Menopause types

There are many types of menopause with distinct hormonal profiles and distinct influences on the brain and cognition.

In humans, the ovarian cycle typically ranges from 21 to 35 days, but can vary in length or be absent owing to pregnancy, use of hormonal contraception, stress, exercise, medication or medical conditions that affect ovarian function (e.g. polycystic ovary syndrome (PCOS) ^{Reference Mishra, Vijay and Tiwari16} ). Eventually, ovarian function declines because of the natural depletion of primordial ovarian follicles, resulting in changes in cycle length, and marking the beginning of perimenopause. In some individuals, cycles may initially become more regular before they become more irregular later on in the transition; ^{Reference Jain and Santoro17} this heterogeneity is important to consider when pinpointing the onset of perimenopause.

The perimenopausal stage is characterised by the initiation of menopausal symptoms (further described in section ‘The many menopausal symptoms’), and substantial change in ovarian hormones and gonadotropins. During perimenopause, there is heightened FSH release from the pituitary gland, low levels of anti-Mullerian hormone (a marker of ovarian reserve) and highly variable/fluctuating 17β-oestradiol (E2) release from the ovaries. ^{Reference Davis, Lambrinoudaki, Lumsden, Mishra, Pal and Rees7,Reference Harlow, Gass, Hall, Lobo, Maki and Rebar18,Reference Bulun19} Over time, the ovaries become less responsive to FSH. This results in reduced E2 and luteinising hormone production, more frequent anovulation (i.e. less frequent menses) and, in turn, reduced P4 production until the final menstrual period (i.e. 12 months without menstruation). ^{Reference Davis, Lambrinoudaki, Lumsden, Mishra, Pal and Rees7,Reference Harlow, Gass, Hall, Lobo, Maki and Rebar18,Reference Bulun19} However, as noted below, there is substantial variation in the levels of these hormones, making it challenging to determine thresholds. ^{Reference Randolph, Zheng, Sowers, Crandall, Crawford and Gold12,Reference Stellato, Crawford, McKinlay and Longcope20}

Although menopause is often referred to as the cessation of ovarian hormones, the postmenopausal ovaries continue to produce testosterone up to 10 years after the final menstrual period. ^{Reference Fogle, Stanczyk, Zhang and Paulson21} In addition, there are other sources of oestrogens aside from the ovaries. Adipose tissue, which can increase in postmenopause, can produce another of the oestrogens, oestrone (E1). ^{Reference Bulun19} Furthermore, the adrenal glands continue to produce androstenedione, a precursor of oestrogens, into older age, but to a much lesser extent than the ovaries. ^{Reference Bulun19} There is substantial interindividual variability in the concentration of FSH and luteinising hormone as well as in the length of the perimenopausal period, making it challenging to model. On average, perimenopause lasts approximately 4 years, but in some individuals it can be as short as 2 years or upward of 10 years, ^{Reference Li, Lanuza, Gulanick, Penckofer and Holm14} with ethnic/geographic disparities ²² and differences in racial disparities: Black and Hispanic females have been found to experience menopause earlier, on average, than White females. ^{Reference Appiah, Nwabuo, Ebong, Wellons and Winters23} These intersectional factors are important to consider for understanding the effects of menopause on brain and cognitive outcomes.

Menopause can vary in the age at which it occurs and in the abruptness of its transition, and based on these parameters, can be categorised into spontaneous, early, premature or induced (chemically or surgically). However, the type of menopause is not always accounted for in studies, but can dramatically affect brain health outcomes. ^{Reference Edwards, Duchesne, Au and Einstein24} Early and premature menopause follow the same transition and progression of hormonal changes as spontaneous menopause, but occur earlier in time. The definition of early menopause varies, but it is generally defined as menopause before age 45 years and after age 40 years, and premature menopause is defined as menopause before age 40 years. Primary ovarian insufficiency ((POI), also referred to as premature ovarian failure) is distinct from premature menopause. In POI, the ovaries stop functioning optimally before age 40, but individuals with POI can still become pregnant and may experience occasional menstrual periods. ^{Reference Stuenkel and Gompel25} Thus, POI should not be confused with premature menopause. ²⁶ Both early and premature menopause are experienced by one in ten females. ^{Reference Rocca, Gazzuola Rocca, Smith, Kapoor, Faubion and Stewart27} An earlier age of menopause is linked with poorer cognitive and brain outcomes, as well as increased risk for dementias. ^{Reference Gong, Harris, Peters and Woodward28,Reference Wood Alexander, Wu, Coughlan, Puri, Buckley and Palta29} Early menopause has also been associated with increased neuropathology, reduced cognition and increases in biological ageing markers. There are region-specific increases in levels of phosphorylated tau, a neuropathological feature of Alzheimer’s disease, in those with high amyloid-β load in the entorhinal cortex in people that have experienced an earlier age of menopause compared with older ages of menopause. ^{Reference Coughlan, Betthauser, Boyle, Koscik, Klinger and Chibnik30} In addition, early age of menopause increases epigenetic ageing, a marker of biological ageing. ^{Reference Levine, Lu, Chen, Hernandez, Singleton and Ferrucci4} Earlier age of menopause is associated with lower verbal memory scores ^{Reference Kuh, Cooper, Moore, Richards and Hardy31} and lower global cognitive scores in older age, particularly in those with heightened vascular risk factors, ^{Reference Wood Alexander, Wu, Coughlan, Puri, Buckley and Palta29} compared with those with later age of menopause.

Menopause can also vary in the abruptness of its transition. Unlike the spontaneous menopausal transition, which typically lasts several years, the menopause transition is abrupt when induced. Menopause can be temporarily induced chemically with gonadotropin-releasing hormone analogues (e.g. leuprolide for endometriosis), but is reversible with the cessation of treatment. Within 2–4 weeks of leuprolide acetate treatment, ovarian hormones are within the postmenopausal range, ^{Reference Grigorova, Sherwin and Tulandi32} whereas gonadotropins significantly decrease. Leuprolide acetate has been associated with worse verbal memory, executive functions and working memory in younger, premenopausal females. ^{Reference Grigorova, Sherwin and Tulandi32,Reference Sherwin and Tulandi33} In response to chemotherapy and/or radiotherapy, most females will experience a loss of menstrual periods owing to reduced ovarian functioning, and additionally may experience treatment-induced menopause. ^{Reference Mauri, Gazouli, Zarkavelis, Papadaki, Mavroeidis and Gkoura34} The hormonal profile of those with treatment-induced menopause is not well documented and is highly variable depending on the age of the individual, as well as the dose, frequency, duration and type of chemotherapy and/or radiotherapy treatment. ^{Reference Mauri, Gazouli, Zarkavelis, Papadaki, Mavroeidis and Gkoura34} The brain changes and cognitive deficits related to treatment-induced menopause are challenging to disentangle from the neurotoxic effects of the chemotherapy or radiotherapy itself. ^{Reference Li and Caeyenberghs35} Aromatase inhibitors (e.g. letrozole, anastrozole, exemestane) are often administered as adjuvant therapy to those with hormone receptor-positive breast cancers, as they block the conversion of androgens to oestrogens. ^{Reference Fardell, Walker, Chan and Vardy36} Aromatase inhibitors significantly decrease E2 levels, and significantly increase FSH and luteinising hormone levels. Both the human and non-human literature has suggested that aromatase inhibitors may negatively affect brain health and cognition, but the evidence is mixed. ^{Reference Fardell, Walker, Chan and Vardy36–Reference Gervais, Remage-Healey, Starrett, Pollak, Mong and Lacreuse39} Previous tamoxifen use may modulate the influence of aromatase inhibitors on cognition, but this is often overlooked in the human literature. ^{Reference Fardell, Walker, Chan and Vardy36,Reference Lee Meeuw Kjoe, Kieffer, Small, Boogerd, Schilder and Van Der Wall37}

Finally, menopause can be induced surgically by bilateral oophorectomy (i.e. the surgical removal of both ovaries), which results in an immediate and steep decline in E2, P4 and testosterone, and a sharp increase in luteinising hormone and FSH at the time of surgery. Compared with ovarian conservation, individuals with premenopausal bilateral oophorectomy have increased symptoms of anxiety and depression, ^{Reference Rocca, Grossardt, Geda, Gostout, Bower and Maraganore40} greater risk for dementias and accelerated cognitive decline. ^{Reference Crawford41} Bilateral oophorectomy is a risk-reducing procedure for women with the breast cancer gene (BRCA) mutation or other genetic risk factors for ovarian cancer (e.g. BRIP1, RAD51C/D, PALB2 and ATM). ^{Reference Edwards, Duchesne, Au and Einstein24} Approximately 300 000 people in the USA experience surgically induced menopause caused by bilateral oophorectomy each year. ^{Reference Parker, Jacoby, Shoupe and Rocca42} Furthermore, bilateral oophorectomy can be accompanied by removal of the uterus (hysterectomy) and/or fallopian tubes (salpingectomy), which can have differing effects on brain health outcomes. ^{Reference Koebele, Palmer, Hadder, Melikian, Fox and Strouse43} These differing causes of menopause are worth investigating because transition differences and age differences may matter in terms of outcomes on brain health, and yet are not always accounted for in the literature.

The many menopausal symptoms

All types of menopause are often accompanied by the presence of menopausal symptoms. The Menopause Rating Scale-II (MRS-II) is a validated questionnaire for assessing the impact of menopausal symptoms on health-related quality of life. ^{Reference Potthoff, Heinemann, Schneider, Rosemeier and Hauser49} The MRS-II groups 11 menopausal symptoms into three main categories: somatic-vegetative, psychological and urogenital. However, there are many other symptoms that people can experience during and after the menopausal transition than the 11 categorised. ^{Reference Honermann, Knabben, Weidlinger, Bitterlich and Stute50} A report from the Menopause Foundation of Canada ⁵¹ lists up to 30 symptoms, with most people reporting seven symptoms at a given time. Vasomotor symptoms ((VMS) hot flashes and night sweats), which can both be present or solely present, are reported with high frequency (around 80%) during the menopausal transition. ^{Reference Gold, Colvin, Avis, Bromberger, Greendale and Powell52} Less commonly discussed symptoms of menopause include tinnitus, ^{Reference Lai, Liu and Liu53} burning mouth syndrome (among several other oral health symptoms), ^{Reference Suri and Suri54} dry eyes, musculoskeletal symptoms (arthritis) and itchiness of the skin, all of which may or may not co-occur with VMS. The sheer variety of symptoms also makes modelling menopause challenging, with many combinations of symptoms possible. Although the variety of symptoms are daunting to study, it should not be surprising considering the widespread abundance of oestrogen and P4 receptors throughout all organs, glands and systems, including the brain. Importantly, these less common symptoms can influence quality of life, but are not captured in questionnaires like the MRS-II.

Even among the more common menopausal symptoms, there is a large variability in their reported prevalence. For example, the prevalence of mood symptoms ranges from 15 to 78% and the prevalence of hot flashes and night sweats ranges from 36 to 87% depending on the study. ^{Reference Baker, Lampio, Saaresranta and Polo-Kantola55} This substantial variability is likely attributable to a combination of biological and sociocultural factors ^{Reference Sievert56} and the severity and type of menopausal symptoms depends on type of menopause. Females with surgically induced menopause experience more severe menopausal symptoms relative to women with spontaneous menopause, and score significantly higher on the MRS-II. ^{Reference Bhattacharya and Jha57} Moreover, severity and type of menopausal symptoms also vary by geographic region. For example, East Asian countries experience fewer and less severe hot flashes compared with European countries. ^{Reference Freeman and Sherif58} Another study of Omani females found that they are more likely to self-report physical symptoms compared with cognitive symptoms. ^{Reference El Shafie, Al Farsi, Al Zadjali, Al Adawi, Al Busaidi and Al Shafaee59} Collectively, these findings suggest that culture may influence openness to endorse/self-report menopausal symptoms and underlying genetic differences may influence the experience of menopausal symptoms. Other studies have also reported that education level, socioeconomic status and working status can influence the severity and type of menopausal symptoms. ^{Reference Kakkar, Kaur, Chopra, Kaur and Kaur60} Questionnaires like the MRS-II are useful to evaluate the influence of menopausal symptoms on quality of life, but may not accurately capture the menopausal experience of all females, nor, as discussed above, does it capture all symptoms of menopause. This symptom variability makes menopause difficult to model clinically and preclinically, especially given the lack of understanding of how these hormones can influence health in females.

The Stages for Reproductive Aging Workshop (STRAW) criteria – the gold standard for determining menopause – states that using menopausal symptoms to determine menopausal status is not reliable, given their high variability in presentation. ^{Reference Harlow, Gass, Hall, Lobo, Maki and Rebar18} This is an interesting conclusion, given that the STRAW criteria also recommend that FSH levels, in conjunction with frequency of menstrual periods, can be used to determine menopausal status. Although a cut-off of 40 IU/L FSH may be useful for determining menopausal status, repeated measurements of FSH are needed because there is substantial interindividual variability in FSH levels. ^{Reference Stellato, Crawford, McKinlay and Longcope20} Whether a single FSH cut-off point is useful is debatable, given literature showing FSH levels can remain below this cut off point even well into the menopausal transition. ^{Reference Stellato, Crawford, McKinlay and Longcope20,Reference Maki and Weber45} Additionally, FSH measurements are not routinely offered in clinical practice for determining menopausal status unless ruling out PCOS, premature menopause, treatment-induced menopause or assessing fertility, ^{Reference De Vos, van Laarhoven, Laven, Themmen, Beex and Sweep61} further decreasing their utility. In addition, not including symptoms is limiting menopausal research as phenotypic variability likely has differential needs in terms of treatment and possibly repercussions for brain health, given that VMS are more likely to be tied to white matter hyperintensities. ^{Reference Thurston, Wu, Chang, Aizenstein, Derby and Barinas-Mitchell62}

Another challenge for pinpointing the menopausal transition is that the frequency of menstrual periods is highly variable from person to person. For example, individuals with hysterectomy (i.e. removal of the uterus) have a complete cessation of menstrual periods following surgery, as there is no endometrial lining to be shed each month. Although menopause is known to occur on average 2–3 years earlier in those with hysterectomy compared with those without it, ^{Reference Farquhar, Sadler, Harvey and Stewart63} ovarian function remains normal and cycling occurs, but without menstrual bleeding. Additionally, with hormonal contraception being taken by approximately 400 million individuals worldwide, ⁶⁴ it can be increasingly challenging to determine the onset of perimenopause. Depending on the type of hormonal contraception (e.g. hormonal intrauterine device or continuous use of oral contraception without placebo/inactive pills), complete cessation of menstrual periods may be experienced before menopause. Worldwide, 182 million (19%) women aged 15–45 years use long-acting, reversible contraception (2% implant and 17% intrauterine device). ⁶⁴ In the USA, use of an intrauterine device increased from 7.1% in 2006–2010 to 21.4% in 2015–2019, ^{Reference Daniels and Abma65} suggesting that it is becoming a more popular choice of contraception. It is also challenging to determine perimenopause onset in women with POI and PCOS, where frequency of menstrual periods is diminished, but not necessarily reflective of ovarian function.

Another challenge is that menopause remains a taboo subject. A recent 2023 report published by the Menopause Foundation of Canada ⁵¹ suggests that menopause is viewed in a negative light, and initiating conversations related to menopause is a challenge; 50% of females worry that menopausal symptoms will affect how they are perceived at work, and 30% believe that if they disclosed their menopausal symptoms then others will perceive them as weak, old or ‘past their prime’. In addition, people report dissatisfaction in their menopausal care. Informed menopause care from health care providers is often challenging, given the lack of knowledge and awareness of the variety of menopausal symptoms (see section ‘Recommendations and future directions’).

What is MHT?

MHT is used to supplement the decline in ovarian hormones during the menopausal transition, to offset symptoms. For females with an intact uterus, concurrent progestin/P4 is needed to prevent endometrial hyperplasia and cancer. ^{Reference Furness, Roberts, Marjoribanks and Lethaby66} Although serum testosterone levels remain relatively stable throughout perimenopause and several years postmenopause, concurrent testosterone may be recommended as part of the MHT regimen, especially when lower sexual drive, without an underlying cause, is observed. ^{Reference Marko and Simon67} Testosterone may also be offered as part of the MHT regimen to females with POI or ovarian removal, who may have reduced or no testosterone output from the ovaries. Females with premenopausal oophorectomy have 25% lower circulating testosterone compared with spontaneously menopausal females with intact ovaries. ^{Reference Kotsopoulos, Shafrir, Rice, Hankinson, Eliassen and Tworoger68} It is important to note that many types of MHT are available, with varying doses and combinations of oestrogens, progestins and possibly androgens.

MHT can vary in several dimensions that affects its benefits – not only for alleviating menopausal symptoms, but also for ensuring healthy brain ageing into later life. These dimensions include its formulation, dose, route of administration (transdermal, oral, vaginal, intramuscular), cyclicity (continuous versus cyclic), time of initiation and duration. Additionally, individual health histories and biologies need to be considered in assessing which type of MHT may be the best, including the health status of the person receiving MHT, pregnancy history, genetic predispositions for dementia, as well as their age and type of menopause (healthy cell bias hypothesis, timing hypothesis). In the following sections, we will review the evidence pertaining to these factors, highlighting gaps in our knowledge and important research areas to consider moving forward.

To date, unfortunately, one of the most influential MHT studies is the Women’s Health Initiative Study (WHIS), a large randomised clinical trial that administered continuous oral conjugated equine oestrogen (CEE) and medroxyprogesterone acetate (MPA), or continuous oral CEE-alone therapy (to those with a hysterectomy), or placebo to females aged 50–79 years. ⁶⁹ Initially, they reported increased risk for breast cancer, dementia, stroke and pulmonary embolisms after combined treatment. ^{Reference Rossouw, Anderson, Prentice, LaCroix, Kooperberg and Stefanick70,Reference Shumaker, Legault, Rapp, Thal, Wallace and Ockene71} These findings resulted in the premature discontinuation of the WHIS and massive changes in MHT prescribing practices. Prescriptions for combined CEE and MPA decreased by 66% in the USA from January to June 2003 relative to January to June 2002. ^{Reference Hersh, Stefanick and Stafford72} The estimated number of MHT users was approximately 35 million before the WHIS, halved in the early 2000s and stabilised to 12 million MHT users in the 2010s. ⁷³ However, the study design has been widely criticised on a number of factors.

The WHIS results pertain to one type, dose, dosing schedule and route of administration of MHT, and one type of menopause. In addition, the WHIS had few exclusions for participants, as it included individuals at high risk for cardiovascular disease and approximately a third of participants were obese or smoked. Also, people with more severe VMS, who may be most likely to benefit from MHT (see section ‘VMS and brain health’), were discouraged from participating in the trial. ⁶⁹ Follow-up studies stratifying by age and time since menopause observed that increased risk of heart disease was only observed in those ages 60 years and beyond 10 years since menopause. ^{Reference Hsia, Langer, Manson, Kuller, Johnson and Hendrix74} Additionally, an evaluation of younger women who were aged 50–55 years at the start of the trial showed that taking CEE was not linked to adverse cognitive outcomes. ^{Reference Espeland, Shumaker, Leng, Manson, Brown and LeBlanc75} Moreover, brain atrophy and dementia risk was greatest in those with worse baseline cognitive scores, as measured by the Mini-Mental State Examination, ^{Reference Resnick, Espeland, Jaramillo, Hirsch, Stefanick and Murray76,Reference Espeland, Rapp, Shumaker, Brunner, Manson and Sherwin77} suggesting that the benefits from MHT may be limited to prevention of cognitive decline rather than the treatment of cognitive decline.

These findings are in line with the ‘window of opportunity hypothesis’ or ‘timing hypothesis’, which posits that initiating hormone therapy close to menopause onset has cognitive benefits, whereas initiating it further from menopause onset may have no effects or produce cognitive decrements. Meta-analyses support the initiation of hormone therapy during the early menopausal period. ^{Reference Ryan, Scali, Carriere, Ritchie and Ancelin78,Reference Hogervorst, Williams, Budge, Riedel and Jolles79} Indeed, studies show that the brain’s responsiveness to exogenous hormones declines after an extensive period of ovarian hormone deprivation. Consistent with this, in rodents a greater time since ovariectomy was associated with reduced neural stem cells within the dentate gyrus. ^{Reference Yagi, Mohammad, Wen, Batallán Burrowes, Blankers and Galea80} Moreover, in humans, oestrogen receptor expression in the brain may change over the course of reproductive senescence, as postmenopausal females have greater oestrogen receptor density in certain regions compared with premenopausal females; ^{Reference Mosconi, Nerattini, Matthews, Jett, Andy and Williams81} however, using ¹⁸F-fluoroestradiol positron emission tomography imaging for in vivo detection of brain oestrogen receptors has been challenged. ^{Reference Biegon, Jagust and Mankoff82} The potential upregulation of brain oestrogen receptors in postmenopause may reflect a compensation for reduced circulating levels of oestrogens. However, the Kronos Early Estrogen Prevention Study (KEEPS) did not find that early MHT initiation positively influenced cognition. ^{Reference Gleason, Dowling, Kara, James, Salazar and Ferrer Simo83} Importantly, the transdermal E2-treated group had more baseline white matter hyperintensities and a larger proportion of APOE4 carriers than the placebo group in the KEEPS trial, ^{Reference Kantarci, Tosakulwong, Lesnick, Zuk, Lowe and Fields84} which may have masked the cognitive benefits of E2. Additionally, MHT may have cognitive benefits in older age, depending on the type of hormones, route of administration and cyclicity/dosing schedule. ^{Reference Duka, Tasker and McGowan85,Reference Puri, Gravelsins, Alexander, McGovern, Duarte-Guterman and Rabin86} MHT has been postulated to be preventative not curative, such that MHT may be beneficial only in healthy individuals, and more likely be detrimental in unhealthy individuals. According to the ‘healthy cell bias of oestrogen action’, uncompromised neurons respond positively to exogenous hormones. ^{Reference Brinton87} This may explain why MHT has been associated with brain and cognitive benefits in older individuals, but is more likely to be beneficial if initiated closer to the onset of menopause.

MHT formulations containing E2 are more likely associated with beneficial effects compared with formulations containing E1, perhaps because of the more potent effects of E2 on ERs. In fact, E2 binds with nearly two-thirds as much strength as E1 to oestrogen receptors. ^{Reference Gleason, Carlsson, Johnson, Atwood and Asthana88} Interestingly, E1 is the main component of CEE, which was administered in the WHIS. Other studies have shown greater verbal memory benefits with E2 over CEE in younger postmenopausal females ^{Reference Wroolie, Kenna, Williams, Powers, Holcomb and Khaylis89} (controlled for menopause type), as well as decreased rates of ventricular expansion and white matter hyperintensity volume compared with placebo in younger, spontaneously menopausal females. ^{Reference Kantarci, Tosakulwong, Lesnick, Zuk, Lowe and Fields84} Work in ovariectomised female rodents, a surgical menopause model, has demonstrated that E2 increases, whereas E1 decreases, hippocampal neurogenesis, ^{Reference Yagi, Mohammad, Wen, Batallán Burrowes, Blankers and Galea80,Reference McClure, Barha and Galea90} and E2 facilitates, whereas premarin (E1) impairs, working memory. ^{Reference Holmes, Wide and Galea91–Reference Bimonte-Nelson, Bernaud and Koebele93} Collectively, these findings suggest E2 may be the more beneficial MHT type to promote brain plasticity and memory.

Although E2-based MHT may be more beneficial, it is important to consider that when it is taken orally, it is largely converted to E1 because of a high first-pass metabolism effect, meaning it is broken down by the liver and gut (Fig. 3). Both transdermal and vaginal routes of administration of E2 bypass first-pass metabolism, resulting in more constant hormone levels, more bioavailable E2 and an E1:E2 ratio that more closely approximates premenopausal levels compared with oral oestrogens. ^{Reference Gleason, Carlsson, Johnson, Atwood and Asthana88} Moreover, transdermal and vaginal oestrogens pose a lower risk for venous thromboembolism, as they result in lower clotting factors and have a more favourable risk profile of markers of cardiovascular heart disease than oral oestrogens. ^{Reference Goldštajn, Mikuš, Ferrari, Bosco, Uccella and Noventa94} Female rodent studies of CEE have shown that its negative effects on cognition are dose-dependent, only occurring with high CEE doses. ^{Reference Barha and Galea92,Reference Walf and Frye95} There is a strong biological rationale that transdermal and vaginal HT preparations likely have brain and cognitive benefits over oral preparations. One study found oral preparations reduced risk for several neurodegenerative diseases, ^{Reference Kim, Soto, Branigan, Rodgers and Brinton96} but type of hormones was not considered in this analysis. It remains unclear whether transdermal preparations of E2 specifically provide superior protection against neurodegenerative diseases than oral preparations.

Fig. 3 Oral versus transdermal routes of administration of oestradiol-based menopausal hormone therapy. When taken orally, oestradiol is subject to a large first-pass metabolism effect; it undergoes chemical breakdown by the gut as well as enzymatic breakdown by the liver, and is largely converted to oestrone. Transdermal routes of administration largely bypass first-pass metabolism, resulting in more bioavailable oestradiol. Several lines of research suggest that oestradiol outperforms oestrone in its brain and cognitive benefits.^{Reference Yagi, Mohammad, Wen, Batallán Burrowes, Blankers and Galea80, Reference Kantarci, Tosakulwong, Lesnick, Zuk, Lowe and Fields84, Reference Wroolie, Kenna, Williams, Powers, Holcomb and Khaylis89–Reference Holmes, Wide and Galea91, Reference Diaz-Ruano, Martinez-Alarcon, Perán, Benabdellah, Garcia-Martinez, Preda, Gonzalez-Hernandez, Marchal and Picon-Ruiz97} E1, oestrone; E2, oestradiol; NF-κB, nuclear factor-kappa B. Figure created using BioRender.com.

Similar to the type of oestrogens, the type of progestogen treatment may influence cognition. P4 is associated with greater brain and cognitive benefits than MPA, a progestin (i.e. synthetic P4). Unlike P4, MPA is associated with reduced antioxidant defences within the hippocampus, ^{Reference Irwin, Yao, Ahmed, Hamilton, Cadenas and Brinton98} and fails to increase brain-derived neurotrophic factor (BDNF) within the cortex. ^{Reference Jodhka, Kaur, Underwood, Lydon and Singh99} Moreover, unlike P4, MPA has been linked to worse spatial reference memory and worse spatial working memory in ovariectomised female rats. ^{Reference Lowry, Pardon, Yates and Juraska100} Furthermore, MPA co-administration blocks E2’s neuroprotective effects in the hippocampus and dentate gyrus. ^{Reference Chan, Chow, Hamson, Lieblich and Galea101,Reference Nilsen and Brinton102} MPA is also associated with a higher risk of venous thromboembolism compared with P4. ^{Reference Cockrum, Soo, Ham, Cohen and Snow103} These findings stress that not all MHT regimens are equal in terms of their cognitive and brain benefits.

Nearly two decades after the WHIS, its study findings continue to influence perception of MHT risks, warning labels displayed on MHT monographs and MHT prescribing practices. One lasting negative outcome of the WHIS trial includes amplified worry surrounding increased risk of breast cancer, stroke and dementia with MHT use. In fact, even warning labels on low-dose vaginal oestrogen products (e.g. 0.01 mg doses of Vagifem) list increased risk of stroke, blood clots and dementia as potential risks associated with their use (https://pdf.hres.ca/dpd_pm/00073342.PDF). However, there has been pushback from medical professionals to remove these warnings ^{Reference Manson, Goldstein, Kagan, Kaunitz, Liu and Pinkerton104} in light of scientific evidence to suggest the contrary. Vaginal oestrogens result in significantly lower systemic oestrogens compared with oral oestrogens. Moreover, a large cohort study of women found no increased risk of breast cancer-specific mortality with the use of vaginal oestrogens following breast cancer diagnosis, even in women with past oestrogen receptor-positive breast cancers. ^{Reference McVicker, Labeit, Coupland, Hicks, Hughes and McMenamin105} Similarly, another large cohort study found no increased risk of breast cancer with the use of low-dose transdermal oestrogens, ^{Reference Baik, Baye and McDonald106} as well as several benefits associated with taking oestrogens alone beyond 65 years of age. These benefits included reductions in dementia, mortality, lung cancer, colorectal cancer, congestive heart failure, venous thromboembolism, acute myocardial infarction and recurrent urinary tract infections. ^{Reference Manson, Goldstein, Kagan, Kaunitz, Liu and Pinkerton104,Reference Baik, Baye and McDonald106} These studies suggest that lower-dose oestrogens with vaginal or transdermal routes of administration have several benefits, even in older menopausal females. Thus, the benefits of MHT must be appropriately balanced with the risks for the given individual.

VMS and brain health

As discussed, there are many symptoms of menopause, all of which can influence quality of life and potentially influence cognition and the brain. VMS, night sweats and hot flashes, are the most studied, which are hypothesised to arise from declining ovarian hormone levels, particularly oestrogens, that drive changes in neurotransmitter levels and instability in the hypothalamic thermoregulatory center. ^{Reference Baker, Lampio, Saaresranta and Polo-Kantola55} Hot flashes are distinct from night sweats, but they often co-occur. ^{Reference Leidy Sievert, Makhlouf Obermeyer and Price107} The prevalence of VMS increases over the menopausal transition, with as many as 80% of postmenopausal women reporting these symptoms at a moderate to severe level. ^{Reference Avis, Crawford and Green108} Most females self-report four to five hot flashes per day, but this can be as high as 20 per day. ^{Reference Maki, Drogos, Rubin, Banuvar, Shulman and Geller109} The number of objective night-time hot flashes is approximately 3.5, and approximately 90% are accompanied by night sweats. ^{Reference Leidy Sievert, Makhlouf Obermeyer and Price107,Reference de Zambotti, Colrain, Javitz and Baker110} VMS can contribute to sleep disruption. ^{Reference de Zambotti, Colrain, Javitz and Baker110} VMS can be measured subjectively by asking participants to self-report how many hot flashes and night sweats they experienced, as well as measured objectively/physiologically via sternal skin conductance. Although there is good correspondence between these two methods of quantifying VMS, females generally underreport (rather than overreport) the number of physiologically measured hot flashes experienced. ^{Reference Maki, Drogos, Rubin, Banuvar, Shulman and Geller109} VMS are associated with reduced quality of life, but also measurable cognitive and brain changes.

Physiologically measured VMS have been associated with worse verbal memory, independent of sleep, age, ethnicity and mood, and some evidence suggests that treating VMS can improve memory performance. ^{Reference Maki and Thurston111} Beyond cognition, VMS have also been associated with biomarker and brain changes. ^{Reference Thurston, Wu, Chang, Aizenstein, Derby and Barinas-Mitchell62,Reference Thurston, Maki, Chang, Wu, Aizenstein and Derby112} In ovary-intact females not taking MHT, more severe and later-occurring self-reported VMS (i.e. VMS that appeared further from menopause onset) were associated with accelerated epigenetic ageing – a marker associated with greater physical ageing and premature death – compared with females without VMS. ^{Reference Thurston, Carroll, Levine, Chang, Crandall and Manson113} Additionally, later-occurring self-reported VMS were also associated with an increased risk of cardiovascular disease and all-cause mortality. ^{Reference Szmuilowicz, Manson, Rossouw, Howard, Margolis and Greep114} Interestingly, if VMS appeared closer to the onset of menopause, these associations did not persist. Thus, there are many temporal trajectories of VMS onset, as well as varying degrees of VMS severity, that influence cognition, Alzheimer’s disease brain biomarkers and cardiovascular health.

Additionally, experiencing more VMS has been linked to greater white matter hyperintensity burden in women currently not taking MHT. More frequent physiologically recorded hot flashes were associated with greater whole brain white matter hyperintensities that were not explained by cardiovascular disease risk factors, depressive symptoms, endogenous E2 or sleep fragmentation. ^{Reference Thurston, Wu, Chang, Aizenstein, Derby and Barinas-Mitchell62} A subregion analysis showed that VMS were most strongly associated with white matter hyperintensity volume in the frontal lobe. On the contrary, self-reported hot flashes were not associated with white matter hyperintensity volume. ^{Reference Thurston, Wu, Chang, Aizenstein, Derby and Barinas-Mitchell62} Together, these findings suggest that the frontal lobe may be most vulnerable to VMS-induced brain changes and physiologically recorded VMS may be a more sensitive predictor of brain changes. Although MHT is considered the most effective first-line treatment for alleviating VMS by the Society for Obstetricians and Gynecologists of Canada, the American College of Obstetricians and Gynecologists and the Canadian Cancer Society of Canada, there is insufficient evidence as to whether MHT is helpful for preventing white matter hyperintensity burden. One recent study found that any MHT use beyond 5 years postmenopause was associated with greater white matter hyperintensity burden; ^{Reference Lee, Raghavan, Christenson, Frank, Kantarci and Rocca115} however, this seemingly contradictory linkage of MHT to greater white matter hyperintensity burden may be related to the later initiation of MHT and lack of consideration of MHT types. More objectively measured VMS have also been associated with greater brain amyloid-β pathology in women not taking MHT, ^{Reference Thurston, Maki, Chang, Wu, Aizenstein and Derby112} suggesting that more VMS may increase Alzheimer’s disease pathology.

In some females, persistent hot flashes are experienced beyond 60 years of age; one study found that 10% of females continue to experience hot flashes in their 60s and 5% in their 70s. ^{Reference Brunner, Aragaki, Barnabei, Cochrane, Gass and Hendrix116} This suggests that VMS can persist beyond the recommended time of MHT cessation, which is age 60 years or >10 years postmenopause. Given the association of VMS with brain changes, this guideline should be reconsidered. Importantly, VMS also cause sleep disruption, and greater sleep disruption has been associated with increased white matter hyperintensities independent of VMS. ^{Reference Thurston, Wu, Aizenstein, Chang, Barinas Mitchell and Derby117} It is possible that VMS may negatively influence cognition and the brain through their effects on sleep. ^{Reference Maki and Thurston111} Sleep problems are common in the menopause transition, and great care should be taken to treat these sleep problems given their strong associations with neurodegenerative disease. ^{Reference Baker, Lampio, Saaresranta and Polo-Kantola55} Furthermore, the presence of severe menopausal symptoms was related to an increased odds for developing mild cognitive impairment (MCI), a prodromal state to Alzheimer’s disease. ^{Reference Calle, Blümel, Chedraui, Vallejo, Belardo and Dextre118} Females with MCI experience more severe menopausal symptoms across all three subscales (somatic-vegetative, psychological and urogenital symptoms) of the MRS-II than females without MCI (mixture of menopause types), ^{Reference Calle, Blümel, Chedraui, Vallejo, Belardo and Dextre118} independent of age, years of education and body mass index. Importantly, this suggests that menopausal symptoms beyond hot flashes and night sweats should be investigated for their potential role in cognitive and brain changes and treated aggressively. Knowing which menopausal symptoms might be driving increased risk for MCI is key for developing effective therapeutics.

Apolipoprotein 4 genotype

Apolipoprotein E (APOE) plays a role in fat metabolism and cholesterol transport. Carriers of the APOE4 variant, which is associated with disrupted cholesterol homeostasis, ^{Reference Mahley119} are at a significantly greater risk for sporadic (i.e. non-familial) Alzheimer’s disease. Homozygous APOE4 carriers show as high as a 15-fold increased risk for Alzheimer’s disease compared with non-carriers, with heterozygous and homozygous female carriers at even greater risk than male carriers. ^{Reference Farrer, Cupples, Haines, Hyman, Kukull and Mayeux120} In addition, female carriers show more Alzheimer’s disease neuropathology (i.e. greater cerebrospinal fluid amyloid-β, greater total tau and elevated tau-to-amyloid-β ratio) compared with male carriers. ^{Reference Duarte-Guterman, Albert, Barha and Galea121} Thus, APOE4 status is a non-modifiable risk factor for Alzheimer’s disease that confers a greater risk in women.

Interestingly, APOE4 status may influence MHT’s effect on the brain and cognition. Some findings suggest that female APOE4 carriers may benefit more from MHT than non-carriers. Current or past MHT (oestrogens alone or combined with progestogens) was associated with greater entorhinal and amygdala volumes, as well as higher delayed memory score, in APOE4 carriers only. ^{Reference Saleh, Hornberger, Ritchie and Minihane122} An earlier age of MHT initiation was also associated with larger hippocampal volumes in APOE4 carriers only, suggesting that they may be more responsive to MHT. ^{Reference Saleh, Hornberger, Ritchie and Minihane122} Indeed, MHT initiation before menopause is associated with less brain ageing, but only in APOE4 carriers. ^{Reference de Lange, Barth, Kaufmann, Maximov, van der Meer and Agartz123} Moreover, a recent prospective clinical study showed that APOE4 carriers had significantly lower pathophysiological Alzheimer’s disease plasma biomarkers if they took MHT, compared with those who did not. ^{Reference Depypere, Vergallo, Lemercier, Lista, Benedet and Ashton124} Together, these studies suggest that MHT may have more brain benefits for APOE4 carriers.

Despite these findings, other studies have found the opposite. APOE4 carriers with past or current MHT showed faster cognitive decline than those who never used MHT. ^{Reference Kang and Grodstein125} Furthermore, a few rodent studies suggest that E2 has more benefits for neuroprotection following ovariectomy in non-carriers as compared with APOE4 carriers. ^{Reference Balu, Valencia-Olvera, Deshpande, Narayanam, Konasani and Pattisapu126,Reference Taxier, Philippi, Fleischer, York, LaDu and Frick127} It remains to be determined why these differential effects are seen. One potential explanation is that these studies in rodents were done in young animals, suggesting perhaps a biphasic MHT effect in APOE4 carriers; oestrogens may be detrimental earlier in life, but beneficial later in life. In fact, the effects of the APOE4 allele itself on cognition vary over time. Carrying APOE4 is associated with cognitive benefits in midlife, particularly for executive functions, but with cognitive risks at age 65 years and beyond. ^{Reference Gharbi-Meliani, Dugravot, Sabia, Regy, Fayosse and Schnitzler128} Given that APOE4’s effects on cognition may vary over time, it is not surprising that MHT may have differential brain effects in APOE4 carriers over time. Moreover, menopause type has not been investigated. Most rodent studies of menopause use the ovariectomy model of menopause, a form of induced menopause that has more severe cognitive consequences than natural/spontaneous menopause. Also, one study found that only carriers with two APOE4 alleles – but not one APOE4 allele – had negative responses to MHT, suggesting that the number of APOE4 alleles may influence MHT effects. ^{Reference Taxier, Philippi, Fleischer, York, LaDu and Frick127} Future work that considers age, menopause type and number of APOE4 alleles – and their interactions – is needed to further our understanding of MHT’s effects on the brain and cognition.

Modelling human menopause in rodents

Having appropriate, well-designed models of menopause in non-human animals is necessary for furthering research on MHT and developing tailored MHT treatments that translate to the best ageing outcomes in females. In the following section, we discuss rodent models of human menopause. In rodents, natural ageing, ovariectomy and chemical induction of menopause are used to model different types of human menopause. The model chosen has an influence on outcomes, much like the differential effects of menopause type on MHT effects or disease risk (Table 1). Although it is beyond the scope of this review, non-human primate models of menopause are also serve as valuable means to further our knowledge on MHT and female brain ageing. ^{Reference Lacreuse, Mong and Hara129,Reference Hara, Waters, McEwen and Morrison130} We urge the research community to consider both the menopause model and type of menopause in humans to further scientific discovery.

Table 1 Summary of rodent menopause models and their important considerations

VCD, 4-Vinylcyclohexene diepoxide.

Recommendations and future directions

There is not one type of perimenopause, menopause or postmenopause. Similarly, there is not one type of MHT. Given this, it is unrealistic to expect that one type of MHT will equally benefit all females, who have different genetics and varied life experiences. Females and healthcare practitioners alike must appreciate these nuances to ensure all females receive informed, individualised menopausal care to support their healthy brain ageing.

This starts with improved menopausal education. Approximately 65% of females do not feel prepared for menopause, ^{Reference Marlatt, Beyl and Redman152} and most general practitioners acknowledge that they are lacking menopause education, especially education surrounding menopause management and various forms of MHT. ^{Reference Davis, Herbert, Reading and Bell153} One survey found that only 6.8% of USA medical practitioners felt sufficiently prepared to support females through the menopause transition. ^{Reference Kling, MacLaughlin, Schnatz, Crandall, Skinner and Stuenkel154} In fact, only 31.3% of USA obstetrics and gynaecology residency programmes reported having a menopause curriculum. ^{Reference Allen, Laks, Zahler-Miller, Rungruang, Braun and Goldstein155} This is concerning, knowing that 38% of females feel that their menopausal symptoms are undertreated, ⁵¹ and in Canada alone, 540 000 days of work are lost each year because of ineffective menopausal symptom management. ⁵¹ Unmanaged menopausal symptoms negatively influence quality of life and have negative economic consequences. This lack of menopause education increases the possibility that menopausal symptoms, especially if they deviate from the commonly reported VMS, may be left unrecognised or mistakenly attributed to other causes. In current medical practice, the onus falls on the patients to determine when their menopause transition is commencing and to advocate for treatments to help ease this transition. There are significant delays in diagnosis across disorders in females compared with males, and females feel their symptoms are dismissed in healthcare practitioner offices. ^{Reference Westergaard, Moseley, Sørup, Baldi and Brunak156} Coupled with the findings that healthcare practitioners are not equipped with knowledge on menopause, it is imperative that more research and education of healthcare practitioners takes place. Given that 50% of the population will go through menopause and that just less than half of the female lifespan will be spent in a perimenopausal or postmenopausal state, it is shocking so little attention has been paid to menopause care.

Compounded bioidentical hormone therapy (cBHT) has exploded in popularity, partly because of the lack of appropriate care by healthcare practitioners and physician prescriptions are not needed to obtain it. cBHT is taken by up to 40% of females in the USA. ^{Reference Stuenkel and Manson157} cBHT is marketed as a more ‘natural’ option; however, it undergoes chemical extraction and stabilisation processes. ^{Reference Jackson, Parker and Mattison158} It is overlooked is that many prescribed MHTs also use natural hormones of oestradiol, E1 or even oestriol and P4. Moreover, cBHT is not approved by Health Canada, meaning that their safety and efficacy have not been evaluated with the same standards as prescribed MHT. In fact, there is no procedure to ensure cBHT contains the listed amount of active ingredient, and many cBHT formulations use E1. ^{Reference Jackson, Parker and Mattison158} The safety and efficacy of prescribed MHT is influenced by the type, route of administration and dose of oestrogens and progestogens. These same factors influence the safety and efficacy of cBHT, but they have not been formally regulated. Current guidelines suggest that more research is needed on the efficacy of cBHT, and it is recommended that their use be restricted to patients with a known allergy to an active pharmaceutical agent. ^{Reference Jackson, Parker and Mattison158}