 Arousal	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.4.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.32 [‐0.11, 0.75]
2.5 Lubrication	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.5.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.45 [0.13, 0.77]
2.6 Orgasm	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.6.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.04 [‐0.53, 0.61]
2.7 Satisfaction	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.7.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.13 [‐0.28, 0.54]
2.8 Pain	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
2.8.1 Unselected post‐menopausal women	1	104	Mean Difference (IV, Fixed, 95% CI)	0.93 [0.02, 1.84]

Comparison 3. Synthetic steroids versus placebo/no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Sexual function composite score	2		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
3.1.1 Symptomatic or early post‐menopausal	1	883	Std. Mean Difference (IV, Random, 95% CI)	1.32 [1.18, 1.47]
3.1.2 Unselected post‐menopausal women	1	105	Std. Mean Difference (IV, Random, 95% CI)	0.46 [0.07, 0.85]
3.2 Desire at 12‐month follow‐up	2		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.2.1 Symptomatic or early post‐menopausal	1	57	Std. Mean Difference (IV, Fixed, 95% CI)	1.19 [0.62, 1.75]
3.2.2 Unselected post‐menopausal women	1	39	Std. Mean Difference (IV, Fixed, 95% CI)	0.19 [‐0.44, 0.82]
3.3 Desire at 24‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.3.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.35 [0.21, 0.49]
3.4 Arousal at 12‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.4.1 Unselected post‐menopausal women	1	39	Mean Difference (IV, Fixed, 95% CI)	0.19 [‐0.43, 0.81]
3.5 Arousal at 24‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.5.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.87 [0.47, 1.27]
3.6 Lubrication at 12‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.6.1 Unselected post‐menopausal women	1	39	Mean Difference (IV, Fixed, 95% CI)	0.56 [0.04, 1.08]
3.7 Lubrication at 24‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.7.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.62 [0.28, 0.96]
3.8 Orgasm at 12‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.8.1 Unselected post‐menopausal women	1	39	Mean Difference (IV, Fixed, 95% CI)	0.25 [‐0.98, 1.48]
3.9 Orgasm at 24‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.9.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.31 [‐0.19, 0.81]
3.10 Satisfaction	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.10.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.51 [0.12, 0.90]
3.11 Pain at 12‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.11.1 Unselected post‐menopausal women	1	39	Mean Difference (IV, Fixed, 95% CI)	0.12 [‐0.18, 0.42]
3.12 Pain at 24‐month follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
3.12.1 Unselected post‐menopausal women	1	105	Mean Difference (IV, Fixed, 95% CI)	0.90 [0.01, 1.79]

Comparison 4. Selective estrogen receptor modulators (SERMs) versus placebo/no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Sexual function composite score	3	1740	Mean Difference (IV, Fixed, 95% CI)	0.84 [0.18, 1.49]
4.1.1 Symptomatic or early post‐menopausal women	1	215	Mean Difference (IV, Fixed, 95% CI)	‐1.00 [‐2.00, ‐0.00]
4.1.2 Unselected post‐menopuasal women	2	1525	Mean Difference (IV, Fixed, 95% CI)	2.24 [1.37, 3.11]
4.2 Desire	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.2.1 Unselected post‐menopausal women	2	1545	Mean Difference (IV, Fixed, 95% CI)	0.25 [0.14, 0.36]
4.3 Arousal	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.3.1 Symptomatic or early post‐menopausal	1	215	Mean Difference (IV, Fixed, 95% CI)	‐0.22 [‐0.44, ‐0.00]
4.3.2 Unselected post‐menopausal women	2	1543	Mean Difference (IV, Fixed, 95% CI)	0.28 [0.11, 0.46]
4.4 Lubrication	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.4.1 Symptomatic or early post‐menopausal	1	215	Mean Difference (IV, Fixed, 95% CI)	‐0.17 [‐1.44, 1.10]
4.4.2 Unselected post‐menopausal women	2	1546	Mean Difference (IV, Fixed, 95% CI)	0.62 [0.42, 0.82]
4.5 Orgasm	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.5.1 Symptomatic or early post‐menopausal	1	215	Mean Difference (IV, Fixed, 95% CI)	‐0.38 [‐0.66, ‐0.10]
4.5.2 Unselected post‐menopausal women	2	1546	Mean Difference (IV, Fixed, 95% CI)	0.35 [0.15, 0.55]
4.6 Satisfaction	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.6.1 Symptomatic or early post‐menopausal	1	215	Mean Difference (IV, Fixed, 95% CI)	‐0.28 [‐0.56, ‐0.00]
4.6.2 Unselected post‐menopausal women	2	1527	Mean Difference (IV, Fixed, 95% CI)	0.23 [0.07, 0.39]
4.7 Pain	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
4.7.1 Unselected post‐menopausal women	2	1543	Mean Difference (IV, Fixed, 95% CI)	0.53 [0.33, 0.74]

Comparison 5. Selective estrogen receptor modulators (SERMs) + estrogens versus placebo/no treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Sexual function composite score	2		Std. Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.1.1 Symptomatic or early post‐menopausal	1	542	Std. Mean Difference (IV, Fixed, 95% CI)	0.22 [0.00, 0.43]
5.1.2 Unselected post‐menopausal women	1	272	Std. Mean Difference (IV, Fixed, 95% CI)	2.79 [2.41, 3.18]
5.3 Arousal	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.3.1 Symptomatic or early post‐menopausal	1	542	Mean Difference (IV, Fixed, 95% CI)	0.07 [0.03, 0.10]
5.4 Lubrication	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.4.1 Symptomatic or early post‐menopausal	1	542	Mean Difference (IV, Fixed, 95% CI)	‐0.33 [‐0.51, ‐0.16]
5.5 Orgasm	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
5.5.1 Symptomatic or early post‐menopausal	1	542	Mean Difference (IV, Random, 95% CI)	‐0.05 [‐0.42, 0.32]
5.6 Satisfaction	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.6.1 Symptomatic or early post‐menopausal	1	542	Mean Difference (IV, Fixed, 95% CI)	‐0.02 [‐0.27, 0.22]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Archer 2019.

Study characteristics
Methods	RCT phase 3, multicentre, 68 centres in the USA
Participants	Inclusion criteria: "postmenopausal women aged 40‐80 years with intact uterus, or having a serum follicle‐stimulating hormone level > 40 IU/L for hysterectomized women with intact ovaries, or 6 weeks since bilateral oophorectomy, or women aged 45 years with an unknown date of their last spontaneous menstrual bleed reporting moderate or severe vaginal dryness as their MBS with ≤ 5% superficial cells on their vaginal wall smear and a vaginal pH Inclusion criteria > 5.0 were eligible to participate".
Interventions	Intervention for up to 12 weeks: oral ospemifene 60 mg once daily Control: placebo (local‐based lubricant K‐Y Jelly as needed during penis‐vagina sexual relation)
Outcomes	FSFI scores were also secondary endpoint, collected at weeks 4, 8, and 12
Notes	Dr Archer has received research support from Actavis, Bayer Healthcare, Endoceutics, Glenmark, Merck, Radius Health, Shionogi, and Therapeutics MD; has served as consultant to AbbVie, Actavis, Agile Therapeutics, Bayer Healthcare, Endoceutics, Exeltis, InnovaGyn, Merck, Pfizer, Radius Health, Sermonix, Shionogi, Teva Women’s Healthcare, and Therapeutics. Dr Goldstein serves on the advisory board for AbbVie, Allergan, IBSA, Pfizer, and Therapeutics MD; and is a consultant for Cook ObGyn and Cooper Surgical. Dr Simon has served as consultant/advisor to AbbVie, Allergan, AMAG, Amgen, Ascend Therapeutics, Azure Biotech, Bayer Healthcare, CEEK Enterprises, Covance, Millendo Therapeutics, Mitsubishi Tanage
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomization schedule prepared by an independent statistician before study start
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	All study staff and participants were blinded throughout the study; ospemifene 60 mg and placebo tablets were identical in appearance and packaging
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	FSFI total scores and scores for all domains of FSFI
Selective reporting (reporting bias)	Low risk	Participant withdrawal (ospemifene, 4.1%; placebo, 5.1%). Study discontinuation due to AEs was low (ospemifene, 1.9%; placebo, 3.2%). The majority of women who received ospemifene and placebo were included in the intention to treat (ITT) population (99.1% vs 99.7%) and completed the 12‐week study (89.6% vs 88.6%)
Other bias	Low risk	Not suspected

Bumphenkiatikul 2020.

Study characteristics
Methods	RCT, single centre, Thailand
Participants	Inclusion criteria: women aged 45–70 years who experienced spontaneous menopause with penile‐vaginal penetrative sexual intercourse at least once a month and reported at least one self‐assessed vaginal atrophy symptom in moderate or severe intensity
Interventions	Intervention: intravaginal ovule containing 0.03 mg estriol Control: matching placebo
Outcomes	FSFI total scores and scores for all domains of FSFI
Notes	The study was supported by THE 90TH ANNIVERSARY OF CHULALONGKORN UNIVERSITY FUND (Ratchadaphiseksomphot Endowment Fund). The funding body had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized randomization
Allocation concealment (selection bias)	Low risk	The allocation and concealment of placebo and the drugs into identical opaque envelopes were done by a nurse...
Blinding of participants and personnel (performance bias) All outcomes	Low risk	...who was not involved in contacting patients or analysing any data
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding complete and unlikely that the blinding could have been broken
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasosns for missing data not related to outcome
Selective reporting (reporting bias)	Low risk	The protocol is available
Other bias	Low risk	The study appears to be free from sources of risk

Bushmakin 2014.

Study characteristics
Methods	RCT, multi centre, 43 sites in the USA
Participants	Inclusion criteria: postmenopausal women aged 40‐65 years with an intact uterus who had ≥ 7 moderate‐severe hot flushes/d at baseline (or ≥ 50/week)
Interventions	Intervention 12 weeks: bazedoxifene 20 mg + conjugate estrogens 0.45 mg, bazedoxifene 20 mg + conjugate estrogens 0.625 mg Control: matching placebo
Outcomes	MENQOL, sexual function domain score
Notes	Funding and conflict of interest was copied in full from the manuscript: "The demographic data and the risk of bias has been reported in the first SMART‐2 study. This study was sponsored by Pfizer. The SMART‐2 trial was sponsored by Wyeth Research, which was acquired by Pfizer in October 2009. The medical writing support was provided by Katie McClendon, PhD, of MedErgy, and Diane M. Sloan, PharmD, of Peloton Advantage and was funded by Pfizer. Three authors of the study (AGB, JCC, and SM) are employees of Pfizer Inc. One author of the study (LA) is an employee of Pfizer Ltd. One author of the study (JVP) has served as a consultant (fees to the University of Virginia) for Pfizer, Noven Pharmaceuticals, Shionogi, and DepoMed; has received grants/research support (fees to the University of Virginia) from DepoMed, Bionova, and Endoceutics; and has received travel funds from Pfizer, Noven Pharmaceuticals, Shionogi, and DepoMed"
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized randomization/enrolment system generated by Wyeth Research
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It is not clear how many women answered the final questionnaires, and if the women who withdrew from treatment had the final evaluation
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Caruso 2020.

Study characteristics
Methods	RCT, single‐centre, Italy
Participants	Inclusion criteria: "Ninety‐six postmenopausal women ranging in age from 47 to 73 years (mean age 57.5 ± 6.8), affected by SUI requiring surgical treatment, were eligible for this study" Exclusion: "body mass index (BMI) ≥ 35 kg/m2, endometrial thickness ≥ 4 mm, abnormal uterine bleeding,hormone‐dependent malignancies, history of thromboembolic disease,liver disease, and/or using systemic or local hormone therapy for less than 3 months, phytoestrogens use, use of nonhormonal lubricants or moisturizers within 30 days before the start of the study, without sexual relation for more than a month, having a partner with a sexual disorder"
Interventions	Intervention: 0.03 mg daily estriol vaginal ovule for 8 weeks before and 8 weeks after transobturator tape (TOT) surgical placement.
Outcomes	FSFI total scores and scores for all domains of FSFI are reported
Notes	Funding and financial disclosure/conflicts of interest: none reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The participants were enroled by a single operator
Allocation concealment (selection bias)	Unclear risk	We are not sure how the assignments were hidden
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	We are not sure how the participants and personnel were blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	We are not sure how the outcome assessment was blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasosns for missing data not related to outcomes
Selective reporting (reporting bias)	High risk	The protocol was not available
Other bias	High risk	Possible biases in randomization or blinding

Constantine 2015.

Study characteristics
Methods	RCT, parallel‐group, phase‐3, USA
Participants	Inclusion criteria: women with symptoms of vaginal atrophy and bothersome symptoms of either dyspareunia or vaginal dryness. Exclusion criteria are unclear.
Interventions	Intervention: ospemifene 60 mg/day administered for 12 weeks Control: placebo
Outcomes	The total FSFI score is the sum of the scores for the 6 individual domains
Notes	Study does not report RCT protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	High risk	Does not report demographic data after losses. Suspected unbalance of losses in the groups
Selective reporting (reporting bias)	High risk	Does not report baseline scores of FSFI domains
Other bias	Unclear risk	Does not report details of randomization in recruitment. Recruitment bias suspected

Czarnecka 2000.

Study characteristics
Methods	RCT, single centre, Poland
Participants	Inclusion: postmenopausal women (amenorrhoea > 6 months and estradiol < 50 pg/mL + FSH 21U/L) with mild and moderate essential hypertension Exclusion criteria are unclear.
Interventions	Intervention: 17‐β‐estradiol and norethisterone acetate (Estracomb TTS, Norvatis) Control: no treatment
Outcomes	SSA‐P Final scores
Notes	Age: intervention (49.4 ± 5.8 years), control (52.0 ± 6.4 years) BMI: intervention (27.4 ± 4.7kg/m²), control (25.0 ± 3.6kg/m²) Follow‐up: 12 weeks Do not report conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described in detail
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open study
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described in detail
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	High risk	The groups differed in baseline characteristics. We observed relevant differences in the baseline sexual function (74.8 ± 24.7 vs 66.0 ± 22.4, treatment and control respectively), QoL (12.6 ± 2.3 vs 11.4 ± 2.5), and hot flushes (74.8 ± 24.7 vs 66.0 ± 22.4) scores. Study authors did not present the results as change from baseline.

Dayal 2005.

Study characteristics
Methods	RCT, single‐centre, USA
Participants	Inclusion: menopausal women, aged 44–70 years, with a history of no menstrual cycle for at least 1 year or at least 6 months of amenorrhoea with a documented FSH level > 40 mIU/mL. "These women had no exposure to HT (specifically estrogen therapy or DHEA supplementation) for at least 60 days prior to enrolment, had a normal Pap smear and mammogram within the last year, and had normal liver transaminase levels, renal function, total cholesterol, and triglyceride levels". Exclusion: "any contraindication to estrogen therapy (known or suspected cancer of the breast, endometrial hyperplasia/carcinoma, undiagnosed vaginal bleeding, active thromboembolic disorders, or history of cerebrovascular disease, coronary artery disease, or myocardial infarction), a current diagnosis of diabetes mellitus, uncontrolled hypertension, abnormal liver or renal function, major psychiatric disorder (major depression, bipolar disorder, psychotic disorder, drug addiction), or any contraindication to the use of magnetic resonance imaging (pacemaker, magnetic aneurysm clip, severe claustrophobia)"
Interventions	Intervention: conjugated equine estrogen 0.625 mg daily Control: placebo (We included only the arms without DHEA in this review)
Outcomes	WHQ (bigger values correspond to better conditions)
Notes	We excluded 2 arms with use of DHEA from the analysis Follow‐up: 3 months Mean age 56.6; ranging from 44‐77 Corresponding author did not respond to our contact via email Article presents the differences from baseline (in %) without SD or P values For sexual function, depression, and QoL, we determined the final value based on the (%) difference and considered the final SD to be equal to the baseline SD This study was supported by an unrestricted grant from the Berlex Foundation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized program generating random numbers
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss after allocation
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	High risk	There was a relevant difference in baseline values between intervention and control groups for sexual function (4.33 ± 1.53 vs 5.00 ± 2.12, intervention and control respectively), and QoL (15.25 ± 3.40 vs 13.00 ± 4.90). Although study authors provided relative change from baseline (in %) we could not use these data because we were not able to estimate SDs.

Diem 2018.

Study characteristics
Methods	RCT, 2 centres, USA
Participants	302 women were randomized to receive vaginal estradiol tablet plus placebo gel (n = 102), placebo tablet plus vaginal moisturizer (n = 100), or dual placebo (n = 100) Inclusion: women aged 45‐70 years old, ≥ 2 years since last menses, reported ≥ 1 moderate–severe symptom of vulvovaginal itching, pain, dryness, or irritation experienced at least weekly within the past 30 days; or pain with penetration at least once monthly. Exclusion: current vaginal infection; use of hormonal medication in past 2 months; use of antibiotics, vaginal moisturizer, probiotic, prebiotic, or douche in past month; and chronic premenopausal vulvovaginal symptoms
Interventions	Intervention: vaginal estradiol 10 mcg tablet (Vagifem) + placebo vaginal gel (hydroxyethyl cellulose); placebo vaginal tablet + vaginal moisturizer (Replens); or placebo tablet + placebo gel
Outcomes	MENQOL questionnaire (higher scores indicated poorer QoL), PHQ‐8, FSFI, and the Female Sexual Distress Scale‐Revised
Notes	Data on ≥ 1 domains of the MENQOL were available on 301 women (> 99%) at baseline and 298 (99%) at 4 and/or 12 weeks of follow‐up. PHQ‐8 and GAD‐7, data were available on 301 (> 99%) women at baseline and 297 (97%) at follow‐up. Women with both baseline and follow‐up MENQOL data, 83% women randomized to estradiol tablets were inclusion criteria 80% adherent, compared to 80% women randomized to placebo tablets. 74% women randomized to vaginal moisturizer were inclusion criteria 80% adherent, compared to 79% of women randomized to placebo gel Total MENQOL scores declined, that is, menopause‐related QoL improved, in all treatment groups Vaginal estradiol group significantly improved the sexual function domain (Table 2) (mean difference at 12 weeks of 0.4 (95% CI 1.0, 0.1; P = 0.005) PHQ‐8 scores improved in all treatment groups at week 12 No evidence of interactions between treatment assignment and selected characteristics (i.e. presence or absence of sexual function–related MBS and baseline FSFI) on total MENQOL score or sexual function domain score (data not shown)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization by permuted blocks
Allocation concealment (selection bias)	Low risk	Used a computerized inventory system for dispensing identical‐appearing tablets in bottles and gel in tubes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, study personnel, and clinical providers were blinded to treatment assignments
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Unknown if there was blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Lost data were reported
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Contamination and recruitment bias not suspected

Fonseca 2007.

Study characteristics
Methods	Cross‐over, single centre, Brazil March 2001‐February 2003
Participants	Inclusion: postmenopausal women aged 42–61 years; with uterus and no contraindications to HT Exclusion: systemic, psychiatric or endocrine disease; smoking > 10 cigarettes/d; use of any medication that could interfere with climacteric symptoms or with their sexuality
Interventions	Intervention: 17β‐estradiol 2 mg with norethisterone acetate 1 mg (Cliane, Schering) Control: matching placebo
Outcomes	GRISS Data not suitable for meta‐analysis
Notes	Follow‐up: 6 months each part of the cross‐over trial No available data to report Financial support for this study from Schering do Brasil Quimica e Farmaceutica Ltda
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, details not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	High risk	The study authors did not report all the results obtained. They reported only FSFI domains with statistically significant differences
Other bias	Low risk	Not suspected

Gambacciani 2011.

Study characteristics
Methods	RCT, parallel‐group, single centre, Italy
Participants	Inclusion: amenorrhoea for at least 12 months before treatment, plasma gonadotropin and estradiol levels in the postmenopausal range, office BPs in the normal rang, vasomotor symptoms. Exclusion: age > 60 years; previous intolerance to HT; absence of vasomotor symptoms; history of hypertension and/or treatment with antihypertensive drugs within the last 3 months; diabetes or any other metabolic disorder including dyslipidaemia, history of CVD, venous thromboembolic disease or haemostatic disorder predisposing to thromboembolic complications; abnormal uterine bleeding; any unstable medical condition; or history of alcoholism, drug abuse, psychosis, or other emotional or intellectual problems likely to invalidate informed consent or to limit compliance with protocol requirements.
Interventions	Intervention: estradiol 1 mg + drosperinone 2 mg (Angeliq, Bayer Schering Pharma AG, Berlin, Germany) Control: calcium (1 g/d)
Outcomes	WHQ Data not suitable for meta‐analysis
Notes	Follow‐up: 3 months 70 women evaluated; 35 allocated Data from WHQ not given in numbers, just presented as a graphic No third‐party sources provided financial support for the study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	70 consecutive eligible women were randomized to treatment using a randomization list (yes – no)
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Dropouts: 3/23 in the control group; 0/32 in the intervention group
Selective reporting (reporting bias)	High risk	Although sexual function was measured by a questionnaire, the study authors did not provide the numeric results. They stated that the scores improved after intervention
Other bias	Low risk	Not suspected

Gast 2009.

Study characteristics
Methods	RCT, double‐blind, placebo‐controlled, multi‐centre (25 sites), Brazil and USA May 2001‐April 2004
Participants	Inclusion: women aged 45‐65 years at enrolment, had a sexual partner or partners, and had an intact uterus, their last natural menstrual cycle at least 12 months before screening (or 6 months and high FSH). Maximum of 10% superficial cells on the Vaginal Maturation Index. Endometrial thickness ≤ 5 mm by TVUS. Exclusion: history or presence of neoplasia or endometrial hyperplasia, history of thrombophlebitis, thrombosis, or thromboembolic disorders (cerebrovascular accident, stroke, or transient ischaemic attack) related to estrogen use; neuro‐ocular disorders; myocardial infarction or ischaemic heart disease; chronic renal or hepatic disease; or a known hypersensitivity to estrogen, progestogen, or other ingredients of the hormonal preparations used in the study. Also, women were excluded if they had used any oral or vaginal estrogen‐, progestogen‐, or androgen‐containing medications within 8 weeks of screening or any transdermal hormone therapy within 4 weeks before screening. In addition women diagnosed with sexual dysfunctions other than dyspareunia were excluded.
Interventions	Intervention: conjugated estrogen 0.45 mg + 1.5 mg MPA for six 28‐d cycles along with initial vaginal priming with 1 g CEE (Premarin) cream (0.625 mg CEE/g) intravaginally for the first 6 weeks Control: matching placebo
Outcomes	MFSQ BISF‐W Data not suitable for meta‐analysis
Notes	Reported the results of the questionnaire in figures, no numeric data presented This study was funded by Wyeth, Collegeville, PA. Premarin and Premarin Vaginal Cream are manufactured and marketed by Wyeth. Alberta Vieweg and Michael Gast are current or former employees of Wyeth Follow‐up: 3 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Claim to be double‐blind, no details described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropouts: 41/144 in the intervention group; 29/112 in the control group
Selective reporting (reporting bias)	High risk	Although sexual function was evaluated by a validated questionnaire the scores are not reported, the results are shown in graphics
Other bias	Low risk	None

Haines 2009.

Study characteristics
Methods	RCT, multicentre: Thailand, the Philippines, Singapore, Hong Kong and Malaysia
Participants	Inclusion: aged 40‐65 years, had undergone natural menopause (12 months’ amenorrhoea or 6 months’ amenorrhoea with FSH > 40 mIU/mL) or bilateral oophorectomy (6 weeks postsurgery) and had at least 24 hot flushes (of any severity) within a 7‐d screening period Exclusion: recently used estrogen‐containing products, an abnormal cervical smear test, endometrial thickness of 5.0 mm, any condition that could interfere with study medication or interpretation of results, concomitant use of inducers or inhibitors of CYP3A4 or drugs effective in treating hot flushes, received anticoagulant treatment for the past 6 months, or known severe dyslipoproteinaemia"
Interventions	Intervention: transdermal patch delivering micro‐dose E2 (0.014 mg/d) Control: placebo
Outcomes	MENQOL questionnaire Change from baseline scores
Notes	Follow‐up: 12 weeks Age: intervention (52.6 ± 3.99 years); control (52.2 ± 4.73 years) The following outcomes could not be pooled because of lack of details: BMI: intervention (24.0 ± 3.69 kg/m²); control (24.0 ± 3.70kg/m²) hot flushes: reduction in mean weekly number of hot flushes ‐ data not described in detail, could not be pooled involuntary urination: intervention (28.6%); control (18.7%) urge‐incontinence: intervention (10.4%); control (6.7%) vaginal dryness: intervention (37.7%); control (46.7%) This study was supported financially by Bayer Schering Pharma AG. Editorial support was provided by Parexel MMS
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomization was by a centrally provided computer‐generated list
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	Although it is not clear how many women were assessed for sexual function; we assumed it was only those who completed the study. Those participants had > 75% overall study drug compliance; thus not respecting the ITT principle
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Hemminki 2009.

Study characteristics
Methods	RCT, parallel‐group, 3 clinical centres in Estonia 777 women randomized, January 1999‐December 2001
Participants	Inclusion: aged 50–64 years, postmenopausal Exclusion: none
Interventions	Intervention: CEE 0.625 mg/d + MPA 5.0 mg/d Control: matching placebo
Outcomes	WHQ Percentage of women satisfied and who lost interest in sex
Notes	Follow‐up: 1 year for the included evaluation (3.6 years total) Trial register: ISRCTN35338757 Potentially eligible women were randomized into 4 trial groups resulting in 2 substudies: blind HT group vs blind placebo group (Estonian blind trial) non‐blind HT group vs non‐treatment group (Estonian non‐blind trial) 1001 women were recruited into the non‐blind substudy and 777 in the blind substudy. The Estonian Postmenopausal Hormone Therapy trial was originally planned to be a part of the Women’s International Study of Long Duration Oestrogen After Menopause (WISDOM) The study was financially supported by grants from the Academy of Finland (grants 69838, 201490, 115088), STAKES (National Research and Development Centre for Welfare and Health), Finland, and the Estonian Ministry of Education and Research Drugs were donated by Wyeth Ayerst Company via the Women's International Study of Long Duration Oestrogen After Menopause (WISDOM) in the UK This study did not contribute data to meta‐analyses
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization in permuted blocks, each of a size 16 and each block of the 3 clinics separately, at the National Research and Development Centre for Welfare and Health in Finland
Allocation concealment (selection bias)	Low risk	The treatment allocation was enclosed in a non‐transparent sealed envelope and sent to trial clinics
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Use of matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Balanced dropouts Blind trial ‐ intervention 98/404; control 101/372
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Huang 2008.

Study characteristics
Methods	RCT, multicentre, 9 clinical centres in USA ‐ ULTRA trial
Participants	Inclusion: aged 60‐80 years, with an intact uterus but who had not had a menstrual period in at least 5 years, normal bone mineral density for age (z‐score ≥ 2.0 at the lumbar spine) and could not have used estrogen or progestogen in the 3 previous months Exclusion: unexplained uterine bleeding, endometrial hyperplasia, abnormal mammogram suggestive of breast cancer, CVD, venous thromboembolism, uncontrolled hypertension, uncontrolled thyroid disease, liver disease, fasting triglyceride level > 300 mg/dL, or fasting glucose level > 180 mg/dL
Interventions	Intervention: 3.25 cm² patch releasing 0.014 mg/d estradiol Control: matching placebo
Outcomes	MOS Sexual Problems Index Final scores, data sent by the study authors
Notes	Follow‐up: 24 months Age: intervention = 66.8 (± 5.1); control = 66.7 (± 4.8) years This research was supported in part by a grant from Berlex Laboratories Inc and manufacturer of the estradiol patch used; and by Grant KLZRR024130 from the National Center for Research Resource, a component of the NIH
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomly permuted blocks of size 4
Allocation concealment (selection bias)	Low risk	Participants, investigators and outcome assessors were blinded to treatment assignment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, investigators and outcome assessors were blinded to treatment assignment, and no blinding occurred during the trial
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Participants, investigators and outcome assessors were blinded to treatment assignment, and no blinding occurred during the trial
Incomplete outcome data (attrition bias) All outcomes	High risk	Sexual function evaluated only in women who were sexually active during in the previous 4 weeks: sexual function was evaluated on 186/208 participants in the intervention group, and on 180/209 in the control group
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Hudita 2003.

Study characteristics
Methods	RCT, parallel‐group, single centre, Romania
Participants	Inclusion: healthy women aged 40‐65 years, non‐obese, postmenopausal, with an intact uterus Exclusion: previous deep thrombophlebitis or cerebral apoplexia; mammary or gynaecological neoplasia; uncontrolled diabetes; abnormal mammography; abnormal values of C‐reactive and S‐coagulative protein; endometrial thickness > 4 mm or an abnormal Pap smear; abnormal bleeding of undetermined origin; history of hepatic or renal diseases; use of estrogenic, progestogen or androgenic drugs in the 8 weeks preceding enrolment into the study; use of IUD in the preceding 3 months; use of medications known to affect vasomotor system in the 2 weeks prior to enrolment; known hypersensitivity to study drugs; glucose intolerance (blood glucose > 125 mg/dL), hypercholesterolaemia (total cholesterol > 300 mg/dL) or hypertriglyceridaemia (triglycerides > 300 mg/dL), hypertension (systolic BP > 160 mm Hg or diastolic BP > 90 mm Hg)"
Interventions	Intervention 1: tibolone 1.25 mg Intervention 2: tibolone 2.5 mg Control > matching placebo
Outcomes	MFSQ Data not suitable for meta‐analysis
Notes	Follow‐up: 24 weeks Tibolone 1.25 and 2.50 mg tablets were supplied by Libbs as Libiam
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described in detail
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Used matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	42/162 women dropped out after inclusion, groups not described
Selective reporting (reporting bias)	High risk	Although sexual function was evaluated by a validated questionnaire, the scores are not reported. The study authors reported only the proportion of women with acceptable quality of sexual life
Other bias	Low risk	Not reported

Laan 2001.

Study characteristics
Methods	Cross‐over trial, single centre, the Netherlands
Participants	Inclusion: postmenopausal women with intact uteri and ovaries, aged < 65 years with BMI of 18–29 kg/m², last menstrual period ≥ 12 months prior to the study Exclusion: women using medication known to influence sexual desire and/or response or who had used HT in the 3 months prior to enrolment
Interventions	Intervention: tibolone 2.5 mg/d for 3 months Control: placebo
Outcomes	Greene Climacteric Scale Change from baseline scores of the first phase of the cross‐over; provided by the study author
Notes	We included only the results concerning the first phase (before cross‐over): these data were obtained by correspondence with the study authors. This work was supported by NV Organon Follow‐up: 3 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	Data were available from only 37 women (17 in the intervention group and 20 in the control group): 44 women were randomly allocated
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Liberate 2011.

Study characteristics
Methods	RCT, multicentre, 245 sites from 31 countries. Sexual function was examined only in women from 8 countries (Austria, Belgium, Germany, Spain, France, UK, Italy, The Netherlands) June 2002‐July 2007
Participants	Inclusion: women with climacteric symptoms were eligible if they had been surgically treated within the previous 5 years for histologically confirmed breast cancer. Participants had to be postmenopausal and < 75 years. Exclusion: non‐hysterectomized women with endometrial abnormalities
Interventions	Intervention: tibolone 2.5 mg daily, orally Control: matching placebo About 75% of participants in both groups were using either tamoxifen, aromatase inhibitor, or GnRH analogues
Outcomes	WHQ Change from baseline scores WHQ 'sexual' and 'depression': the smaller the values the better the condition
Notes	Follow‐up: 26 weeks Mean age: 52.7 (± 7.3) years For sexual function and vasomotor domain/hot flushes, SD was calculated from P values Depression could not be included because P values were lacking Vaginal dryness could not be included because only stratified data were reported (no difference observed on any strata) Funding: Schering‐Plough (formerly NV Organon, Oss, Netherlands). The sponsor conducted the trial and collected the data. An Advisory Board had overall scientific responsibility for trial design and protocol, and advised the sponsor as to the conduct of the trial. The corresponding author had full access to all data relevant for the article publication and all authors were involved in its submission.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Centralized interactive voice response system, stratified by centre, with a block size of 4
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, investigators, sponsor personnel, and outcome assessors were blinded to treatment assignment
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	Sexual function was evaluated in only 438/1575 women in the treatment group, and on 445/1558 in the placebo group
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Maki 2007.

Study characteristics
Methods	RCT, multi‐centre, 19 medical centres, USA
Participants	Inclusion: healthy postmenopausal women with intact uteri and whose last natural menstrual cycle had been completed ≥ 12 and ≤ 36 months prior to screening, or last natural menstrual cycle between 6 and ≤ 12 months before screening, provided their serum FSH more than the lower limit reference of normal level for postmenopausal women. Participants spoke English as their primary language and were required to demonstrate verbal knowledge with a total score 10 on the Primary Mental Abilities Vocabulary Test. Women were required to have reported at least 1 cognitive complaint on the Self‐Reported Cognitive Function Questionnaire, a brief subjective survey of memory and concentration, and to be free of neurologic, systemic, or psychiatric disease that would influence cognition. Exclusion: known or suspected estrogen‐dependent neoplasia, endometrial hyperplasia, thrombophlebitis, thrombosis, thromboembolic disorders, myocardial infarction, ischaemic heart disease, cerebrovascular accident, stroke, transient ischaemic attack, or hypersensitivity to estrogen or progestogens were not included in the study. Other exclusion criteria included recent (≤ 8 weeks prior to screening) use of estrogen or progestogen as estrogen modulates memory‐related synaptic plasticity, excessive smoking (≥ 0.5 pack of cigarettes/d), a history of alcohol abuse (based on a Michigan Alcohol Screening Test score of ≥ 5), evidence of cognitive impairment on the Mini‐Mental State Examination (total score < 24), or use of any medication that might influence central nervous system function.
Interventions	Intervention: CEE 0.625 mg + medroxiprogesterone acetate 2.5 mg daily Control: matching placebo
Outcomes	Greene Climacteric Scale Change from baseline
Notes	SD calculated from the P values of the differences between the groups MFSQ–level of interest in sex was considered as sexual function Mean age: intervention (51.9 ± 3.6 years); control (52.4 ± 3.3 years) Weight: intervention (70 ± 16.1 kg); control (67.8 ± 16.4 kg) Follow‐up: 4 months This study was funded by Wyeth, Collegeville, PA. Prempro is manufactured and marketed by Wyeth. Dr. Maki has received grants (in excess of USD 10,000) and honoraria from Wyeth. Dr. Yaffe has received grants (in excess of USD 10,000) and honoraria from Wyeth. Dr. Gast, A.J. Vieweg, and S.W. Burriss are current or former employees of Wyeth. Dr. Gast has ownership interest in the company. Dr. Yaffe received compensation through the University of California for her role as a consultant throughout the trial.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Permuted block design
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropouts: intervention (11), control (11)
Selective reporting (reporting bias)	High risk	Sexuality, not sexual function, was assessed by the Greene Climateric Scale, sexual QoL (not sexual function) was assessed by Utian Quality, and the levels of sexual thoughts and interest were assessed by the MFSQ. However, only the partial results (total score and some domains) of the MFSQ were reported, which do not represent the sexual function itself.
Other bias	Unclear risk	Statistical analyses were performed by Wyeth Research, whose conflict of interest was stated in the article.

Mameri Filho 2005.

Study characteristics
Methods	RCT, single centre, Brazil
Participants	The diagnosis of menopause was based on clinical data, i.e. amenorrhoea for at least 12 months confirmed by elevated gonadotropin, FSH > 40 mIU/mL). Inclusion: women with vasomotor symptoms or moderate or marked Kupperman index > 39 points, with complaints of sexual function. Exclusion: women with absolute contraindication to HT, BMI > 30 (obese); Ferriman, Gallwey and Lorenzo > 8 (hirsute). In addition, women who were not on HT prior and did not use herbal or diet high in soy‐based foods (use prior 6 months) were selected for this study.
Interventions	Intervention 1: CEE 0.625 mg/d Control: placebo Intervention 2: arm not included: CEE (0.625 mg/d) associated with methyltestosterone (2.5 mg/d)
Outcomes	WHQ: composite scores MFSQ: sexual domains Final scores
Notes	We did not include the group using intervention 2 in this review Study authors provided baseline and final values for WHQ without SD. We estimated the SD from P values The article does not report any conflict of interest disclosure The drugs were donated by the university hospital Follow‐up: 3 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐based
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Use of placebo matching pills
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Mitchell 2018.

Study characteristics
Methods	RCT, multicentre, 2 centres USA
Participants	"Women aged 50‐70 years within 20 miles of the clinical sites at least 2 years since last menses, report of at least 1 moderate‐severe symptom of vulvovaginal itching, pain, irritation, or dryness experienced at least weekly within the past 30 days; or pain with penetration at least once monthly." Exclusion criteria included "current vaginal infection, use of hormonal medication in past 2 months, use of antibiotics or vaginal moisturiser in past month, and chronic premenopausal vulvovaginal symptoms"
Interventions	Intervention "1:1:1 to Vagifem 10‐μg tablet + placebo vaginal gel, placebo vaginal tablet + Replens vaginal moisturiser, or placebo tablet + placebo gel. Women were instructed to use the vaginal tablet daily for 2 weeks, then twice weekly for the remaining 10 weeks, and the vaginal moisturiser every 3 days throughout the trial". "During the first 2 weeks, participants were advised to use the tablet in the morning and gel in the evening. After that, participants were instructed to use products on alternate days."
Outcomes	The primary outcome was change in severity of the MBS Secondary outcomes were composite VSI, satisfaction with treatment received (Likert scale: 0 = not satisfied to 10 = completely satisfied), meaningful benefit from the study medications (yes or no), FSFI
Notes	The 302 women had a mean (SD) age of 61 (4) years Mean total FSFI improvement was similar between estradiol (5.4; 95% CI, 4.0 to 6.9) and placebo (4.5, 95% CI 2.8 to 6.1; P = 0.64), and between moisturizer (3.1, 95% CI 1.7 to 4.5) and placebo (P = 0.17)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Stratification through permuted blocks
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, study personnel, and clinicians were blinded to treatment assignments
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding, and unlikely that the blinding could have been broken
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing data balanced across groups, and reasons similar
Selective reporting (reporting bias)	Low risk	Protocol is available
Other bias	Low risk	Study appears to be free of other sources of risk

Modugno 2003.

Study characteristics
Methods	RCT, multicentre, USA and Canada This trial is a subset of older postmenopausal women participating in the Multiple Outcomes of Raloxifene Evaluation study, a multi‐centre, randomized, double‐blind, placebo‐controlled trial. However, only English‐speaking participants at selected study sites in the USA and Canada were administered the sexual function questionnaire (n = 1218).
Participants	Inclusion: women aged 31–80 years, postmenopausal for at least 2 years, meet the WHO criteria for having osteoporosis, English as a native language, completing the sexual function questionnaire at both the baseline and the 36‐month follow‐up visits
Interventions	Intervention 1: raloxifene 60 mg/d Intervention 2: raloxifene 120 mg/d Control: matching placebo
Outcomes	MSFQ Percentage of women with improved scores Data not suitable for meta‐analysis
Notes	Follow‐up: 36 months This work was supported by Eli Lilly and Company, maker of raloxifene. The company provided salary support to SE and research support to JAC, who is also a member of the speaker’s bureau of Eli Lilly and Company.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Block randomization by clinical site
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	High risk	7705 women were randomized in 25 countries, data of women from only 2 countries were included in this analysis (1218). Women with unsatisfactory questionnaire answers were excluded from analysis (275)
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Morais‐Socorro 2012.

Study characteristics
Methods	RCT, single centre, Brazil
Participants	Inclusion: aged 40‐55 years; menstrual irregularity during the previous 6 months but < 12 months of amenorrhoea; the presence of a uterus without anomalies in an initial vaginal ultrasonography evaluation and an endometrial thickness measurement ≤ 10 mm; plasma FSH level ≥ 30 mUI/mL; total cholesterol < 250 mg/dL; triglycerides ≤ 200 mg/dL; glycaemia ≤ 100 mg/dL; a Blatt‐Kupperman Menopausal Index score ≥ 14; mammography with BI‐RADS = 1 or 2 Exclusion criteria: current smoker; use of any hormonal, psychotropic or other medications that could interfere with the lipid glycaemic profile within 90 days of trial; history of clinical hepatic or kidney disease or diabetes mellitus; cerebrovascular, thyroid, CVD or thromboembolic disorders; any neoplastic disease; hypertensive disorder (systolic BP ≥ 160 mmHg and/or diastolic BP ≥ 105 mmHg); BMI > 30 kg/m²; and mammography with BI‐RADS = 3 or 4
Interventions	Intervention: tibolone 2.5 mg/d (ReduClim®, Farmoquímica, Brazil) Control: placebo (identical capsules containing lactose)
Outcomes	Greene Climacteric Scale Final scores
Notes	Age: intervention (48.0 ± 3.6 years), control (49.0 ± 3.6 years) BMI: intervention (26.2 ± 0.6 kg/m²), control (26.0 ± 0.6 kg/m²) Follow‐up: 12 weeks Loss to follow‐up: 3/30 in the intervention group; 5/35 in the control group Reduclin 2.5 mg was donated by the laboratory Farmoquímica, Brazil
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Use of placebo matching pills
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	Balanced dropouts (3/30 in the intervention group; 5/35 in the control group)
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Nathorst‐Böös 1993.

Study characteristics
Methods	RCT, parallel‐group, 15 centres in Sweden
Participants	Inclusion: postmenopausal women aged 45‐65 years requiring HT for climacteric symptoms Exclusion: malignant tumours, surgical menopause, CVD or respiratory diseases, diabetes, psychiatric illness, endometriosis or vaginal bleeding
Interventions	Intervention: estradiol therapy (Estraderm 50 mcg/d) Control: matching placebo
Outcomes	MSFQ WHQ Change form baseline scores
Notes	Follow‐up: 12 weeks No conflict of interests disclosure
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described in detail
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Not reported
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Dropouts: 3/242; groups not specified
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Nielsen 2006.

Study characteristics
Methods	RCT, multicentre, 2 centres, Denmark
Participants	Inclusion: 40–65 years of age at baseline, were randomized if they were < 5 years past menopause at study entry. Menopause was defined as amenorrhoea > 12 months or > 6 months and a concomitant serum level of estradiol < 0.16 nmol/L and a FSH > 42 IU/L. All women who had undergone hysterectomy had menopause confirmed by measurement of serum estradiol and FSH at least 2 months before study entry. Women who had undergone surgical menopause (bilateral oophorectomy) could be included, if the operation was performed at least 6 weeks before study entry. Women who previously had received estrogen replacement therapy, had to go through a 6‐month wash‐out period before randomization. To be in good general health with no clinical or laboratory evidence of systemic disease or conditions with known influence on bone metabolism. Women had to be osteopenic (BMD T score < −1) Exclusions: women with nasal disease incompatible with nasal therapy such as frequent epistaxis, chronic rhinitis or sinusitis, severe allergic rhinitis, or frequent nasal treatment; women with contraindications for the use of HT.
Interventions	Wash out: 6 months Intervention: intranasal sprays estrogen (Aerodiol, France) 150 mcg/d (n = 114) or 300 mcg/d. Women with an intact uterus also received 200 mg of micronized progestogen (Effik, Spain), combined with the nasal spray for the last 14 days of each 28‐day cycle Control: matching placebo (n = 118)
Outcomes	WHQ Change from baseline scores
Notes	Age: mean 53 years BMI: mean 25.2 kg/m² 6.3% of participants did not have a uterus and therefore used only estrogen Follow‐up: 2 years The article does not report any conflict of interest disclosure Included women up to 5 years of amenorrhoea
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization was performed according to the order of inclusion. Thus, the first participant received the lowest numbered therapeutic unit in the block of units supplied to the centre concerned
Allocation concealment (selection bias)	Low risk	The treatment allocation lists were drawn up and encoded by Servier Laboratoires
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	High risk	Only 268 women from the 335 allocated completed the evaluation questionnaires
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Osmanagaoglu 2006.

Study characteristics
Methods	RCT, single centre, Turkey
Participants	Inclusion: postmenopausal women (amenorrhoea > 1 year, confirmed by FSH ≥ 30 IU/L), had not undergone any gynaecological operation, and no absolute contraindication for HT Exclusion: aged > 60 years, smokers, with diabetes, cancer, liver, renal or haematological, disease or other medical disorders or with a, history of sexual pain disorders or disharmony with their husbands
Interventions	Intervention 1: tibolone 2.5 mg (Livial, Organon) Intervention 2: estradiol valerate 2 mg + dienogest 2 mg (Climodien, Schering) Control: no intervention
Outcomes	Rosen’s FSFI Final scores
Notes	Age: intervention 1 (50 ± 1.87 years); intervention 2 (50 ± 2.20 years); control (51+1.94 years) BMI: intervention 1 (26 ± 0.84 kg/m²); intervention (25 ± 0.97 kg/m²); control (25 ± 0.97 kg/m²) Follow‐up: 6 months Report no conflict of interest and no external funding
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described in detail
Allocation concealment (selection bias)	Unclear risk	Not described in detail
Blinding of participants and personnel (performance bias) All outcomes	High risk	Although the authors claim study to be single‐blind, the study's limitations make it an open‐label design
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Few and balanced dropouts
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	High risk	The groups differed in baseline characteristics. The baseline sexual function score in tibolone (13.23 ± 6.63) group was higher than in control group (12.60 ± 3.58) and results were presented only as final score (not change from baseline)

Seyyedi 2016.

Study characteristics
Methods	RCT, single centre, Iran January 2013 to January 2014
Participants	Inclusion: "lack of vaginal douching in the last 48 hours, refraining from using vaginal cream within the last week, 'lack of having luggage vaginal' within the last 24 hours, not having spotting, age between 50 and 65 years old, at least 12 months of amenorrhea and/or having FSH test more than 40 IU, normal Pap smear within the last 3 years, symptoms of vaginal atrophy, vaginal pH>5, having sexual activity, and naturally menopause women". Exclusion criteria: "dissatisfaction to participate in the study, prolapse grade 2 and more of bladder and uterus, surgery of anterior posterior repair (APR), history of sexual problems, chronic or systemic disease (e.g., diabetes, personal and family [first‐degree family] history of breast or endometrial cancer, history of endometriosis, personal and family [first‐degree family] history of thromboembolism, optic nerve disease, chronic liver disease, migraine, epilepsy and hereditary dyslipidemia), use of hormones, sensitivity to estrogenic compounds, allergy to honey and its products and divorced women, non‐normal TVS"
Interventions	Intervention: vaginal estrogen 0.625 mg (Premarin, Aburaihan Pharmaceutical Company) Control: matching placebo (KY Jelly, Aburaihan Pharmaceutical Company)
Outcomes	MENQOL, sexual function domain score
Notes	Age: intervention (49.44 ± 2.52 years), control (49.16 ± 3.8 years) BMI: intervention (26.97 ± 4.16 kg/m²), control (29.43 ± 3.95 kg/m²) Follow‐up: 12 weeks Do not report conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random allocation sequence: the women were randomly assigned in this clinical trial. The women were randomly distributed into three groups of 30 royal jelly vaginal cream 15%, vaginal Premarin, and placebo (lubricant), for three months
Allocation concealment (selection bias)	Low risk	To conceal allocation, all the tubes were prepared the same and were named A, B, C. Those who entered the study randomly received the creams.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for missing data not related to outcome. Missing data balanced across groups, and reasons similar.
Selective reporting (reporting bias)	Unclear risk	Protocol not available but it is clear that all prespecified and expected outcomes of interest are reported
Other bias	Low risk	Study appears to be free of other sources of risk

Simon 2007.

Study characteristics
Methods	RCT, multicentre: 28 sites in USA and Canada
Participants	Inclusion: healthy women aged ≥ 18 years who had undergone natural (amenorrhoea for ≥ 12 months) or surgical menopause (bilateral oophorectomy with or without hysterectomy ≥ 6 months before screening), who had serum E2 < 20 pg/mL, FSH > 40 mUI/mL, and BMI of 18‐35 kg/m² Exclusion: history of estrogen‐dependent neoplasia; endometrial hyperplasia; active hepatic, gallbladder, renal, or endocrine disease other than controlled thyroid abnormalities; or if they were receiving concomitant medications that could potentially interfere with hot flush frequency, severity, or their assessment
Interventions	Intervention 1: estradiol gel 0.87 g/d transdermal Intervention 2: estradiol gel 1.7 g/d transdermal Intervention 3: estradiol gel 2.6 g/d transdermal. This regimen was discontinued from further enrolment in the middle of the study because FDA advised so, in light of WHI results Control: placebo
Outcomes	Utian Quality of Life Change from baseline scores
Notes	Follow‐up: 12 weeks Age: intervention 1 (54.4 ± 6.3 years); intervention 2 (53.9 ± 6.2 years); intervention 3 (55.3 ± 8.5 years); control group (54.4 ± 5.8 years); range 28‐74 BMI: intervention 1 (26.4 ± 4.0 kg/m²); intervention 2 (26.2 ± 3.8 kg/m²); intervention 3 (26.6 ± 3.6 kg/m²); control (25.8 ± 3.8 kg/m²) As intervention 3 was considered unsafe in the middle of the study and is not used in clinical practice, we did not include this group in the analysis. For sexual function only data from intervention 2 (1.7 g/d) could be retrieved, data from intervention 1 (0.87 g/d) was omitted from the report. For vaginal maturation and pain outcomes, we combined data from intervention groups 1 and 2. This study was supported by a grant From BioSante Pharmaceuticals, Inc. (Lincolnshire, IL). Drs Bouchard, Simon, Waldbaum, and Utian have participated in clinical research as principal investigators for estradiol gel supported by BioSante Pharmaceuticals. They received payment under a research contract for the conduct of the study. Dr Simon was a consultant on this research and was paid by BioSante Pharmaceuticals. Ms Zborowski and Dr Snabes are paid independent consultants for BioSante Pharmaceuticals. Writing support was paid for by BioSante Pharmaceuticals.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomization list
Allocation concealment (selection bias)	Unclear risk	Study co‐ordinators assigned treatment numbers to participants entering the 12‐week double‐blind treatment period
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, investigators, and outcome assessors were blinded to treatment assignment, and no unblinding occurred during the trial
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Participants, investigators, and outcome assessors were blinded to treatment assignment, and no blinding occurred during the trial
Incomplete outcome data (attrition bias) All outcomes	High risk	The assessment of sexual function is only reported for the higher doses, 1.7 g/d and 2.6 g/d. The intervention group of 2.6 g/d discontinued enrolment in the middle of the study because of safety issues. The results for sexual function for the lower‐dose group "E2 gel 0.87 G/day" is omitted.
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

SMART‐3.

Study characteristics
Methods	RCT, multi‐centre, 66 sites in USA
Participants	Inclusion: "healthy women with an intact uterus, aged 40‐65 years, BMI ≤ 34.0 kg/m², postmenopausal ‐ defined as ≥ 12 months of spontaneous amenorrhoea or ≥ 6 months of spontaneous amenorrhoea with FSH > 40 mIU/mL. At screening, all women had to have a vaginal cytological smear showing ≤ 5% superficial cells and vaginal pH > 5 and had to identify on the symptom questionnaire at least 1 moderate‐severe vulvar/vaginal symptom that was most bothersome to them ‐ either vaginal dryness, irritation/itching, or pain with intercourse". Exclusion: "history of or clinically active medical disease such as endometrial hyperplasia; known or suspected estrogen dependent neoplasia; undiagnosed vaginal bleeding; chronic renal or hepatic disease; thrombophlebitis, thrombosis, or thromboembolic disorders; cerebrovascular accident; neuro‐ocular disorders; myocardial infarction or ischaemic heart disease; gallbladder disease; malignancy or treatment for malignancy within the previous 5 years (with the exception of carcinoma of the skin); or history of breast cancer, melanoma, or any gynaecologic cancer at any time. In addition, women with a transvaginal ultrasound screening that could not measure endometrial thickness or indicating double‐walled endometrial thickness > 4 mm, focal endometrial abnormality (other than fluid), complex ovarian cyst of any diameter, or simple ovarian cyst > 20 mm were excluded. Women who were hypersensitive to estrogen, had an active endocrine disease, had a known alcohol or drug abuse, were heavy smokers (> 15 cigarettes/d), had unresolved or abnormal cervical cytologic smear reports, had unresolved findings suspicious for malignancy on breast examination, and had uncontrolled hypertension were also excluded. All women should have had a normal endometrial biopsy at baseline."
Interventions	Wash out: "use of oral estrogen‐, progestin‐, androgen‐, or SERM‐containing drugs, transdermal hormone products, or intrauterine progestins within 8 weeks before screening; vaginal hormone products within 4 weeks; progestin implants/injectables or estrogen pellets/injectables within 6 months; intrauterine device within 12 weeks; or any investigational drug within 60 days" as these conditions may interfere with the basal dosage of endogenous hormones." Intervention 1: bazedoxifene 20 mg + conjugated estrogens 0.45 mg daily Intervention 2: bazedoxifene 20 mg + conjugated estrogens 0.625 mg Intervention 3: bazedoxifene 20 mg Control: placebo
Outcomes	Arizona Sexual Experiences (ASEX) scale domains and MENQOL
Notes	Follow‐up: 12 weeks Mean age in each group: 56.4 (± 4.7); 56.3 (± 4.4); 56.4 (± 4.5); 56.1 (± 4.2) BMI: intervention 1 (25.4 ± 3.8); 2 (25.2 ± 3.8); 3 (25.3 ± 3.9); and placebo (25.7 ± 4.1) We merged intervention groups 1 and 2 for analysis and calculated SDs from the standard errors using the calculator in RevMan. Wyeth Research, Collegeville, PA, sponsored the study and supported the medical writing assistance of Kathleen Ohleth PhD, and Chastity Bradley PhD. "G Bachmann has served as a Research Consultant‐PI for Wyeth, Bayer, GlaxoSmithKline, Duramed, Novartis, Pfizer, Boehringer‐Ingelheim, Johnson & Johnson, Roche, Boston Scientific, Novo Nordisk, Procter & Gamble, Merck, Xanodyne, Hormos, and Femme Pharma. J Bobula and S Mirkin are current employees of Wyeth. Dr Risa Kagan received research grants from Proctor & Gamble, Lilly, Boehringer Ingelheim, Depomed, and Wyeth. Dr Kagan is a consultant for Proctor & Gamble, Lilly, Medtronic, and Wyeth and is a speaker for GlaxoSmithKline, Roche, Novogyne, Novartis, and Lilly. Dr Williams received research grants from Wyeth, Xanodyne, and Organon and is also a speaker for Wyeth. Drs Mirkin and Pickar and Kaijie Pan are full‐time employees of Wyeth Research, Collegeville, PA."
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized randomization/enrolment system generated by Wyeth Research
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It is not clear how many women in fact answered the final questionnaires, and if the women who withdrew from treatment had the final evaluation
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Speroff 2003.

Study characteristics
Methods	RCT, multicentre, 35 sites in USA
Participants	Inclusion: postmenopausal woman, with or without a uterus, with at least 7 moderate‐severe hot flushes per day or an average of at least 56 moderate‐severe vasomotor symptoms per week for the 2 weeks before randomization. In addition, a woman with a uterus was required to have had amenorrhoea > 12 months before randomization; if amenorrhoea < 12 months and > 6 months, she was also required to have a FSH > 40 IU and an E2 < 20 pg/mL. A woman was also eligible if she had a hysterectomy and bilateral oophorectomy performed 6 weeks before randomization. Finally, a woman who had a hysterectomy without bilateral oophorectomy was eligible if she had FSH > 40 IU and an E2 < 20 pg/mL. Exclusion: past or current thromboembolic disorder, or cerebrovascular accident; endometriosis; allergy or intolerance to previous estrogen therapy only or HT, including disabling breakthrough bleeding; past or current estrogen‐dependent neoplasia; abnormal uninvestigated vaginal bleeding within 6 months of randomization; and known or suspected pregnancy. Previous treatment with any of the following was also reason for exclusion: estrogen, progestogen, androgen, or systemic corticosteroids by the oral route within 8 weeks of screening, by transdermal or buccal delivery within 4 weeks of screening, or by injection within 6 months of screening; hormone pellets or implants inserted within the previous 5 years or an implant removed within the past 3 months; unopposed estrogen therapy only for ≥ 6 months in women with an intact uterus; or selective estrogen receptor modulators within 8 weeks of screening.
Interventions	Intervention: vaginal ring (Femring, UK) delivering estradiol 50 mcg/d (n = 113) or 100 mcg/d (n = 112). The ring contains E2 acetate in a reservoir system within the ring polymer, which is released steadily over 3 months Control: matching placebo (n = 108)
Outcomes	Greene Climacteric Scale Change from baseline scores
Notes	Follow‐up: 13 weeks Mean age: 51.7 (± 7.2) years BMI 28 kg/m² SD calculated from P values of the "changes from baseline" values between the groups There were no differences on 'vaginal dryness' and 'urinary leakage' between groups, but there were not sufficient details to pool data. Not enough data to pool the outcome 'vaginal maturation'; the study authors claim it was significantly improved with the treatment. This study was supported by Warner Chilcott, a division of Galen Holdings PLC, which has developed this product. Research support was provided for costs of the study. Study authors have received speaking and consulting honoraria from the company. Dr Speroff owns stock in the company.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated schedule
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Few and balanced dropouts
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Strickler 2000.

Study characteristics
Methods	RCT, multicentre (USA), 32 sites
Participants	Inclusion: 47‐60 years, 2‐8 years after final menstrual period, serum E2 < 73 pmol/L (< 20 pg/mL), and lumbar spine bone mineral density 2‐2.5 SD, the mean peak lumbar spine bone mineral density of premenopausal women. Exclusion: intolerable menopausal symptoms requiring therapy; uterine bleeding of unknown cause; BMI < 18 kg/m² or > 31 kg/m², history of deep venous thromboembolic disease; use of corticosteroids, estrogen, or progestogen within 6 months before beginning the study; and chronic illness
Interventions	Washout: 6 months (because of the influence of estrogen on some subscales of WHQ assessing physical capacity that indirectly is related to bone health status) Intervention 1: CEE 0.625 mg/d Intervention 2: raloxifene 60 mg/d (n = 97) or 150 mg/d (n = 100) Control: placebo
Outcomes	WHQ Change from baseline scores
Notes	Age: 54.7 ± 3.6 years BMI: intervention groups 25.9 ± 3.3kg/m², 25.5 ± 3.5kg/m², 25.4 ± 3.4kg/m²; control group 25.5 ± 3.5kg/m² Follow‐up: 12 months SD calculated from P values Sponsored study: the research was supported by Eli Lilly and Company, maker of raloxifene. Doctors Shen, Wong, and Silfen are employees of and hold stock in Eli Lilly and Company. Doctors Strickler, Stovall, and Merritt have received honoraria for their services to Eli Lilly and Company.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were assigned randomly in blocks of four to 1 of 4 therapy groups
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, placebo was similar to the drug
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT approach on data from women who had at least 1 follow‐up visit after randomization. Missing values were replaced with values from a previous post‐randomization visit (last observation carried forward)
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

Taylor 2017.

Study characteristics
Methods	RCT, double‐blinded, placebo‐controlled
Participants	Inclusion: "women between 42 and 58 years of age and at least 6 months and no more than 36 months from last menstrual period, with plasma follicle stimulating hormone level at least 35 mIU/mL (to convert to IU per liter, multiply by 1.0) and/or E2 levels less than 40 pg/mL". Exclusion: use of HT or supplement, endometrial thickness > 5 mm, current smoking > 10 cigarettes/d, bone mass index > 35 kg/m2, history of clinical CVD, history of cerebrovascular and/or thromboembolic disease, coronary calcium score ≥ 50 units, dyslipidemia‐LDL cholesterol > 190 mg/dL, hypertriglyceridemia‐triglycerides > 400 mg/dL, lipid‐lowering medication, nut allergy, uncontrolled hypertension‐systolic, hysterectomy, history of, or prevalent, chronic diseases including any cancer (other than basal cell skin cancers), renal failure, cirrhosis, diabetes mellitus, and endocrinopathies other than adequately treated thyroid disease, known HIV infection and/or medications for HIV infection results of any safety laboratory test chemistries, inclusion criteria 20% abnormal
Interventions	727 participants enroled in the parent trial; of these 670 agreed to participate in the sexual function ancillary study (n = 209 in the o‐CEE group, n = 204 in the t‐E2 group, and n = 257 in the placebo group). Intervention: "oral o‐CEE, 0.45 mg/d and placebo patches with micronized progesterone, 200 mg, for 12 days each month; t‐E2, 50 μg/d, and placebo pills with micronized progesterone, 200 mg, for 12 days each month”.
Outcomes	Primary: rate of change of carotid intimal medial thickness by ultrasound. Secondary: change in coronary calcium score by X‐ray tomography, plasma lipid profiles, blood clotting factors, serum inflammatory factors, hormone levels, cognitive and affective scores on standard psychometric tests, and QoL FSFI
Notes	ClinicalTrials.gov Identifier: NCT00154180 We failed to obtain data from 12 weeks of follow‐up
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number table
Allocation concealment (selection bias)	Low risk	Hidden allocation sequence
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding, and unlikely that the blinding could have been broken
Blinding of outcome assessment (detection bias) All outcomes	Low risk	There is no suspicion that the blinding has been broken
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing data balanced across groups, and reasons similar
Selective reporting (reporting bias)	Low risk	Protocol is available
Other bias	Low risk	Study appears to be free of other sources of risk

Veerus 2006.

Study characteristics
Methods	RCT, parallel‐group, 3 clinical centres in Estonia 1001 women randomized January 1999‐December 2001
Participants	Inclusion: aged 50–64 years, postmenopausal Exclusion: none Potentially eligible women were randomized into 4 trial groups: blind HT group blind placebo group non‐blind HT group non‐treatment group
Interventions	Intervention: CEE 0.625 mg + MPA 5.0 mg daily (2 groups: open and blind) Control: matching placebo in the blind arm and no intervention in the open arm
Outcomes	WHQ Percentage of women satisfied and who lost interest in sex
Notes	Follow‐up: 1 year for the included evaluation (3.6 years total) Trial register: ISRCTN35338757 Potentially eligible women were randomized into 4 trial groups resulting in two substudies blind HT group vs blind placebo group (Estonian blind trial) non‐blind HT group vs non‐treatment group (Estonian non‐blind trial) 1001 women were recruited into the non‐blind subtrial and 777 in the blind subtrial. The Estonian Postmenopausal Hormone Therapy trial was originally planned to be a part of the Women’s International Study of Long Duration Oestrogen After Menopause (WISDOM) The study was financially supported by grants from the Academy of Finland (grants 69838, 201490, 115088), STAKES (National Research and Development Centre for Welfare and Health), Finland, and the Estonian Ministry of Education and Research Drugs were donated by Wyeth Ayerst Company via the Women's International Study of Long Duration Oestrogen After Menopause (WISDOM) in the UK
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization in permuted blocks, each of 16 and each block of the 3 clinics separately, at the National Research and Development Centre for Welfare and Health in Finland
Allocation concealment (selection bias)	Low risk	The treatment allocation was enclosed in a non‐transparent sealed envelope and sent to trial clinics
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open study compared with no treatment
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Balanced dropouts Blind trial ‐ intervention 98/404; control 101/372 Open trial ‐ intervention 130/494; control 125/507
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	Not suspected

WHI 1999.

Study characteristics
Methods	RCT, multicentre (40 clinical centres), USA
Participants	Inclusion criteria: women aged 50‐79 years with either an intact uterus or with hysterectomy 3 months earlier (with or without oophorectomy), and were likely to reside in the area for 3 years. Exclusion criteria: medical condition predicted survival of < 3 years or if there were diagnoses of previous breast cancer or melanoma, other cancer within the past 10 years (except non‐melanoma skin cancer), low haematocrit or platelet counts, or any condition that would interfere with acceptable adherence and retention (e.g. alcoholism or dementia). Failing the 4‐week compliance schedule of placebo pill taking. Potential participants who were using postmenopausal hormones at the initial screening visit were required to undergo a 3‐month 'washout' period before enrolment. Women who reported moderate or severe menopausal symptoms during the washout period were discouraged from participating in the study but were not excluded.
Interventions	Intervention 1: CEE 0.625 mg, (Premarin; Wyeth, St Davids, Pa) orally, daily Intervention 2: CEE 0.625 mg + MPA 2.5 mg (Prempro, Wyeth) Control: matching placebo pills
Outcomes	Modified Mini‐Mental State Examination (QoL): sexual satisfaction (range from 1/worst to 4/best) Change from baseline scores
Notes	Part of WHI study; included data from 1‐year follow‐up period Intervention: 5310 women allocated Control: 5429 women allocated Funding: WHI is funded by the National Heart, Lung, and Blood Institute, US Department of Health and Human Services, National Institutes of Health (NIH), Bethesda, MD, USA. Drs Shumaker and LaCroix report having served as paid consultants to Pfizer. Dr Shumaker reports having served as a paid lecturer for Wyeth–Ayerst and Pfizer. Drs Hays, Brunner, and Shumaker report having received grant support from Wyeth–Ayerst. Dr Shumaker reports having received grant support from Pfizer. Dr Brzyski reports having received grant support from Berlex.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomized permuted block algorithm, stratified by clinical centre site and age group
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind. All study medication bottles had a unique bottle number and bar code to allow for blinded dispensing
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias) All outcomes	High risk	About half of the allocated participants were evaluated for sexual function
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	None

Wiklund 1993.

Study characteristics
Methods	RCT, multicentre, 15 sites, Sweden
Participants	Inclusion: postmenopausal women between 45 and 65 years old requiring hormone replacement therapy for climacteric symptoms, amenorrhoea > 6 months, no hormone replacement therapy within the last 6 months. Exclusion: surgically induced menopause, previous or current estrogen‐dependent tumours, other current malignant or life‐threatening disease, severe metabolic, endocrine, or gastrointestinal disease, concomitant heart disease, insulin‐treated diabetes, uncontrolled hypertension, endometriosis, undiagnosed vaginal bleeding, active skin disease, and unstable medical conditions such as rheumatoid arthritis or chronic obstructive lung disease. Women with psychiatric disorders and/or those receiving continuous tranquillizer or antidepressant therapy.
Interventions	Washout: 6 months Intervention: transdermal estradiol therapy, 50mcg/24 hours, changed twice a week Control: matching placebo
Outcomes	Simplified McCoy Sex Scale ('sexual problems' was included as 'dyspareunia', because it refers to the question 'do you currently experience any pain during intercourse?') Change from baseline
Notes	Follow‐up: 12 weeks Mean age: intervention = 52.4 (±4.3) years; control = 52.9 (±4.0) years Do not report conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, matching placebo
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Described in details
Incomplete outcome data (attrition bias) All outcomes	Low risk	Dropouts: intervention (5), control (10)
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Unclear risk	The article does not report any conflict of interest disclosure

WISDOM 2008.

Study characteristics
Methods	RCT, multi‐centre (499 general practices): UK, Australia, New Zealand (WISDOM trial)
Participants	Inclusion: postmenopausal women aged 50‐69 years at randomization were eligible for the trial. Postmenopausal was defined as no menstrual period in the last 12 months or having had a hysterectomy. Women who had started taking HT before their menstrual periods had stopped were considered eligible when they had been taking HT for 12 months. Exclusion " For the placebo‐controlled components: oral or transdermal HT use in the last 6 months, ever used HT implant in women with a uterus, HT implant inserted in last 8 months in women with a hysterectomy History of endometriosis (hysterectomized or non‐hysterectomized woman) or endometrial hyperplasia in a woman with a uterus History of invasive breast cancer, lobular carcinoma in‐situ (LCIS), ductal carcinoma in‐situ (DCIS), Paget's disease of the nipple or atypical hyperplasia of the breast Known BRCA1 or BRCA2 mutation carrier History of melanoma Invasive cancer at any other site apart from basal and squamous cell skin cancer within the last 10 years History of meningioma Myocardial infarction, cerebrovascular accident, sub‐arachnoid haemorrhage or transient ischaemic attack within the last 6 months History of currently active liver disease or chronic liver disease but excluding Hepatitis A unless currently active Evidence of severe current renal impairment History of gall bladder disease in a woman who had not had a cholecystectomy or of gallstones following a cholecystectomy History of deep vein thrombosis, pulmonary embolism or retinal vein occlusion Positive thrombophilia screen (Factor V Leiden or prothrombin mutations, Protein C, Protein S or anti‐thrombin III deficiencies, Activated Protein C resistance, dysfibrinogenaemia or antiphospholipid antibodies) Otosclerosis Porphyria History of hepatitis B, hepatitis C or HIV (not an exclusion in New Zealand) Currently pregnant Currently taking or has taken contraceptive drugs in the last 12 months Current triglyceride level (fasting) > 5.5 mmol/L Active participant in any other intervention trial likely to affect trial outcomes Taking tamoxifen, toremifene, raloxifene or any other SERM Other conditions/circumstances where the general practitioner judged that it would not be possible to obtain fully informed consent and/or successfully complete trial procedures"
Interventions	Intervention: oral CEE 0.625 mg daily, plus oral MPA 2.5/5.0 mg daily (Prempro, Wyeth Ayerst, USA) Control: matching placebo
Outcomes	WHQ Final scores
Notes	Follow‐up: 1 year Mean age at randomization: 63.8 years Drugs were donated by Wyeth Ayerst Company
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐based, stratified block randomization program
Allocation concealment (selection bias)	Low risk	Randomized centrally and appropriate medication dispatched to the practice
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No‐one involved with assessment of outcomes was aware of treatment assignment
Incomplete outcome data (attrition bias) All outcomes	High risk	Sexual function was evaluated in only 588/1043 participants in the intervention group and in only 569/1087 participants in the control group
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Low risk	None

Yang 2004.

Study characteristics
Methods	RCT, single centre, Taiwan
Participants	Inclusion: postmenopausal women with intact uterus and natural menopause of at least 6 months’ duration, FSH > 18 mIU/L, and estradiol level < 30 pg/mL Exclusion: history of liver disease, breast cancer, endometrial cancer, thrombophlebitis, thromboembolic disorders related to estrogen use, myocardial infarction or ischaemic heart disease, chronic renal disease, cerebrovascular accident, uncontrolled hypertension, diabetes or metabolic bone disease, or had received any hormone treatment within 1 year previously, chronic use of steroids, fluoride, Vitamin D, rifampin, barbiturates, sulfenamide, and herbal medications, or the use of estrogen and progestogen other than the study medications.
Interventions	Intervention: CEE 0.625 mg + MPA 2.5 mg orally daily for 5.5 months Control: matching placebo
Outcomes	Greene Climacteric Scale Change from baseline
Notes	Follow‐up: 5.5 months Age: intervention (53.2 ± 3.1 years); control (53.3 ± 3.4 years) Weight: intervention (56.0 ± 8.3 kg); control (55.7 ± 6.6 kg) Do not report conflict of interest
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported in detail
Allocation concealment (selection bias)	Unclear risk	Not reported in detail
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	High rates, but balanced, dropouts: intervention (6/33); control (8/33)
Selective reporting (reporting bias)	Low risk	Not suspected
Other bias	Unclear risk	We observed relevant difference between intervention and control groups at baseline (0.6 ± 0.8 vs 1.0 ± 0.9, intervention vs control respectively). However, we were able to analyse data as change from baseline. There is no conflict of interest disclosure.

AE: adverse effect; BISF‐W: Brief Index of Sexual Functioning‐Women; BMI: body mass index; BP: blood pressure; CE: conjugated estrogen; CEE: conjugated equine estrogen; CVD: cardiovascular disease; DHEA: dehydroepiandrosterone; FDA: Food and Drug Administration; FSFI: Female Sexual Function Index; FSH: follicle stimulating hormone; GAD‐7: Generalised Anxiety Disorder Assessment; GRISS: Golombok‐Rust Inventory of Sexual Satisfaction; HT: hormone therapy; ITT: intention‐to‐treat; IUD: intrauterine device; LDL: low‐density lipoprotein; MBS: most bothersome symptom; MENQOL: Menopause‐specific Quality of Life; MFSQ: McCoy Female Sexuality Questionnaire; MOS: Medical Outcomes Study; MPA: medroxyprogesterone acetate; NIH: National Institutes of Health; PHQ: Patient Health Questionnaire; QoL: quality of life; RCT: randomized controlled trial; SD: standard deviation; SERM: selective oestrogen receptor modulator; SSA‐P: Subjective Symptoms Assessment Profile; SUI: stress urinary incontinence; TVUS: transvaginal ultrasound; VSI: Vaginal Symptom Index; WHI: Women's Health Initiative; WHO: World Health Organization; WHQ: Women's Health Questionnaire

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
2010‐024603‐26	Did not evaluate sexual function
2014‐000179‐18	Did not evaluate sexual function
Abedi 2018	Did not use HT
Abedi 2020	Did not use HT
Abraham 2014	Did not evaluate sexual function
Adler 2005	Not randomized
Almeida 2006	Did not evaluate sexual function
Aptecar 2021	Did not evaluate sexual function by a validated tool
Archer 2011	Did not evaluate sexual function
Argyroudis 1997	Not randomized
Arêas 2019	Did not use HT
Bachmann 2008	Did not evaluate sexual function
Bachmann 2009	Did not evaluate sexual function
Bachmann 2010	Did not evaluate sexual function
Bachmann 2011	Did not evaluate sexual function with a validated instrument
Barnabei 2002	Did not evaluate sexual function
Bennink 2016	Did not evaluate sexual function
Bianchi‐Ferraro 2022	Had another intervention for a comparator
Binder 2001	Did not evaluate sexual function
Blumel 2008	Addition of testosterone to estrogen/progesterone in the study group
Bosak 2022	Did not evaluate sexual function
Bride 2014	Did not evaluate sexual function
Bruno 2012	Pseudo‐randomized
Cano 2012	Did not evaluate sexual function
Cardozo 2001	Did not evaluate sexual function
Casper 1999	Did not evaluate sexual function
Castelo‐Branco 2000	Pseudo‐randomized
Cayan 2008	Pseudo‐randomized
Ceausu 2011	Did not evaluate sexual function
Concepcion 2011	Did not evaluate sexual function
Cruz 2018	Comparator is another intervention (does not use placebo or no intervention)
Dalil 2020	Did not use HT
Daneshmand 2015	Did not evaluate sexual function
Dennerstein 1980	Did not evaluate sexual function
Dessole 2004	Did not evaluate sexual function
Diem 2006	Did not evaluate sexual function
Ditkoff 1991	Did not evaluate sexual function
DRKS00006888	Did not evaluate sexual function
Duka 2000	Did not evaluate sexual function
Eftekhar 2009	Not randomized, compares before and after treatment
Eftekhar 2020	Control group used Premarin
Eriksen 1992	Did not evaluate sexual function
Eriksen 1999	Did not evaluate sexual function
Fabian 2019	Did not evaluate sexual function
Fernandes 2014	Comparator is another intervention (does not use placebo or no intervention)
Freedman 2009	Did not evaluate sexual function
Galesanu 2010	Did not evaluate sexual function
Galhardo 2006	Did not evaluate sexual function by a validated questionnaire
Garcia de Arriba 2022	Did not evaluate sexual function
Gaspard 2021	Did not evaluate sexual function
Gonenne 2006	Did not evaluate sexual function
Gorenstein 2011	Did not evaluate sexual function
Griesser 2012	Did not evaluate sexual function
Guan 2010	Did not evaluate sexual function
Hachul 2008	Did not evaluate sexual function
Harman 2014	Did not evaluate sexual function
Hedrick 2009	Did not evaluate sexual function
Hendrix 2005	Did not evaluate sexual function
Hodis 2001	Did not evaluate sexual function
Hofling 2005	Did not evaluate sexual function
IRCT138809032766N1	Did not evaluate sexual function
IRCT2017050433817N1	Did not evaluate sexual function
Islam 2021	Were not or pseudo‐randomized
ISRCTN46661996	Did not evaluate sexual function
Jensen 1983	Did not evaluate sexual function
Ju 2021	Were not or pseudo‐randomized
Kagan 2014	Did not evaluate sexual function
Kenny 2005	Did not evaluate sexual function
Kessel 2003	No group without HT
Kimura 1995	Not randomized
Kingsberg 2017	Did not evaluate sexual function
Kocoska‐Maras 2011	Did not evaluate sexual function
Kornafel 1992	Did not evaluate sexual function
Kroll 2018	Did not evaluate sexual function
Krychman 2016	Did not evaluate sexual function
Kumru 2008	Did not evaluate sexual function
L'Hermite 2022	Did not evaluate sexual function
Lin 2011	Did not evaluate sexual function
Lindsay 2002	Did not evaluate sexual function
Liu 2012	Did not evaluate sexual function
Lobo 2009	No group without HT
Lobo 2017	Did not evaluate sexual function
MacDonald 1994	Did not evaluate sexual function
Marschalek 2022	Were a different population
Marsden 2000	Did not evaluate sexual function
Marx 2004	Did not evaluate sexual function
Miller 2019	Did not evaluate sexual function
Mittal 2020	Did not evaluate sexual function by a validated tool
Mizunuma 2011	Did not evaluate sexual function
Naessen 2001	Did not evaluate sexual function
NCT01543399	Did not evaluate sexual function
NCT01585558	Did not evaluate sexual function
NCT01779947	Did not evaluate sexual function
NCT01822288	It is not a randomized controlled study (RCT)
NCT02253173	Did not evaluate sexual function
NCT02638337	Did not evaluate sexual function
NCT03018106	Not randomized
Nevinny‐Stickel 1983	Did not evaluate sexual function by a validated questionnaire
Nieto 2011	Did not evaluate sexual function
Noguchi 2018	Did not evaluate sexual function
Notelovitz 2000	Did not evaluate sexual function
Obel 1993	Did not evaluate sexual function
Palacios 1995	Control group used a different intervention, did not use a validated questionnaire
Palacios 2005	Did not evaluate sexual function
Paraiso 2020	Did not evaluate sexual function
Pinkerton 2014	No positive response after contacting the authors
Poetry 2011	Did not evaluate sexual function
Polo‐Kantola 1998	Did not evaluate sexual function
Portman 2013	Did not evaluate sexual function
Portman 2014a	Did not evaluate sexual function
Pradeep 2016	Did not evaluate sexual function
Prestwood 2003	Did not evaluate sexual function
Raz 2014	Without HT
RBR‐94dx93/1	Without HT
Rebar 2000	A questionnaire that includes assessment of sexual function is applied; however, the study authors do not report the results for the sexual domains
Reindollar 2002	Did not evaluate sexual function
Ribom 2011	Did not evaluate sexual function
Rigano 2001	Did not evaluate sexual function by a validated questionnaire
Russu 2011	Did not evaluate sexual function
Saeideh 2010	Did not evaluate sexual function by a validated questionnaire
Saletu 2002	Did not evaluate sexual function
Samaras 1999	Did not evaluate sexual function by a validated questionnaire
Santofimia 2022	Did not evaluate sexual function
Santoro 2012	Did not evaluate sexual function
Serebani 2020	Did not use HT
Shah 2006	Did not evaluate sexual function
Sherman 2003	Did not evaluate sexual function
Shulman 2002	Did not evaluate sexual function
Silverman 2012	Did not evaluate sexual function
Simon 1999	Did not evaluate sexual function
Simon 2008a	Did not evaluate sexual function
Simon 2008b	Did not evaluate sexual function
Simon 2012	Did not evaluate sexual function
Simunic 2003	Did not evaluate sexual function
SMART‐5	Did not evaluate sexual function
Sorensen 2001	Did not evaluate sexual function
Speroff 2006	Did not evaluate sexual function
Stening 2011	Did not evaluate sexual function
Stevenson 2010	Did not evaluate sexual function
Sun 2016	Did not evaluate sexual function
Swanson 2006	Did not evaluate sexual function
Tansupswatdikul 2012	Dis not evaluate sexual function
Tinelli 2002	Did not evaluate sexual function by a validated questionnaire
Tungmunsakulchai 2015	Control group included an intervention (not or pseudo‐randomized)
Tuomikoski 2010	Did not evaluate sexual function
Usall 2011	Did not evaluate sexual function
Utian 2001	Did not evaluate sexual function
Utian 2004	Did not evaluate sexual function
Vardy 2003	Did not evaluate sexual function
Veldhuis 2011a	Did not evaluate sexual function
Veldhuis 2011b	Did not evaluate sexual function
Vestergaard 2003	Did not evaluate sexual function
Voipio 2002	Did not evaluate sexual function
Warton 2014	No response to author contact
Wolff 1982	Did not evaluate sexual function
Woo 2003	Did not evaluate sexual function
Yeo 2020	Did not use HT
Zervoudis 2009	Did not evaluate sexual function
Zhou 2011	Did not evaluate sexual function
Zullo 2005	Did not evaluate sexual function

HT: hormone therapy

Characteristics of studies awaiting classification [ordered by study ID]

Hirschberg 2020.

Methods	Randomized, double‐blind, placebo‐controlled, international, multicenter study
Participants	Inclusion criteria Postmenopausal women with hormone receptor‐positive (and any HER2 status) early breast cancer (stage I‐IIIA) who had been treated with NSAI (i.e. either anastrozole or letrozole) for at least 6 months. Exclusion criteria Women treated with another antitumoural therapy and those who received vulvovaginal treatment in the 15 days before the study start, and/or had used hormones, natural (phytoestrogens) or herbal products to treat menopausal symptoms in the 3 months before the study start.
Interventions	Intervention: 0.005% estriol vaginal gel Control: placebo (moisturizing gel)
Outcomes	FSFI score
Notes	The study authors do not report how study was funded

NCT01845649.

Methods	RCT, multicentre in 57 locations, USA
Participants	Inclusion criteria Sexually active with self‐identified dyspareunia (pain with sexual activity), at least moderate to severe and most bothersome symptom of VVA Postmenopausal and meets 1 of the following: 12 months spontaneous amenorrhea, 6 months spontaneous amenorrhea with FSH > 40 mIU/mL, 6 weeks postsurgical bilateral oophorectomy with confirmation by FSH > 40 mIU/mL, surgical report or ultrasound, 6 weeks postsurgical hysterectomy with ovary failure confirmed by FSH > 40 mIU/mL Age ≥ 40 years; ≥ 35 with bilateral oophorectomy Vaginal pH > 5.0 ≤ 5% superficial cells on vaginal wall cytologic smear Normal clinical breast exam or negative mammogram if ≥ 40 years of age Negative urine pregnancy test (non‐hysterectomized and < 12 months amenorrhea) Exclusion criteria Enrollment in Sponsor's Study PR‐04409 or PR‐05812 Participation in clinical trial or use of investigational drug within 30 days prior to screening Known hypersensitivity to estrogen and/or progesterone therapy Known or suspected premalignant or malignant disease or history steroid‐dependent malignancy Manifestation or treatment for significant CVD (congestive heart failure, stroke or ischemic attack, myocardial infarction, coronary artery bypass, percutaneous angioplasty or > 50% angiographic narrowing of coronary artery) Thrombophlebitis or thromboembolic disorder or history of Insulin‐dependent diabetes mellitus Increased frequency or severity of headaches while on hormone or estrogen therapy Currently taking St. John's Wort Drug/alcohol addiction within past 2 years Treatment with anticoagulants (heparin or warfarin) Smoking ≥ 15 cigarettes/day
Interventions	Estradiol vaginal cream applied daily for 14 days followed by dosing 3 times per week for 10 weeks
Outcomes	Self‐assessment of the symptoms of VVA, other than dyspareunia and vaginal bleeding associated with sexual activity
Notes	Last ipdate posted date: 9 May 2022

NCT02460302.

Methods	RCT, single centre, Mount Sinai Hospital, Toronto, Canada
Participants	Inclusion criteria Age ≥ 45, unless the patient has had a bilateral oophorectomy Amenorrheic for > 1 year and/or FSH > 40 IU/mL Women who have had a bilateral oophorectomy will be eligible at any age Patients must have ≥ 2 urogenital atrophy symptoms: vaginal dryness vaginal irritation vaginal soreness dyspareunia Exclusion criteria Age < 45 Systemic or local HT use within 3 months of entry into study Known or suspected history of breast cancer Hormone‐dependent tumour Genital bleeding of unknown cause Current vaginal infection requiring treatment Known allergy to test constituents Serious disease or chronic condition that could interfere with compliance Active thrombophlebitis, or history of a hormone‐associated thrombophlebitis Kidney disease Liver dysfunction or disease High LDL levels
Interventions	Micronized vaginal progesterone 100 mg inserted vaginally, or placebo suppository inserted vaginally
Outcomes	Improvement in patient‐identified most bothersome symptom (vaginal dryness, itching, soreness, pain with sexual activity)
Notes	Study in progress

NCT02834312.

Methods	RCT, single centre, Belgium
Participants	Inclusion criteria Women presenting at least 7 moderate to severe hot flushes/day or at least 50 moderate to severe hot flushes/week in the week preceding randomization BMI between 18.0 and 35.0 kg/m², inclusive Post‐menopausal status Intact uterus Negative pregnancy test Good physical and mental health Has provided signed and dated written informed consent before admission to the study Able to understand and comply with the protocol requirements, instructions, and protocol‐stated restrictions Exclusion criteria Uterine disease or any medical conditions associated with an increase in endometrial thickness Any history of malignancy with the exception of basal cell (excluded if within the prior 2 years) or squamous cell (excluded if within the prior year) carcinoma of the skin. Any clinically significant findings at the breast examination and/or on mammography suspicious of breast malignancy that would require additional clinical testing to rule out breast cancer. Abnormal cervical Pap smear Systolic BP outside the range 90‐140 mmHg, diastolic BP outside the range 60‐90 mmHg, and/or heart rate outside the range 40‐100 beats/min Any clinically significant abnormality identified on the screening 12‐lead ECG History of venous or arterial thromboembolic disease, history of known coagulopathy or abnormal coagulation factors Diabetes mellitus with poor glycaemic control Dyslipoproteinaemia at screening Smoking > 10 cigarettes/day Presence or history of gallbladder disease, unless cholecystectomy has been performed Systemic lupus erythematosus Multiple sclerosis Acute or chronic liver disease Acute or chronic renal impairment Uncontrolled thyroid disorders Use of oestrogen or progesterone containing drug(s) Use of non‐hormonal treatments to reduce hot flushes History or presence of allergy or intolerance to any component of the investigational product History of alcohol or substance abuse or dependence in the 12 months before screening as determined by the Investigator Sponsor, CRO or Investigator's site personnel or their relatives directly affiliated with this study Women with known or suspected history of a clinically significant systemic diseases, unstable medical disorders, life‐threatening disease or current malignancies that would pose a risk to them in the opinion of the Investigator Participation in another investigational drug clinical study within 1 month (30 days) or have received an investigational drug within the last 3 months (90 days) prior to study entry Is judged by the Investigator to be unsuitable for any reason
Interventions	2.5 mg, or 5 mg, 10 mg, or 15 mg of oral estetrol
Outcomes	Change in the genitourinary symptoms (GSM) of menopause (time frame: from baseline to week 12): vaginal pain associated with sexual activity (none, mild, moderate or severe); vaginal bleeding associated with sexual activity (presence vs absence)
Notes	Study in progress

NCT03363997.

Methods	RCT, single centre, Campinas, SP, Brazil
Participants	Inclusion criteria BMI ≥ 18.5 kg/m² and ≤ 30.0 kg/m² Postmenopausal state: FSH (plasma) ≥ 40 IU/L, estradiol (serum) ≤ 20 pg/mL last spontaneous menstruation at least 12 months ago Normal transvaginal endometrial scan (endometrial thickness < 5 mm) Good state of health Non‐smoker or ex‐smoker for at least 6 months Written informed consent, after having been informed about benefits and potential risks of the clinical trial, as well as details of the insurance taken out to cover the women participating in the clinical trial Exclusion criteria Existing cardiac, hepatic and/or haematological diseases or pathological findings, which might interfere with the safety or tolerability and/or pharmacokinetics and/or pharmacodynamics of the active ingredient History of relevant CNS and/or psychiatric disorders and/or currently treated CNS and/or psychiatric disorders Known allergic reactions to the active ingredients used or to constituents of the pharmaceutical preparations Women with severe allergies or multiple drug allergies, unless it is judged as not relevant for the clinical trial by the investigator Positive anti‐HIV‐test (if positive to be verified by western blot), HBs‐AG‐test (if positive to be verified by test for HBc‐IgM) or anti‐HCV‐test Presence or history of venous or arterial thrombosis (e.g. deep venous thrombosis, pulmonary embolism) Known, past or suspected breast cancer or increased familiar risk for development Known or suspected estrogen‐dependent malignant tumours (e.g. endometrial or breast cancer) Undiagnosed genital bleeding Acute vaginal infection or other diseases prohibiting the placement of vaginal ring History of endometrial hyperplasia Migraine or frequent episodes of severe headache History of or current drug or alcohol dependence Women who are on a diet which could affect the pharmacokinetics of the active ingredient Regular intake of caffeine containing food or beverages of ≥ 500 mg caffeine per day Blood donation or other blood loss of > 400 mL within the last 3 months prior to individual enrolment of the woman Participation in a clinical trial during the last 6 months prior to individual enrolment of the woman Concomitant systemic therapy with antibiotics, which might interfere with enterohepatic recirculation (e.g. cephalosporines, neomycin, ampicillin or tetracyclines) Use of sex hormones within 6 months (oral, transdermal, vaginal) or 8 months (intramuscular administered depot preparations used once per month) or 12 months (intramuscular administered depot preparations used once per 3 months) before screening Use of systemic or topical medications or substances that oppose the study objectives or that might influence them within 8 weeks before screening examination Women who are unable to understand the written and verbal instructions, in particular regarding the risks and inconveniences they will be exposed to during their participation in the clinical trial
Interventions	Estriol (100 mg (Test 1), 300 mg (Test 2) or 600 mg (Test 3) with continuous delivery (0.125, 0.250 or 0.500 mg/day) for 21 days. And also, characterized its safety and tolerability
Outcomes	Number of adverse events, in each treatment group, including clinically relevant alterations of vital signs and laboratory tests results
Notes	Study in progress

Pinkerton 2017.

Methods	RCT, multicentre, USA and Canada
Participants	Inclusion criteria Healthy postmenopausal women Those seeking treatment for menopausal symptoms Those with moderate/severe vasomotor symptoms Postmenopausal women aged 40‐75 years
Interventions	Intervention: conjugated estrogen (CE)/bazedoxifene (BZA) (CE 0.45 or 0.625 mg with BZA 10, 20, or 40 mg), raloxifene 60 mg Control: placebo
Outcomes	Sexual activity and health‐related QoL
Notes	Pending other possible secondary results derived from the study

Tanmahasamut 2020.

Methods	Double‐blind RCT
Participants	Inclusion criteria Postmenopausal women aged ≥ 45 years with amenorrhea for 12 months or < 12 months with serum FSH level > 40 mIU/ mL and estradiol level < 30 pg/mL Or who had undergone bilateral oophorectomy with vaginal symptoms related to postmenopausal status (vaginal dryness, vaginal or vulvar irritation, vaginal soreness, vaginal discharge, and/or dyspareunia) Exclusion criteria The excluded known or suspected history of breast cancer, cardiovascular disease, liver disease, gallbladder disease, genital tract infection, abnormal uterine bleeding with unknown cause, thromboembolism, previous hormone use within 90 days, hypersensitivity to the study medication and/or any contraindication to the administration of estrogen
Interventions	Intervention group Estradiol (Estrofem) 25 μg in K‐Y Jelly 2 mL, prepared by dissolving a 1 mg estradiol tablet in 5 mL of sterile water for injection. That solution was then added to 75 mL of K‐Y Jelly and mixed well Control group Solution in the control group was prepared by adding 5 mL of sterile water for injection to 75 mL of K‐Y Jelly with thorough mixing.
Outcomes	FSFI score
Notes	This study was registered with the Thai Clinical Trials Registry (registration number 20170322001). All study authors declare no financial support from the companies that produce and/or distribute the drugs, devices or materials described in this report.

BMI: body mass index; BP: blood pressure; CNS: central nervous system; CVD: cardiovascular disease; ECG: electrocardiogram; FSFI: Female Sexual Function Index; FSH: follicle stimulating hormone; HT: hormone therapy; LDL: low‐density lipoprotein; QoL: quality of life; RCT: randomized controlled trial; VVA: vulvovaginal atrophy

Characteristics of ongoing studies [ordered by study ID]

NCT02668796.

Study name	To study generic estradiol 10 mcg vaginal tablets in the treatment of vulvar and vaginal atrophy in post menopausal women
Methods	RCT
Participants	Postmenopausal women aged > 30 to < 75 years
Interventions	Experimental: estradiol vaginal tablets 10 mcg (Glenmark) apply using the given applicator Active comparator: Vagifem® (estradiol vaginal tablets) 10 mcg (Novo Nordisk) apply using the given applicator
Outcomes	Primary: the proportion of participants with at least 25% reduction from baseline in the sum of % basal/parabasal + % intermediate cells on vaginal cytology and vaginal pH < 5.0 with a change from baseline vaginal pH of at least 0.5 (time frame: day 15) Secondary: the proportion of participants with treatment success at the end of study where treatment success is defined as a women who achieves a score of 0 (none) or 1 (mild) at visit 3/end of study for the MBS (time frame: day 15]
Starting date	January 2016
Contact information
Notes	NCT02668796

NCT02670785.

Study name	A safety and efficacy study of estradiol vaginal capsule in postmenopausal women with vulvovaginal atrophy
Methods	RCT
Participants	Clinical diagnosis of vulvovaginal atrophy due to post menopause; age ≥ 40 years or if bilateral oophorectomy ≥ 35 years; moderate‐severe vaginal dryness; normal breast exam; if > 40 years; vaginal pH > 5.0 < 5% superficial cells on vaginal wall cytology smear
Interventions	Estradiol vaginal: capsule 0.003 mg, capsule 0.01 mg, and capsule 0.02 mg
Outcomes	Primary: change in the percentage of vaginal superficial cells Secondary: change in the investigator assessment of vulvovaginal atrophy measured by scoring the vaginal health in 5 categories using a 4‐point scale change in woman's self‐assessment of vaginal dryness reported as the most bothersome symptom change in woman's self‐assessment of dyspareunia reported as the most bothersome symptom change in woman's self‐assessment of dysuria change in woman's self‐assessment of vaginal and/or vulvar irritation/itching change in woman's self‐assessment of vaginal dryness in women where vaginal dryness was not reported as the most bothersome symptom change in woman's self‐assessment of dyspareunia in women where dyspareunia was not reported as the most bothersome symptom self‐assessment by sexually active participants as to the presence or absence of vaginal bleeding
Starting date	15 February 2016
Contact information
Notes	NCT02670785

NCT02859285.

Study name	The impact of hormone therapy on lower urinary tract symptoms, sexual function and psychosomatic disorder
Methods	RCT
Participants	Women aged 40‐70 years, with intact uterus seeking medical treatment for menopausal syndrome
Interventions	Tibolone (2.5 mg/d) for 12 consecutive weeks
Outcomes	Primary: menopausal symptoms Secondary: sexual function, sexual function, psychosomatic symptoms.
Starting date	16 November 2012
Contact information
Notes	NCT01822288

NCT03116022.

Study name	Use of estriol at distal third of vagina improving coital pain in post‐menopause women
Methods	RCT
Participants	Sexually active postmenopausal women of any age with complaint of vulvar and/or vaginal atrophy who also complain of bothersome decreased libido, and/or arousal, and/or orgasmic dysfunction
Interventions	Experimental: distal estriol group (DEG). This arm is composed of women using estriol 1 mg/1g in the distal third of the vagina every other night. Active comparator: proximal estriol group (PEG). This arm is composed of women using estriol 1 mg/1g in the proximal third of vagina every other night.
Outcomes	Primary: number of women with lower score of pain assessed by the McGill Pain Questionnaire. Secondary: number of participants with risk for sexual dysfunction after estriol use number of participants with risk for anxiety after estriol use number of participants with risk for depression after estriol use reduction of degree of coital pain variation in serum levels of estriol after estriol use variation in serum levels of FSH after estriol use variation in serum levels of thyroid‐stimulating hormone after estriol use number of participants with adverse events that are related to estriol use variation in blood glucose after estriol use variation in weight after estriol use weight variation in BMI after estriol use
Starting date	1 June 2017
Contact information
Notes	NCT03116022

BMI: body mass index; FSH: follicle stimulating hormone; MBS: most bothersome symptoms; RCT: randomized controlled trial

Differences between protocol and review

The previous review version included several changes from the planned protocol.

• Primary outcome: we kept the composite score as the primary outcome to assess the global effect on sexual function. As secondary outcomes, we evaluated the individual domains of sexual function.

• Secondary outcomes: we previously planned to evaluate the genitourinary system, quality of life, depression, and body mass index. However, evidence for the planned secondary outcomes would just be a subset of the overall body of evidence because the present review only includes studies that reported sexual function. In this way, we evaluate the domains of sexual function (desire, arousal, lubrication, orgasm, satisfaction, and pain).

• Structure of data synthesis: we did not plan stratification of data, however, the results were confusing and we therefore stratified by participant characteristics.

• We performed sensitivity analysis only by excluding studies with a high risk of bias. We were not able to perform sensitivity analysis using alternative imputation strategies as initially planned.

• We have added two summary of findings tables to present the results.

• There were not enough studies to perform all planned subgroup analyses by the type of menopause and sexual symptom status at baseline.

At the 2022 update we made the following changes:

• In the previous version, all definitions of female sexual function were established based on Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM‐IV), and we used the North American Menopause Society (NAMS)‐2012 guideline to support hormone therapy definitions. We have updated the current review according to DSM‐V and International Classification of Disease (ICD)‐11 criteria for female sexual dysfunction diagnosis. From the six domains of sexual function (desire; arousal, lubrication, orgasm, satisfaction, and pain) we made the following adjustments. Following the latest recommendations of DSM‐V and ICD‐11, we evaluated the frequency of sexual activity, libido, desire, or interest in sex in the domain of 'desire'; we evaluated arousal and fantasy in the domain of 'arousal'; we evaluated orgasm and the number of orgasms in the domain of 'orgasm'; we evaluated pleasure or enjoyment in sex and satisfaction in the 'satisfaction' domain; we evaluated dyspareunia in the 'pain' domain, and we incorporated an assessment of the lubrication domain.

• We stratified analyses by time points in effect measurements and clinically reasonable time points on effect measurements from available studies (between 12 and 16 weeks; between 22 and 26 weeks).

Contributions of authors

Drafting the protocol: Lucia AS Lara, Denisse Cartagena‐Ramos, Wellington P Martins, Jaqueline BP Figueiredo, Ana Carolina JS Rosa‐e‐Silva, Rui A Ferriani helped, and Miguel Fuentealba‐Torres

Development of search strategy: Lucia AS Lara, Denisse Cartagena‐Ramos, Wellington P Martins and Miguel Fuenteallba‐Tores

Search for trials: Lucia AS Lara, Denisse Cartagena‐Ramos, Jaqueline BP Figueiredo, Wellington P Martins and Miguel Fuenteallba‐Tores

Obtaining copies of trials: Lucia AS Lara, Denisse Cartagena‐Ramos, Jaqueline BP Figueiredo, Wellington P Martins, and Miguel Fuenteallba‐Tores

Contact authors from trials: Lucia AS Lara, Denisse Cartagena‐Ramos, and Miguel Fuenteallba‐Tores

Selection of which trials to include: Lucia AS Lara, Denisse Cartagena‐Ramos, Jaqueline BP Figueiredo, Jaqueline BP Figueiredo, Lucia AS Lara and Miguel Fuenteallba‐Tores

Extraction of data from trials: Lucia AS Lara, Denisse Cartagena and Miguel Fuenteallba‐Tores

Assessment of risk of bias in included studies: Lucia AS Lara, Denisse Cartagena‐Ramos, Wellington P Martins and Miguel Fuenteallba‐Tores

Entry of data into Review Manager: Lucia AS Lara, Wellington P Martins and Miguel Fuenteallba‐Tores

Drafting the review: Lucia AS Lara, Denisse Cartagena‐Ramos, Wellington P Martins, Ana Carolina JS Rosa‐e‐Silva, Jaqueline BP Figueiredo, Rui A Ferriani helped and Miguel Fuenteallba‐Tores

Sources of support

Internal sources

• Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo (FMRP‐USP), Brazil

  Authors' salary

• Facultad de Enfermería y Obtetricia, Universidad de los Andes, Chile, Chile

  Authors' salary

External sources

• Fundação de Apoio ao Ensino, Pesquisa e Assistência do Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo (FAEPA), Brazil

  Funded the purchase of articles

Declarations of interest

Miguel Fuenteallba‐Tores declares no relationships or activities that could have influenced the submitted work.

Lucia A Lara declares no relationships or activities that could have influenced the submitted work.

Denisse Cartagena‐Ramos declares no relationships or activities that could have influenced the submitted work.

Jaqueline BP Figueiredo declares no relationships or activities that could have influenced the submitted work.

Ana Carolina JS Rosa‐e‐Silva declares no relationships or activities that could have influenced the submitted work.

Rui A Ferriani declares no relationships or activities that could have influenced the submitted work.

Wellington P Martins declares no relationships or activities that could have influenced the submitted work.

New search for studies and content updated (no change to conclusions)