The effect of hormone replacement therapy on cognition and mood

Abstract

Objectives

To summarise the available data regarding the effect of hormone replacement therapy (HRT) on cognition and mood in women.

Background

Complaints of impaired cognition and mood are common in the peri‐menopausal and menopausal period. There is debate as to whether HRT can ameliorate this phenomenon.

Design

A literature search of studies using electronic databases was conducted. Both randomised control trials and observational studies were included.

Patients

Perimenopausal and menopausal women.

Results

Due to the heterogenicity of results it is challenging to draw firm conclusions. The preparations used in many of the studies are older regimes no longer routinely used clinically. The notion of a ‘critical window’ for HRT is compelling, suggesting HRT has a positive impact on cognition when administered in the peri‐menopausal or early postmenopausal period but may have negative effects on cognition in the older, postmenopausal woman. The evidence would seem to suggest importance of hormonal replacement in woman undergoing a surgical menopause, especially when young. It remains unclear for how long they ought to continue HRT though until at least the natural age of the menopause seems reasonable. Evidence for a positive effect of HRT on mood is more convincing, though possibly more efficacious in the younger age group. The effect of HRT on anxiety is less clear.

Conclusions

Further study, particularly focusing on the more contemporaneous HRT preparations, is warranted before evidence‐based conclusions can be drawn.

1. INTRODUCTION

The menopause is a retrospective diagnosis following 12 months of amenorrhoea and represents a low oestrogen state. ¹ This is preceded by the peri‐menopause, a period of time of variable length, characterised by fluctuations in hormone levels secondary to a dwindling pool of ovarian follicles. Cognitive and mood changes during the menopausal and perimenopausal period are commonly reported in clinical practice. ¹ Patients report forgetfulness, brain fog and a lack of concentration. ¹ The 2011 SWAN study, involving >16,000 women aged 40–55‐year‐old, found that reported forgetfulness was higher in the postmenopausal compared to premenopausal groups; 41% and 31%, respectively. ² A 2013 meta‐analysis of four cross‐sectional studies found verbal recall to be the most significantly different parameter between pre‐ and postmenopausal women. ¹ Longitudinal studies following individuals through the menopausal transition have also found participant reported worsening of cognitive function following the menopause. ³ , ⁴ These are subjective reports yet importantly they are supported by studies, albeit small, assessing objective cognitive function. This has found premenopausal women perform better than postmenopausal women when formally assessed for cognition. ⁵ A 2014 Nuffield Health Survey of 3725 women aged 40–65‐year‐old reported that 47% of menopausal woman felt depressed and 37% felt anxious. ⁶ There are many factors that plausibly contribute to this phenomenon such as poor sleep related to vasomotor symptoms, a change in body shape, and the familial and social stressors that often occur at this stage of a woman's life for example bereavement, relationship breakdowns and children leaving home. ⁷ However, the argument for waning oestrogen levels being directly causal is compelling. Animal studies show oestrogen regulates metabolism within the brain, increases cerebral blood flow, counters oxidative stress, stimulates neurotransmitter synthesis and increases dendritic growth. ⁸ Given the biological plausibility of the deleterious effects of low oestrogen states on the brain it is postulated that hormone replacement therapy (HRT) could ameliorate this phenomenon. However, the evidence is conflicting and in many areas lacking. Analysis of available data is complicated by the heterogeneity of studies. The National Institute of Clinical Excellence (NICE) recommends offering HRT to address specific symptoms of the menopause, namely vasomotor symptoms, sexual dysfunction and urogenital atrophy. ⁹ Currently it recommends considering HRT to alleviate low mood and makes no specific recommendation regarding the use of HRT for cognitive function. ⁹ This review strives to summarise the current evidence.

2. COGNITIVE FUNCTION AND HRT

Many trials have investigated the effect of HRT on cognition, though often as a secondary outcome measure. The results are inconsistent with some reporting beneficial effects ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ and others showing no ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ or deleterious effects. ³⁰ , ³¹ , ³² , ³³ , ³⁴ It is important to note that women with a uterus who take oestrogen (E) for HRT need to also receive a progestogen (P) for endometrial protection. This to a degree limits studies as only participants who have undergone a hysterectomy can receive E‐only HRT. Thus, if treatment effects are seen it is difficult to establish whether they are driven by the oestrogenic or progestogenic component of the HRT. There is also great variety in the hormonal preparations used in the studies (Table 1), some of which are no longer routinely used in clinical practice.

Table 1.

Hormonal preparations used in hormone replacement therapy

	Name	Abbreviation
Oestrogen	17β‐Oestradiol	17β‐E2
	Oestradiol	E2
	Conjugated equine oestrogen	CEE
	Ethinyl estradiol	EE
	Estradiol valerate	EV
Progestogen	Medroxyprogesterone acetate	MPA
	Norethisterone	NET
	Dienogest	‐
	Dydrogesterone	DYD
	Nomegestrol acetate	NMG
	Micronized progesterone	Micro‐P
Androgen	Testosterone undecanoate	TU
	Methyltestosterone	‐

Note: Tibolone: Oestrogenic, progestogenic and androgenic action.

A 2005 double‐blind randomised control trial (RCT) by Viscoli et al. (n = 461) studied women receiving combined HRT (E + P) versus placebo over a 3‐year period. ¹⁴ The treatment group scored more highly than placebo in all cognitive domains with the exception of delayed recall. The oestrogen used in this trial was 17β‐E2 and the progestogen was MPA. However, this study was carried out in women with a past medical history of stroke. A double‐blind RCT by Tierney et al. ¹⁷ in 2009 (n = 142) compared combined HRT (E + P) versus placebo for a duration of 2 years. Verbal recall was found to be improved in the treatment group compared with placebo. The oestrogen used was E2 and the progestogen was NET. This study was limited by assessing only one aspect of cognition, namely verbal recall. The age range of patients studied was 61–87‐year‐old.

A large study by Shumaker et al. ³⁴ of postmenopausal women randomized to CEE + MPA or placebo for a period of 7 years found significant regression in scores in the treatment arm (n = 2808). This study exclusively recruited women >65 years without probable dementia. Ober et al. ³⁵ compared the use of E‐alone versus E + P versus placebo with regard to recall of lists and stories in postmenopausal women (n = 100). ³⁵ The participants were tested for both immediate recall and delayed recall. The E‐only group showed a significantly more rapid deterioration in immediate list recall, and to a lesser degree story recall than the E + P group. Delayed testing was unchanged between groups. The authors concluded that oestrogen hastens the natural age‐related waning of list recall but slows the natural age‐related decline of story recall relative to placebo. Furthermore, P reduced the effects of E, thereby accounting for the dissimilarity found between E‐only and E + P subgroups.

Pefanco et al. ²⁸ assessed memory, language, mood and executive function at baseline, 3 months and 3 years in women taking 17β‐E2 versus placebo (n = 57). Women with a uterus also concomitantly received oral micronized progesterone. This study exclusively recruited older postmenopausal women. It failed to observe a difference in the scores for women in the treatment arm versus placebo at 3 years. Polo Kantola et al. ¹⁸ studied a group of postmenopausal women randomised to either E or placebo for 3 months (n = 60). These women, aged between 47 and 65‐year‐old, had all previously had hysterectomies allowing the investigators to give E‐only HRT. E failed to show improvement in cognitive domains including speed and accuracy, attention and memory. Almeida et al. performed a double‐blind RCT comparing oral E2 versus placebo in women ≥70‐year‐old (n = 115). ²⁹ After 20 weeks no difference was found in cognitive outcomes between the groups. These findings were consistent with the findings of Henderson et al. ¹⁹ (n = 567), Kocoska‐Maras et al. ²⁰ (n = 200) and Yoon et al. ²¹ (n = 37).

The above studies serve to show the difficulties in reaching firm conclusions when analysing data generated from studies which utilised a variety of different cognitive function assessment tools and in which there is a range of HRT regimes, ages, medical comorbidities and menopausal status of participants. Moreover, the healthy user bias, whereby women of generally better health are more likely to opt to start HRT, confounds observational studies. RCTs seek to overcome the limiting factors such as healthy user bias posed by observational studies. As exemplified above the findings from these are conflicting.

2.1. Treatment regimes

Wroolie et al. ³⁶ compared the effect of different E formulations on verbal memory in healthy postmenopausal women aged 49–68‐year‐old (n = 68). ³⁶ Women were allocated to 17β‐E2 or CEE. The 17β‐E2 group showed considerably higher verbal memory performance in comparison to women receiving CEE. When comparing cognitive outcomes in women administered E2 or placebo (n = 14) Krug et al. ³⁷  found that while convergent thinking and word recall improved with E2, divergent thinking and motor error scores showed a declining pattern in comparison with placebo. Sherwin et al. ³⁸ investigated the differential effects of CEE + placebo versus CEE + MPA versus CEE + micro‐P on cognitive function (n = 24). Cognitive function testing of the CEE + placebo group remained unchanged following treatment. Interestingly whilst the CEE + MPA group was found to have a significantly greater improvement in working memory compared with the other two groups this was the only group where a significant decrease in delayed verbal memory scores was identified. Albertazzi et al. ³⁹ trialled the effect of tibolone versus EV + NET for 6 months and found both yielded improvement in memory (n = 22). Similarly, Kocoska et al. ⁴⁰ compared EV versus TU versus placebo over a 1 month period of women between 50 and 65 years of age (n = 203). Individuals on TU had improvement in verbal and spatial memory, and a decline in verbal fluency. A reduced spatial ability score was obtained in the women allocated to EV. It is worthy of comment that assessment of the effects of the older forms of HRT for example, CEE may be less relevant as they are scarcely used in clinical practice now. The use of synthetics progestins is declining in favour of the bio‐identical micronized progesterone due to its more favourable risk profile. However, current evidence is lacking to draw evidence‐based conclusions regarding its effect on cognition compared with synthetic progestins. While studies on testosterone are included above for completion it is important to note that the current stance of NICE is clear; RCTs have not to date demonstrated beneficial effects of testosterone on energy, mood, cognition or musckuloskeletal health. Until further data is available the only indication for testosterone is hypoactive sexual desire disorder in menopausal women. ⁹

2.2. Timing of treatment

Taking into account the age range of women within these studies the notion of a ‘critical window’ has arisen to potentially explain the conflicting data. This hypothesis suggests that HRT can have a positive impact in the premenopausal or perimenopausal period but a negative impact on cognition in the postmenopausal period. ⁸ , ⁴¹ This conclusion based on the balance of evidence can be seen schematically in Figure 1.

• A. Surgical menopause:

  Women who have been rendered menopausal by surgical removal of their ovaries (oophorectomy) experience a sudden decrease in hormone levels. This is in contrast to the more gradual decline seen during the natural menopause. Rocca et al. ⁴² found that the earlier the age of surgical menopause the greater the degree of cognitive impairment. The large Three Cities Study reported menopause before the age of 40‐year‐old was associated with a 40% increased risk of impaired visual memory and verbal fluency. ³ A recent meta‐analysis found a surgical menopause before the age of 45‐year‐old was associated with a 70% increased risk of dementia. ⁴³

  Recruiting only women who had undergone a surgical menopause, Sherwin compared E alone versus androgen (A) alone versus E + A versus placebo (n = 24). ³⁸ Short and long‐term memory as well as logical reasoning skills were assessed before and after starting treatment. Each treatment group had a statistically significant retention of cognitive function compared to the placebo group. Similarly, Philips et al. ¹¹ enroled women due for oophorectomy (n = 19) and assessed their cognition preoperatively. Following oophorectomy they were allocated to either E2 or placebo. Postoperative testing found that immediate recall of paragraphs was substantially better in the treated group compared with preoperatively; however, scores remained unchanged in the placebo group. HRT influence on the immediate or delayed recall of visual material, delayed recall of paragraphs, or digit span scores was insignificant compared to placebo. The sample size was small and the women were of a younger demographic (mean age of 48) though similar findings have been found in other studies in women of similar ages using CEE postsurgical menopause. ³⁶ In contrast, a study treating women with E2 implants for 10 years postsurgical menopause found that their verbal memory was poorer than women who were not taking HRT. ⁴⁴

  A Canadian study found that women indicated for oophorectomy had a higher Framlingham Risk Score (e.g., obesity, cigarette smoking) for cardiovascular disease and were on average less educated preoperatively than women who underwent a natural menopause. ⁴⁵ A large (n = 1315) British longitudinal study following women from an average age of 43 to an average age of 69 found that an older age of natural menopause was associated with better verbal memory. ⁴⁶

  This would suggest therefore that surgical menopause is associated with an increased risk of impairment to verbal recall, in a dose‐dependent fashion (i.e., the earlier the oophorectomy the greater the deficit). Whether there is an association with surgical menopause and dementia is unclear and potentially confounded by factors such as presurgery increased cardiovascular risk and educational parameters. Access to and acceptance of HRT postoperatively may also be affected by levels of education and other socioeconomic factors. The evidence is unclear but it would seem reasonable to conclude that HRT given up to the natural age of the menopause is beneficial but that there is a period of time/age after which the risk of cognitive impairment may be increased with HRT use.

• B. Older age:

Figure 1.

The effect of hormone replacement therapy on cognition. Graphical representation of the balance of evidence for benefits versus harm of HRT on cognition in perimenopausal or menopausal women in a mixed age group (1A), and women below (1B) and above (1C) 60‐year‐old. Small, medium and large font size of citations is used for studies recruiting ≤100 participants, 100–999, and ≥1000, respectively. [Color figure can be viewed at wileyonlinelibrary.com]

An extensive series of multicentre trials, conducted under the umbrella of the Women's Health Initiative (WHI) on the use of HRT in postmenopausal women, was published in 2002 and 2004. ³⁰ , ⁴⁷ One such trial was the WHI Study of Cognitive Aging which observed a decline in cognition in women receiving a combination HRT (E + P). ³² This decline was not observed in women on E only preparations (no effect). Also under the WHI umbrella, Shumaker et al. ³⁴ performed a long‐term double‐ blind RCT comparing CEE + MPA versus placebo in postmenopausal women ≥65‐year‐old (n = 4344). The placebo group was seen to perform significantly better on cognitive function testing than the treatment group at the end of the study period. These findings suggest that oestrogen has no protective effect on cognition in older postmenopausal women, and when used with a progestogen hastens cognitive decline.

However, Henderson et al. ¹⁹ investigated the effect of HRT on cognition by age of treatment onset failed to show a correlation (n = 567). Women were split into two groups by date of last menstrual period (LMP): LMP < 6 years ago or LMP > 10 years ago. They were assigned to placebo or to treatment which consisted of oral 17β‐E2 with the addition of micronized progesterone vaginally for those with a uterus. With a mean duration of 2.5–5 years of treatment there were no observed differences in the cognitive performance of women with respect to time since menopause before instigation of HRT.

2.3. Dementia

The risk of dementia increases with advancing age and rates are higher in women than men. ⁴⁸ Previously this has been accounted for by the longer life expectancy in women. However even once corrected for this longevity the greater incidence of dementia in women prevails. ² The age‐related increase in risk of dementia may account for this ‘critical window’, though the effect of HRT on dementia risk remains largely unclear.

In general, data from the WHI showed no significant change in cognitive function in postmenopausal women on HRT. However subgroup analysis revealed a statistically increased risk of dementia if E + P HRT was started at 65–75‐year‐old. ³¹ Of note, though the risk was also increased in women taking E alone HRT in this age group, this did not reach statistical significance.

Another subset of the WHI was the Women's Health Initiative Memory Study (WHIMS) (n = 4532). ⁴⁹ This study, assessing women who started HRT > 65‐year‐old, found in women taking CEE + MPA there was an increased rate of dementia. This result was not found in the group taking CEE alone. A large Finnish case–control study compared postmenopausal women with Alzheimer's disease (n = 84,739) with controls (n = 84739) between 1999 and 2013. ⁵⁰ It found that both E and E + P HRT regimes was associated with an increased risk of Alzheimer's disease. If the HRT was initiated before the age of 60 this increased risk was recognised after 10 years of use. However if HRT was initiated after the age of 60 the risk of Alzheimer's disease was increased irrespective of duration of use. In contrast, The KEEPS Cognitive and Affective Study (n = 693) which randomised women to CEE + micronized progesterone versus transdermal E2 + micronized progesterone versus placebo found no positive or negative effect on cognition over the 4‐year study period. ⁵¹

2.4. Cerebrovascular disease

Studies have examined the effect of HRT on cardiovascular risk. Follow‐up data from the WHI did not find HRT to protect against cardiovascular disease. ³⁰ , ⁴⁷ Due to shared risk factors it has been postulated that cerebrovascular disease may follow a similar trend to that of cardiovascular disease. The same WHI follow‐up data found the CEE‐only and the CEE + MPA groups had an increased risk of venous thromboembolism and stroke. ³⁰ , ⁴⁷ The Framlingham Heart Study reported on morbidity from cardiovascular disease in postmenopausal women aged 50–83‐year‐old (n = 1234). ⁵² It found a significantly increased risk of stroke amongst users of oestrogens. Older cohort studies found a 20%–50% decreased risk of stroke with HRT use, though these studies were uncontrolled. ⁵³ , ⁵⁴ The Nurse's Health Study found a 45% higher stroke risk in women using combined E + P HRT compared with women who had never used HRT. ⁵⁵

2.5. Summary of cognition and HRT

Animal models suggest oestrogen is neuroprotective to the aging female brain. However, paradoxically emerging evidence suggests older women, at increased risk of dementia or cardiovascular disease by virtue of age, may experience decline in cognitive function with HRT use. This risk may to a degree depend upon whether a progestogen is used alongside the oestrogen, or indeed which progestogen is used. Further work is needed to clarify this. Women who have undergone a surgical menopause, possibly more representative of animal models of induced menopause, may benefit cognitively from HRT. Likely treatment until the natural age of the menopause is advisable in the context of premature menopause. Limited evidence exists regarding the optimal duration of HRT use before risks may dominate potential benefits on cognition function. ⁵⁶

3. MOOD AND HRT

The evidence for the use of HRT to ameliorate mood symptoms associated with the menopause is conflicting with studies suggesting a positive effect, ¹² , ²⁷ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ no effect ⁷⁴ , ⁷⁵ , ⁷⁶ or negative effects. ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ , ⁸⁷ , ⁸⁸ , ⁸⁹ , ⁹⁰ , ⁹¹ , ⁹² , ⁹³ , ⁹⁴ , ⁹⁵

3.1. Depression

The 2015 NICE guidance to consider HRT for the treatment of low mood in association with the menopause was derived from several RCTs carried out between 1977 and 2014. ⁹ These studies utilised a variety of mood assessment tools (Table 2). In addition to these RCTs there are further contemporaneous studies that demonstrate a positive correlation between improvement in mood and use of HRT. Two such studies, included in the revised NICE guidance are those conducted by Gleason et al. ²⁷  and Gordon et al. ⁷¹  Gleason at al. ²⁷ randomly allocated women aged between 40 and 60 to three groups: oral CEE only, transdermal 17β‐E2 + micro‐P group or placebo (n = 693, follow up over 4 years). They found that the active treatment groups displayed significantly higher mood scores in comparison with placebo. Gordon et al. ⁷¹ compared the effect of transdermal E2 + cyclical micro‐P versus placebo over a 12‐month period in euthymic patients aged 45–60 (n = 172). There was a significantly lower incidence of developing depression in the treatment arm compared with placebo. Furthermore, in 2016, Raz et al. ⁷² observed a possible association between oestrogen and progesterone levels and platelet content of serotonin (5‐hydroxytryptamine, 5‐HT) which in turn is a key determinant of the psychological state of mind (n = 79). The mean platelet content of 5‐HT was used as a surrogate marker for the synthesis and nervous system procurement of 5‐HT. This was measured via enzyme‐linked immunoassay (ELISA). Women were randomised to either oral CEE, transdermal E2 or placebo for a period of 48 months (n = 79). Mood was assessed using the POMS questionnaire. Mean platelet content of 5‐HT showed an enhancement by 84.5%, and 107% in the CEE group and E2 group, respectively. This compared to only an 8.8% rise in the placebo group. An improvement in mood was observed in both the CEE and transdermal E2 groups compared to placebo. Another such study was carried out by Rudolph et al. ⁵⁷ which randomly allocated women aged 48–65‐year‐old to E2 + dienogest versus placebo (n = 129). The E2 + Dienogest group showed a significant improvement in mood in comparison with the placebo group.

Table 2.

Mood assessment tools used in perimenopausal and menopausal women

Assessment tool	Abbreviation
Montgomery‐Asberg Depression Rating Scale	MADRS
Beck Depression Inventory	BDI
Clinical Global Impression	CGI
Clinical Global Impression	CGI
Profile of Mood States	POMS
Geriatric depression score	GDS

However, these results and thus the basis of the recommendation by NICE is not universally supported by other studies. Many older studies found no significant difference in mood symptoms after treatment. Demetrio et al. ⁷⁴ assessed the effect of CEE versus placebo for 6 months on mood and found no difference (n = 50). Yalamanchili et al. ⁷⁵  studied older postmenopausal women (65–77‐year‐old) and found no effect of HRT compared to placebo on GDS (n = 489). A more contemporaneous but small study (n = 66) by Schimdt et al. ⁷⁶ in 2021 replicated these findings, finding no beneficial effect in any treatment arm when assessing transdermal E2, raloxifene (a selective oestrogen receptor modulator), phytoestrogen or isoflavone. Cohen et al. ⁷³ carried out a study on peri‐ or postmenopausal women with a known diagnosis of depression (n = 22). Enroled participants were administered transdermal 17β‐E2 and assessed via a range of mood scores (MADRS, BDI and CGI) which were then repeated following an interval of 1 month. Results suggested a beneficial role of HRT in clinically depressed peri‐menopausal women in comparison with postmenopausal women. It is plausible that HRT is more effective in peri‐menopausal women, or that their effects are more pronounced in this group and therapeutic benefit of HRT depreciates as postmenopausal women age.

Physiologically progesterone is high in the luteal phase of the menstrual cycle, coinciding with premenstrual symptoms and as such it is commonly assumed that progesterone negatively affects mood. Furthermore, oestrogen is known to causes an increase in release of serotonin quantified by urinary 5‐hydroxyindoleacetic level but this is reduced if a progestogen is co‐administered. In support of this concept, Björn et al. ⁹⁶ administered E2 + MPA to women aged 39–55 years for a period of 5 months and found an increase in negative mood states of tension, irritability and depressed mood, and a decrease in positive mood parameters of friendliness (n = 28). Various randomised controlled trials confirm these findings. ⁹⁷ , ⁹⁸ , ⁹⁹ , ¹⁰⁰ In contrast, a study conducted by Cagnacci et al. ⁵⁸ found that specific progestogen preparations may enhance mood (n = 120). Postmenopausal participants were given transdermal E2 daily with the addition of cyclical progesterone. They were allocated to take, as the progestogenic component, either DYD, NMG, MPA or NET. The E2 + DYD group had a significant decrease in anxiety, whilst E2 + MPA groups demonstrated a significant decrease in depressed mood. There was no deterioration in mood symptoms demonstrated with any progestogen preparation. Importantly, a large study conducted by Lee et al. ¹⁰¹ noted an increase in suicidal ideation in postmenopausal women using HRT (n = 2286). It was found that the duration of HRT was directly proportional to the risk of suicidal ideation. This was however an observational study only.

3.2. Anxiety

Baksu et al. ⁷⁰ randomised women following a surgical menopause to receive oral tibolone, transdermal E2 or oral placebo with anxiety scores measured at baseline and after 6 months of treatment (n = 75). A significantly positive response was observed in both treatment arms compared to placebo. These findings have been confirmed by other authors. ¹² , ⁵⁸ However, Bukulmez et al. ⁷⁷  failed to observe a significant effect of any HRT preparation on anxiety when assessing CEE + MPA, CEE + MPA, tibolone or alendronate for a 3‐month duration (n = 60). A randomized controlled trial conducted by Girdler and colleagues comprising of menopausal women aged 48–69 yielded similar results. Participants received either CEE + MPA or placebo. ⁷⁸ At 6 months no significant difference was noted between the two groups. Similarly, Nielsen studied the effect of intranasal HRT preparations on women aged 40–65 years and found no improvement in mood or anxiety after treatment (n = 335). ⁶² It is possible that women with a surgically induced menopause are more likely to experience benefits of HRT with regard to anxiety compared with women who have attained menopause naturally. ⁶²

3.3. The use of antidepressants

The current NICE guidelines specifically state that antidepressants should not be used for menopausal low mood without clinical depression. ⁹ The largest trial on the use of antidepressants with HRT was conducted by Dias et al. ¹⁰² investigating HRT's role as a potential enhancer of Venlaflaxine's antidepressant effect on postmenopausal women. The study randomly allocated postmenopausal women (mean age 53.6) with known depression on Venlaflaxine into 4 study arms lasting 24 weeks: CEE + MPA + methyltestosterone, CEE + MPA, methyltestosterone‐only, and placebo (n = 72). Each group displayed a positive response in comparison with placebo, with the methyltestosterone‐alone group having superior therapeutic efficacy. However the dropout rates were highest in the methyltestosterone group so this should be interpreted with caution. Morgan et al. ¹⁰³ investigated the efficaciousness of oestrogen‐only HRT (CEE) to enhance the effect of antidepressants on mood in peri‐menopausal women. Women taking antidepressant treatment for a major depressive disorder but with only a partial response were randomized to receive CEE or placebo alongside their usual antidepressant (n = 17). Depression scores were significantly reduced in the group allocated to CEE compared with placebo after 6 weeks. The study concluded oestrogen augmented the effect of antidepressants for the treatment of depression in peri‐menopausal and postmenopausal women. ¹⁰⁴ Nagata et al. ¹⁰⁵ randomly allocated surgically menopaused women with hot flushes and depressive symptoms to one of two treatment arms for 8 weeks: CEE‐alone versus CEE + Fluvoxamine (a selective serotonin reuptake inhibitor [SSRI]) (n = 42). The results suggested a superior role of CEE + fluvoxamine compared with CEE‐alone in the improvement of depression scales, although scores improved in both groups compared to baseline. However, no effect was observed in either of the treatment arms with regard to anxiety. The frequency of hot flushes reduced in both groups but most in the CEE + SSRI group. Soares et al. ¹⁰⁶ conducted an open‐labelled RCT among women aged 40–60 years comparing SSRI versus HRT (Escitalopram vs. EE + NET) (n = 40). Depression scores and menopausal symptoms improved in both arms but to a greater degree in the SSRI group. These results are promising however, the trials were all small and significantly they studied the older and now less clinically used HRT preparations. Further studies including oestradiol would be valuable.

3.3.1. Summary of mood and HRT

On balance HRT seems to improve low mood associated with the menopause (Figure 2). This effect may be greater in the younger age group. There is conflicting data regarding the notion that oestrogen positively affects mood and progesterone negatively affects it. The type of progestogen used clinically may be relevant. The effect of HRT on anxiety is less clear though it potentially has a more beneficial role in women who have undergone a surgical menopause than those undergoing a natural menopause at a relatively older age. Evidence to suggest a cumulative effect treatment effect for depression when using antidepressants alongside HRT is promising but further exploration is needed. Importantly further data is needed on the more contemporaneous HRT preparations.

Figure 2.

The effect of hormone replacement therapy on mood. Graphical representation of the balance of evidence for benefits versus harm of hormone replacement therapy in perimenopausal or menopausal women for depression (2A) and anxiety (2B). Small, medium and large font size of citations is used for studies recruiting ≤100 participants, 100–999, and ≥1000, respectively. [Color figure can be viewed at wileyonlinelibrary.com]

4. CONCLUSION

Impaired cognition and mood disturbance is a common complaint amongst perimenopausal and menopausal women. There are both psycho‐social and biological explanations for this phenomenon. Analysis of the evidence regarding the impact of HRT on cognition and mood is challenging due to the heterogenicity of studies, the use of near historical HRT regimes in study protocols, and the age range and variety of comorbidities of women enroled in the trials. Current data is insufficient to draw conclusions of impact between the types of HRT for example, the older conjugated equine oestrogens and oestradiol. Some studies find improvements in cognition with HRT whilst other studies find no effect or even deleterious effects. A notion of a ‘critical window’ of HRT has arisen whereby HRT may improve cognition in the younger age group but be deleterious in older, postmenopausal women. Women who have undergone a surgical menopause are more likely to benefit cognitively from HRT. It remains unclear for how long they ought to continue HRT though until at least the natural age of the menopause seems reasonable. Evidence for a positive effect of HRT on mood is also conflicting but more convincing. It is likely more efficacious in the younger age group. Furthermore, HRT and antidepressants seem to have a positive cumulative effect in clinical depression. The data regarding the effect of HRT on anxiety states is lacking. In summary further study is warranted before evidence‐based conclusions can be drawn.

CONFLICT OF INTEREST

The views expressed are those of the authors and not necessarily those of the above‐mentioned funders, the NHS, the NIHR, or the Department of Health. C. N. J. received a speaker fee for participating in a debate at BES 2016 meeting sponsored by Besins Healthcare.

Sharma A, Davies R, Kapoor A, Islam H, Webber L, Jayasena CN. The effect of hormone replacement therapy on cognition and mood. Clin Endocrinol (Oxf). 2023;98:285‐295. 10.1111/cen.14856

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.