Abstract
Abstract
The objective of this review was to explore the effect of testosterone therapy on objective and subjective cognition during the menopause transition and postmenopause through a systematic review and meta-analysis. A systematic search was conducted in PsycINFO, MEDLINE, Embase, Web of Science and the Cochrane Library from database inception to March 2025 using predefined eligibility criteria. A total of 3,880 records were identified. After screening, five studies were eligible for inclusion. Three studies including 244 postmenopausal participants were included in a random-effects meta-analysis to estimate pooled effects. Three cognitive domains were covered: executive function, verbal learning and memory and visuospatial learning and memory. Studies excluded from the meta-analysis were reported narratively. The meta-analysis found no statistically significant pooled effects of testosterone therapy across the cognitive domains analysed, with effect sizes mostly close to zero. No studies explored the effect of testosterone therapy on subjective cognition or during perimenopause. The findings should be interpreted as insufficient to establish an effect rather than as evidence of no effect, given the small number of available studies and their limited sample sizes. Further adequately powered trials are needed to investigate the effects of testosterone therapy on cognition, particularly during the menopause transition, the period in which cognitive problems appear to be most pronounced.
Lay summary
Menopause is a stage of life that can affect thinking and memory, often colloquially referred to as ‘brain fog’. Brain fog can affect daily life and relationships. Hormone treatments are often used to manage symptoms, but it is still unclear how well they work. Testosterone is sometimes prescribed for low sex drive, but its effects on thinking and memory remain unclear. This study reviews the existing evidence. Five studies were identified. These studies examined different areas of thinking and memory, including attention, planning, decision-making and learning. We found no clear evidence that testosterone affects thinking and memory. However, only a few small studies were available, so the results need to be interpreted carefully. This research highlights a knowledge gap. Further studies are needed on testosterone to guide effective treatment and support women to make informed decisions about their health.
Keywords: menopause, testosterone, menopause-replacement-therapy, cognition
Introduction
Menopause transition (MT) and postmenopause are life stages in a woman’s midlife marked by hormonal changes. The term ‘women’ is used in this paper to maintain consistency and clarity in relation to sex-based aspects of health (Gribble et al. 2022). The authors acknowledge that different terms may be appropriate in different contexts and for different purposes. MT refers to the period between the onset of menstrual changes (perimenopause) and 12 consecutive months without a menstrual period (menopause). Symptoms during these stages may include vasomotor symptoms (e.g. hot flushes and night sweats) (Gold et al. 2006), sleep difficulties (Kravitz et al. 2003), as well as depression (Bromberger et al. 2007) and anxiety (Bromberger et al. 2013). With increasing life expectancy, many women will spend a third of their lives in the postmenopausal stage (Schmidt et al. 2025).
Cognition in menopause transition and postmenopause
MT and postmenopause have also been associated with cognitive complaints, colloquially described as ‘brain fog’. This includes difficulties recalling words and numbers, misplacing items, trouble concentrating and forgetfulness (Sullivan & Woods 2001). These may be subjective (women noticing or worrying about changes in their cognition) or objective (changes observed on neuropsychological tests). Subjective complaints are frequently reported in research (Schaafsma et al. 2010, Bazeley et al. 2022), whereas the research is less unanimous for objective complaints and shows variation across cognitive domains (Unkenstein et al. 2016, Kilpi et al. 2020, Maki & Weber 2021, Naysmith et al. 2026). It has been suggested that subjective cognitive symptoms are weakly associated with objective performance, but moderately related to psychological symptoms (Naysmith et al. 2026).
Hormonal changes in menopause transition and postmenopause
Changes in sex hormones characterise menopause. Testosterone levels do not acutely change during menopause but instead begin to decline from a woman’s mid-20s to early 30s (Davis & Wahlin-Jacobsen 2015, Australian Menopause Society 2019). Sex hormone changes during this period are not uniform and are influenced by factors such as race, ethnicity and body mass index (Tepper et al. 2012).
Testosterone contributes to sexual desire in women (Wåhlin-Jacobsen et al. 2015), and research also indicates that it may play a role in cognition (Barrett-Connor & Goodman-Gruen 1999). Rosario et al. (2011) suggested that testosterone may have neuroprotective effects in men. However, corresponding studies in women are limited and not directly comparable (Dratva et al. 2024).
MHT and cognition
Menopause hormone therapy (MHT) is used to manage menopausal symptoms (NICE 2024). In the UK and Europe, MHT includes oestrogen and progestogen treatment. The British Menopause Society (BMS) advises that testosterone be prescribed for low libido when there has been no improvement with systemic MHT and vaginal oestrogen. For this review, testosterone therapy was considered within the category of MHT, acknowledging that clinical regulatory frameworks vary across the world. As hormone levels fluctuate during different stages of menopause, the dose, delivery method and duration of MHT vary between individuals (Talaulikar 2022).
A recent systematic review of oestrogen-only and combined oestrogen–progestogen therapy (MHT) across 34 studies involving 14,914 women receiving MHT and 12,679 controls found no overall effect on cognition (Andy et al. 2024). However, further analysis suggested that oestrogen-only therapy may improve cognition in women following surgical menopause (Andy et al. 2024). This review also suggested that the timing of MHT initiation is important, with improvements observed when MHT is initiated in midlife or close to menopause onset rather than later in life, consistent with the ‘critical window hypothesis’.
Testosterone prescribing has anecdotally increased across NHS and private clinics in the UK, reflecting growing interest in its role during menopause. Women are increasingly asking questions and seeking support during this stage of life. Increased visibility of menopause in social and mainstream media may have helped reduce some elements of stigma associated with menopause more generally and contributed to interest in testosterone therapy. The quality of evidence and information shared through media outlets varies, meaning that some may be helpful while others may be misleading. Current BMS, NICE and IMS guidance only supports testosterone therapy for low libido (NICE 2024, BMS 2026, IMS 2026).
A previous Cochrane review by Somboonporn et al. (2005) explored whether adding testosterone to conventional postmenopausal hormone therapy was effective by comparing testosterone plus MHT with MHT alone. Although an important early contribution, this review treated cognition as only a secondary outcome. The authors did not evaluate testosterone therapy against a non-active control group, which may be important for detecting subtle cognitive effects, and did not explore differences between objective and subjective cognitive outcomes. More recently, Islam et al. (2019) reviewed the safety and efficacy of testosterone therapy in women, with cognitive performance as one of several primary outcomes. This review permitted active comparator trial arms and did not explore differences between objective and subjective cognition.
Given these limitations, a more focused and up-to-date review is warranted.
Rationale and aims
This systematic review and meta-analysis aimed to explore the effect of testosterone therapy on cognition in women during MT and postmenopause.
The key research questions were as follows:
Is testosterone therapy associated with objective or subjective changes in cognition in women during MT and postmenopause?
Do testosterone treatment characteristics (including dose, formulation and delivery method), age at initiation or duration of use, affect cognitive outcomes?
Method
Search strategy and selection criteria
The review was registered on PROSPERO in February 2025 (registration number CRD420250644676) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al. 2021). Five databases were searched from inception to 25 March 2025: PsycINFO, MEDLINE, Embase, Web of Science and the Cochrane Library. As part of the search process, the reference lists of relevant review papers were also checked to identify any additional eligible studies.
Inclusion and exclusion criteria were prespecified and followed the Population, Intervention, Comparison, Outcome (PICO) framework. Searches were limited to peer-reviewed articles involving human participants and published in English. Grey literature was excluded. Full search terms are available in Supplementary Table 1 (see section on Supplementary materials given at the end of the article).
Studies were included if they i) were RCTs, non-randomised intervention studies or observational studies; ii) included women in MT, postmenopause or surgically or medically induced menopause. Menopause status was defined by each study’s criteria, such as confirmation of amenorrhoea/surgical procedure or serum hormone thresholds), iii) administered testosterone therapy alone or in combination with stable background oestrogen therapy across all arms; iv) reported details of testosterone therapy (dose, formulation and delivery method), participant age and intervention duration; v) included a non-active control or comparison group (e.g. waitlist control, standard care or placebo) to ensure that between-group differences reflected testosterone therapy and were not confounded by the effects of another active treatment; and vi) used a valid and reliable measure of objective or subjective cognition. Studies were excluded if they i) fell outside of MT or postmenopause; ii) included women with neurological disorders or neurodegenerative conditions affecting cognition; iii) used medications or psychotherapies targeting cognition (e.g. stimulant medication); iv) were review papers, background information papers or opinion pieces without primary data; or v) used an active hormonal comparison. Eligibility screening was performed in Rayyan (https://www.rayyan.ai) by the first reviewer (AF), with a second reviewer (SAE) cross-checking 10% of the results. Any discrepancies were resolved through discussion.
Quality assessment
The CASP RCT and Cohort Checklist (Critical Appraisal Skills Programme 2023) was used to assess the overall methodological quality of each study. The findings were not used as a scorecard or exclusion criterion but as a guide to assessing each paper, as recommended by the Cochrane handbook (Higgins et al. 2024). See Tables 1 and 2.
Table 1.
CASP randomised controlled trial checklist.
| Davis et al. (2014) | Huang et al. (2015) | Kocoska-Maras et al. (2011) | Sherwin (1988) | |
|---|---|---|---|---|
| Did the study address a clearly formulated research question? | Yes | Yes | Yes | Yes |
| Was the assignment of participants to interventions randomised? | Yes | Yes | Yes | Yes |
| Were all participants who entered the study accounted for at its conclusion? | Yes | Yes | Yes | Yes |
| Were the participants ‘blind’ to intervention they were given? | Yes | Yes | Yes | Yes |
| Were the investigators ‘blind’ to the intervention they were giving to participants? | Yes | Yes | Yes | Yes |
| Were the people assessing/analysing outcome/s ‘blinded’? | Yes | Yes | Yes | Yes |
| Were the study groups similar at the start of the randomised controlled trial? | Yes | Yes | Yes | Yes |
| Apart from the experimental intervention, did each study group receive the same level of care (that is, were they treated equally)? | Yes | Yes | Yes | Yes |
| Were the effects of intervention reported comprehensively? | Yes | Yes | Yes | No |
| Was the precision of the estimate of the intervention or treatment effect reported? | Yes | Yes | Yes | No |
| Do the benefits of the experimental intervention outweigh the harms and costs? | Yes | No | No | Yes |
| Can the results be applied to your local population/in your context? | Yes | No | Yes | No |
| Would the experimental intervention provide greater value to the people in your care than any of the existing interventions? | Yes | No | No | Yes |
Table 2.
CASP cohort studies.
| Checklist | Davison et al. (2011) |
|---|---|
| Are the results valid? | Yes |
| Was the cohort recruited in an acceptable way? | Yes |
| Were all being the exposure accurately measures to minimise bias? | Yes |
| Was the outcome accurately measured to minimise bias? | Yes |
| Have the authors identified all important cofounding factors? | Yes |
| Have they taken account of the confounding factors in the design and/or analysis? | Yes |
| Was the follow-up of subjects complete enough? | Yes |
| What are the results of this study? | Yes |
| How precise are the results? | Yes |
| Do you believe the results? | Yes |
| Can the results be applied to a local population? | Yes |
| Do the results of this study fit with other available evidence? | Yes |
| What are the practical implications of this study? | Yes |
Data extraction
Data were extracted using a standardised extraction template that included study characteristics (e.g. study design, country, menopause stage and menopause stage classification tool), intervention details (e.g. delivery method, dose and duration), comparison group and cognitive outcomes (e.g. objective and subjective measures), domain classification, post-intervention means (M), standard deviations (SD) and sample sizes (n).
Where reported, change scores (M and SD) were extracted. Where studies reported interquartile ranges (IQRs) instead of SD, SD was approximated using the formula SD ≈ IQR/1.35 (Wan et al. 2014). Where medians rather than means were reported, these were extracted without conversion. When data were incomplete, the authors were contacted to request the raw or supplementary data. Papers with missing data were excluded from the meta-analysis and reported narratively. Each data entry was double-checked by a second rater (RD), and any discrepancies were discussed.
Meta-analysis
Cognitive outcomes reported across these studies were categorised into corresponding cognitive domains based on the information provided in the original papers, cross-checked against a reliable compendium (Strauss et al. 2006) and discussed with a neuropsychologist and a research fellow. Cognitive domains represented by a single study were excluded from the meta-analyses and analysed narratively.
This process resulted in three cognitive domains: i) executive function, ii) visuospatial learning and memory and iii) verbal learning and memory. When multiple tests were reported, one measure per domain per study was selected to avoid violating the assumption of independent data points. Predefined ‘decision rules’ were applied to determine which test to include within each domain (Higgins et al. 2024). Specifically, the most similar objective tests were selected for each domain, with similarity defined primarily by the specific cognitive area assessed and the testing procedures used (see Table 3 and Supplementary data). Where multiple testosterone doses were reported in Huang et al. (2015), one dose was selected to maximise comparability with doses used in other studies and reduce heterogeneity by avoiding further dose variation, consistent with an exploratory meta-analysis. A meta-analysis was conducted where two or more non-overlapping studies reported data that could be used to calculate an effect size for an outcome. Hedges’ g effect sizes were calculated using the Campbell Collaboration effect size calculator (Campbell Collaboration 2023). A random-effects model was used to account for between-study heterogeneity. Heterogeneity was assessed using I2 and Q statistics.
Table 3.
Summary of objective cognitive tests used in the meta-analysis.
| Reference | Objective cognitive domain | Objective cognitive test | Description of test |
|---|---|---|---|
| Davis et al. (2014) | Executive function | GMLT | The GMLT is a computerised version of the HPML task paradigm, stepping stone mazes. In this task a pathway must be located and learnt over successive trials within a grid of tiles according to a set of simple searching rules. A lower score equals a better performance |
| Huang et al. (2015) | Executive function | Trail making B test | Participants are asked to connect numbers (1–13) and letters (A–L) alternately (1-A–2-B…) on a piece of paper as quickly as possible (Huang et al. 2015). Scoring is expressed in terms of the time in seconds required for completion of each task (Strauss et al. 2006). A lower score equals a better performance |
| Davis et al. (2014) | Verbal learning and memory | ISLT immediate | The ISLT is part of CogState, a computerised battery of cognitive tests. The ISLT is a 12 word, three trial word list learning test in which individuals are instructed to remember a list of items that are found commonly in their local stores. A higher score equals better performance |
| Huang et al. (2015) | Verbal learning and memory | BSRT immediate | Participants are read 12 words and asked to recall as many words as possible. Subsequent trials include only those words that were not recalled on the immediately preceding trial (Huang et al. 2015). A higher score equals better performance |
| Kocoska-Maras et al. (2011) | Verbal learning and memory | Noun recall | Noun recall is a test of episodic verbal memory. A list of 12 common unrelated nouns are presented verbally with the instruction to remember as many as possible for an immediate free recall test. The words should be presented at two second intervals. A higher score equals better performance |
| Davis et al. (2014) | Visuospatial learning and memory | Continuous paired associate learning | The continuous paired associate learning is also from part of the CogState, a computerised battery of cognitive tests. In this task participants must match abstract patterns to their locations on the screen. A higher score equals better performance |
| Huang et al. (2015) | Visuospatial learning and memory | VSLT immediate | The VSLT requires the participant to learn to recognise seven abstract designs and recall the correct placement of each design on 6 × 4 matrix grid. After being shown the grid and correct placement, designs are removed and participants place the designs in the correct location. There are five learning trials and a recall administration after a 30-min delay (Huang et al. 2015). A higher score equals better performance |
GMLT, Groton Maze Learning Task; HPML, Hidden Pathway Maze Learning; ISLT, International Shopping List; VSLT, visual spatial learning test; BSRT, Buschke Selective Reminding Test.
Analyses were conducted in R (v4.5.2) using the metafor package (v4.8-0) (Viechtbauer 2010).
Narrative review
A narrative review was conducted when studies were not eligible for quantitative pooling, or when only a single study contributed data to a cognitive domain, making meta-analysis impossible. In addition, when studies contributed multiple tests within the same cognitive domain, only one outcome was included in the meta-analysis, as described above. The remaining outcomes were reported narratively.
Testosterone doses were standardised to 1 mg/day for comparability across studies, following consultation with an endocrinologist (NK), with the acknowledgement that this does not represent biological equivalence but allows for descriptive comparison.
Results
In total, 3,880 studies were identified through database searches (see Fig. 1). After removal of duplicates (n = 1,014), 2,866 studies were screened by title and abstract, and 56 full texts were assessed. A total of five papers were eligible for inclusion, of which three were eligible for inclusion in the meta-analyses. No additional studies were identified through backward reference sampling of relevant reviews. No studies were identified that examined subjective cognition in women during MT. All included studies involved postmenopausal women. Study characteristics are summarised in Table 4.
Figure 1.

PRISMA 2020 flow diagram of the study selection process.
Table 4.
Characteristics of included studies.
| Characteristics | Davis et al. (2014) | Davison et al. (2011) | Huang et al. (2015) | Kocoska-Maras et al. (2011) | Sherwin (1988) |
|---|---|---|---|---|---|
| Country | Australia | Australia | USA | Sweden | Canada |
| Design | RCT | Non-randomised | RCT | RCT | RCT-crossover |
| Menopause stage | Postmenopausal | Postmenopausal | Postmenopausal | Postmenopausal | Postmenopausal |
| Menopause classification | Atleast 12 months of amenorrhoea | 55 years or greater or bilateral oophorectomy or >12 months amenorrhoea and FSH > 20 IU/L, or treated for climacteric symptoms with postmenopausal oestrogen therapy for >12 months | Hysterectomy alone or partial oophorectomy if their FSH levels were ≥30 U/l or if they were already receiving oestrogen therapy | At least 12 months of amenorrhoea or serum levels of FSH >30 U/L | Total abdominal hysterectomy + bilateral salpingo-oophorectomy |
| Recruitment age (years) | 55–65 | 45–60 | 41–62 | 50–62 | N/A |
| Average age (years) | Median: 60 | Mean: 57.2 | Mean: 54 | Range: 50–65; T: 58.3 (mean); C: 5,801 (mean) | 45.4 |
| Intervention | Testosterone therapy for naturally menopausal women | Testosterone therapy for natural and surgical menopausal women | Testosterone therapy for women who had undergone hysterectomy with or without partial or total oophorectomy | Testosterone therapy for naturally menopausal women | Testosterone therapy for surgical menopause |
| Stabilising period pretrial | None | Stable oestrogen therapy for ≥8 weeks prior to testosterone | Transdermal oestrogen for 12 weeks before randomisation | None reported | None reported |
| Testosterone delivery method | Transdermal gel | Transdermal spray | Intramuscular injection | Oral | Injection |
| Testosterone formulation | LibiGel | Unreported | Enanthate | Undecanoate | Enanthate |
| Testosterone dose | 0.3 mg/day | 4.55 mg/day delivered (0.3 mg/day absorbed) | 25 mg*/week | 40 mg/day | 200 mg/28 days |
| Treatment duration | 26 weeks | 26 weeks | 24 weeks | 4 weeks | 2 × 3 months with 1 month placebo washout |
| Comparator | Placebo gel | Treatment as usual control | Placebo injection | Placebo tablet | Placebo injection |
| Total sample size | 89 | 39 | 21* | 134 | 20 |
RCT, randomised controlled trial; T, testosterone therapy; C, comparator; N/A, data not available.
Individual dose and corresponding sample size selected for meta-analysis.
Objective cognition
Three studies that included measures of objective cognition were eligible for inclusion in the meta-analysis, all of which sampled postmenopausal women. Two studies were not eligible for meta-analysis but were eligible for narrative review. Publication bias could not be meaningfully assessed because there were fewer than ten studies in each analysis and funnel plots could not be reliably evaluated.
A separate random-effects meta-analysis was conducted for each of the three predefined cognitive domains, as defined a priori. Forest plots were used to graphically represent the meta-analysis. Forest plots showing standardised mean differences (SMD) of Hedges’ g and 95% confidence intervals are presented in Figs 2 and 3.
Figure 2.

Forest plot for the cognitive domain of executive function.
Figure 3.

Forest plots for the cognitive domains of verbal learning and memory and visuospatial learning and memory.
Across all three cognitive domains, no statistically significant effects of testosterone on cognitive outcomes were observed in postmenopausal women. Heterogeneity varied across the cognitive domains, reflecting between-study differences in intervention and participant characteristics.
Executive function
A meta-analysis of two studies (Davis et al. 2014, Huang et al. 2015) (n = 110) found a pooled Hedges’ g of 0.03 (95% CI: −0.34 to 0.41), indicating no statistically significant pooled effect. Heterogeneity was low, with no significant variation between the samples (Q = 0.73, I2 = 0%, P = 0.39).
Verbal learning and memory
A meta-analysis of three studies (Kocoska-Maras et al. 2011, Davis et al. 2014, Huang et al. 2015) (n = 244) found a pooled Hedges’ g of −0.00 (95% CI: −0.34 to 0.33), indicating no statistically significant pooled effect. Heterogeneity was moderate (Q = 2.73, I2 = 35%, P = 0.26).
Visuospatial learning and memory
A meta-analysis of two studies (Davis et al. 2014, Huang et al. 2015) (n = 110) found a pooled Hedges’ g of 0.00 (95% CI: −0.60 to 0.60), indicating no statistically significant pooled effect. Heterogeneity was moderate (Q = 1.74, I2 = 42%, P = 0.19).
GRADE certainty assessment rated the evidence for all outcomes as very low due to risk of bias, inconsistency, indirectness and imprecision. See Table 5 for the certainty assessment.
Table 5.
GRADE certainty of evidence.
| Outcome | Number of | Effect estimate* (95% CI) | Certainty assessment | Comments | Certainty of evidence | |
|---|---|---|---|---|---|---|
| Studies | Participants | |||||
| Executive function | 2 | 110 | 0.03 (−0.34, 0.41) | Serious risk of bias | Small samples limiting power to detect effects. Some limited information on randomisation procedures in some papers. Different populations (surgical menopause vs natural menopause), different testosterone interventions. Wide confidence interval spanning both sides of the null with different clinical implications | Very low |
| No serious inconsistency | ||||||
| Serious indirectness | ||||||
| Serious imprecision | ||||||
| Cannot assess publication bias | ||||||
| Verbal learning and memory | 3 | 244 | 0.00 (−0.34, 0.33) | Serious risk of bias | Small samples limiting power to detect effects. Some limited information on randomisation procedures in some papers. Moderate heterogeneity and variation in effect size. Variation in menopause populations (surgical vs natural), variation in testosterone treatment. Wide confidence interval spanning both sides of the null with different clinical implications | Very low |
| Serious inconsistency | ||||||
| Serious indirectness | ||||||
| Serious imprecision | ||||||
| Cannot assess publication bias | ||||||
| Visuospatial learning and memory | 2 | 110 | 0.00 (−0.60, 0.60) | Serious risk of bias | Small samples limiting power to detect effects. Some limited information on randomisation procedures in some papers. Moderate heterogeneity and variation in effect size. Variation in menopause populations (surgical vs natural), variation in testosterone treatment. Wide confidence interval spanning both sides of the null with different clinical implications | Very low |
| Serious inconsistency | ||||||
| Serious indirectness | ||||||
| Serious imprecision | ||||||
| Cannot assess publication bias | ||||||
Values are Hedges’ g (95% CI).
Narrative review
Papers that could not be included in the meta-analysis were analysed narratively. Tests included in the meta-analysis were not reported narratively again. Sherwin (1988) was excluded due to the absence of extractable summary statistics and reported separately due to substantial methodological and measurement differences. Davison et al. (2011) was also excluded from the meta-analysis due to its non-randomised design, which did not meet the eligibility criteria for meta-analysis. In cognitive domains where only a single study contributed data, meta-analysis was not possible and findings were therefore reported narratively (Kocoska-Maras et al. 2011, Davis et al. 2014, Huang et al. 2015). Huang et al. (2015) reported multiple testosterone doses. The 25 mg arm was used for the meta-analysis, while the remaining doses (3, 6.25 and 12.5 mg) were included in the narrative review.
Attention and processing speed
Three studies reported outcomes in this domain. Davison et al. (2011) assessed this domain using the CogState detection task. They found no change in scores from baseline to week 26 in the testosterone arm and no difference compared with the normative control group. Bootstrapped regression, adjusted for age and baseline score, confirmed the absence of a treatment effect (β = 0, 95% CI: (−0.03, 0.04), P = 0.872). Davis et al. (2014) used the same CogState detection task and also reported no significant between-group difference in change from baseline to week 26 after adjusted for age and baseline scores (β = −0.01, 95% CI: (−0.02, 0.01), P = 0.38).
Executive function
Huang et al. (2015) additionally assessed executive function using the Stroop Colour Word Interference Test and Mazes. No significant changes (all P > 0.05) were observed in any testosterone treatment arm compared with placebo, even after adjustment for baseline cognitive function, age, and education. Davison et al. (2011) found no significant improvements in the Groton Maze Learning Task after 26 weeks of treatment. Regression analysis adjusted for age and baseline performance confirmed the absence of treatment effects (β = −4.24, 95% CI: (−13.04, 3.60), P = 0.308).
Language and verbal fluency
Only Huang et al. (2015) reported tests in this domain, using the Category Fluency and Phonemic Fluency. There were no significant differences observed between any testosterone treatment arms and the placebo group (all P > 0.05), even after adjustment for age, education and baseline cognitive function.
Spatial cognition
Kocoska-Maras et al. (2011) assessed spatial cognition using the Mental Rotation Task. They reported no significant differences between the testosterone treatment arm and the placebo group after 4 weeks, even after adjustment for age (mean difference: 0.34, 95% CI: (−1.48, 2.17), P > 0.05).
Verbal learning and memory
Huang et al. (2015) additionally assessed verbal learning and memory using Paragraph Recall (immediate and delayed) and the Buschke Selective Reminding Test. No significant changes were observed in any testosterone treatment arm compared with placebo, even after adjustment for baseline cognitive function, age and education (all P > 0.05). Davis et al. (2014) used the International Shopping List Test (delayed recall condition) and reported no statistically significant effects between the testosterone arm and the normative control arm after 26 weeks, following adjustment for age and baseline scores (β = −0.11, 95% CI: (−0.84, 0.61), P = 0.77). Davison et al. (2011) also reported scores for the International Shopping List Test (immediate and delayed recall conditions). In the delayed recall condition, week 26 scores were significantly higher in the testosterone arm than in the normative control group (P < 0.05). Regression analyses adjusted for age and baseline performance confirmed significant treatment effects for both immediate and delayed recall conditions (immediate: β = 2.03, 95% CI: (0.49, 3.43), P = 0.007; delayed recall: β = 1.27, 95% CI: (0.11, 2.18), P = 0.015).
Visuospatial learning and memory
Huang et al. (2015) additionally assessed visuospatial learning and memory using the Puget Sound Route Learning Test (immediate and delayed) and the Visual Spatial Learning Test (delayed). No significant changes were observed in any testosterone treatment arm compared with placebo, even after adjustment for baseline cognitive function, age and education (all P > 0.05). Davison et al. (2011) reported improvements from baseline to 26 weeks in the Groton Maze (recall) in the testosterone arm (P < 0.05), although a regression analysis adjusted for age and baseline performance did not show a significant effect (β = −1.49, 95% CI: (−2.95, 0.40), P = 0.077). Davis et al. (2014) also reported no significant difference between the testosterone arm and placebo on the Groton Maze (recall). Davis et al. (2014) reported no significant difference between the testosterone treatment arm and placebo on the ‘Groton maze task recall’ (β = −0.11, 95% CI: (−0.84, 0.61), P > 0.77).
Working memory
Davison et al. (2011) assessed working memory using the CogState One-Back Task. They reported no improvement from baseline to week 26 in the testosterone arm, and no difference compared with the normative control group. Regression analysis adjusted for age and baseline performance confirmed no significant effect (β = 0.01, 95% CI: (−0.03, 0.06), P = 0.700). Huang et al. (2015) used Letter-Number Sequencing and reported no significant changes in any testosterone treatment arm compared with placebo, even after adjustment for baseline cognitive function, age and education (all P > 0.05).
Sherwin (1988) was included in this narrative review but reported separately due to substantial methodological and measurement differences. The authors investigated the effects of androgen replacement therapy (as well as oestrogen-only and combined oestrogen−androgen therapy) in surgically menopausal women with a mean age of 45.4 years, using a randomised crossover design. Each treatment phase was separated by a one-month placebo washout period to reduce possible carryover effects. Objective cognitive performance was assessed using measures of short- and long-term memory, including i) digit span subtest as a measure of short-term memory, ii) clerical speed and accuracy as measures of response speed, iii) paragraph recall as a measure of memory and iv) abstract reasoning as a measure of reasoning ability. This paper had limited statistical reporting; however, the overall results indicated improvements in cognitive performance in the testosterone intervention group (and other treatment groups) compared with the placebo group.
Subjective cognition
No studies were identified that measured subjective cognition and met the inclusion criteria.
Testosterone characteristics, age and intervention duration
Among the three studies included in the meta-analysis, testosterone characteristics (route of administration, formulation and dose), age, and intervention duration varied substantially. Variability was also observed in studies described narratively. This, along with the use of small datasets, limited the direct comparability of these studies and prevented further meta-analyses. A narrative approach was adopted to describe characteristics that may contribute to variability between studies in objective cognitive outcomes.
Testosterone characteristics
Delivery method
Delivery methods varied considerably. Two studies used transdermal administration (Davison et al. 2011, Davis et al. 2014), one used oral administration (Kocoska-Maras et al. 2011) and two used intramuscular injections (Sherwin 1988, Huang et al. 2015). These delivery methods are associated with different pharmacokinetics, including how quickly the exogenous hormone enters the bloodstream, reaches peak concentration and thereafter the rate of decay of the plasma hormone level.
Formulation
Multiple formulations were used. Two studies (Davison et al. 2011, Davis et al. 2014) reported using a testosterone preparation but did not specify the formulation, compared with testosterone undecanoate reported by Kocoska-Maras et al. (2011) and testosterone enanthate by Huang et al. (2015) and Sherwin (1988).
Dose
Testosterone doses were standardised to 1 mg/day for comparability across studies, following consultation with an endocrinologist, with the acknowledgement that this does not accurately reflect the overall impact of delivery method, formulation, and dose on the body’s metabolism of each testosterone preparation. It does allow for some descriptive comparison. Across the studies, there was variation, ranging from 0.3 mg/day (Davis et al. 2014), 4.55 mg/day delivered with 0.3 mg/day absorbed (Davison et al. 2011), 25 mg/week (≈3.6 mg/day) (Huang et al. 2015), 40 mg/day (Kocoska-Maras et al. 2011) to 200 mg/month (≈6.7 mg/day) (Sherwin 1988).
Age
Across the five studies, participants’ ages ranged from 41 to 65 years, with an approximate mean age of 57 years, consistent with the reporting of all studies that recruited postmenopausal women (Huang et al. 2015). recruited the youngest cohort (M = 54 years) compared to (Davison et al. 2011, Kocoska-Maras et al. 2011, Davis et al. 2014), who reported slightly older cohorts in their mid-to-late 50s. Sherwin (1988) do not clearly report participant ages. There were no eligible studies in MT that met the inclusion criteria.
Duration
Across these studies, intervention duration ranged from four weeks (Kocoska-Maras et al. 2011) to 24–26 weeks (Davison et al. 2011, Davis et al. 2014, Huang et al. 2015). For Sherwin (1988), a crossover design meant that participants experienced two three-month treatment periods with a washout period in between.
Discussion
Summary of main findings
Across three studies of postmenopausal women spanning three cognitive domains, meta-analyses indicated no statistically significant effects, with effect estimates close to zero. Heterogeneity varied across domains, suggesting inconsistency between study findings. There was substantial variability in testosterone characteristics, intervention duration and participant age. No studies were identified that used a population in MT or assessed subjective cognition. In summary, the current review is most likely underpowered (k = 2–3; n = 244) and is insufficient to conclude that there is no effect, rather than providing evidence of no effect.
Interpretation
Given the inconclusive findings, comparisons with the broader literature are limited. Some research suggests improved spatial ability when low testosterone levels are restored (Miller et al. 2005), whereas other research do not (Dong et al. 2021). Furthermore, some research suggests that higher levels of endogenous testosterone and oestrogen can benefit verbal memory. The literature presents mixed findings on the relationship between testosterone and cognition, and this relationship is rarely reported in midlife women, with findings often relying on extrapolated results.
Evidence from Greendale et al. (2009) suggests that cognitive difficulties during menopause may be time-limited. In their longitudinal cohort, postmenopausal women who were not using hormone therapy showed a return to premenopausal-like cognitive performance levels. Similarly, Maki (2015) reported a return to baseline memory state during the postmenopausal period. This is relevant to the current review, as all included studies sampled were already postmenopausal, a stage at which cognitive performance may have begun to recover, as suggested by Greendale et al. (2009) and (Maki 2015). If cognitive difficulties are restricted to MT, sampling women after this period reduces the chances of detecting treatment-related effects.
Another complementary idea is the ‘critical window hypothesis’ of hormone therapy and cognitive function, which states that the effects of hormone therapy depend on the timing of initiation with respect to age or MT, with optimal effects observed with early initiation (Maki 2013). This hypothesis explores treatment responsiveness rather than natural cognitive trajectories. Although the evidence for this hypothesis centres mainly on oestrogen rather than testosterone treatment, it could indicate that initiating testosterone treatment earlier in menopause may increase the likelihood of influencing cognitive outcomes.
Substantial differences in the number of studies measuring objective versus subjective cognition in this review could be understood within a biopsychosocial framework. This framework proposes that hormones interact with psychological factors (e.g. beliefs and symptom attributions), as well as social and contextual factors, to shape cognitive experiences. As a result, subjective experiences may not correlate with objective performance. From this perspective, objective neuropsychological testing measures may lack sensitivity to measure subjective cognitive complaints. Furthermore, this may indicate a lack of appropriate tools to assess subjective cognition in menopause, which could explain why no studies were identified in our search.
Strengths and limitations
This is an up-to-date and focused systematic review and meta-analysis exploring the effects of testosterone on objective and subjective cognition in women in MT and postmenopause. This study was not designed to examine the mechanisms by which testosterone impacts cognitive function, but aimed to explore trends in the data.
A strength of this review is the inclusion of both surgical and natural menopause, increasing the generalisability of the findings. This inclusion also resulted in methodological diversity, as two of the five studies employed an oestrogen stabilisation period of at least 8 weeks. This reflects real-world clinical variation among women.
Another strength of this review is the use of a screening process to ensure that reliable and valid cognitive measures were included. Predefined decision rules were used to select a single test per cognitive domain for studies that reported multiple tests. This test selection was intended to reduce violations of statistical independence.
Limitations of this review include methodological constraints. Small sample sizes and differences in study designs and comparator groups may affect the validity and comparability of this review. Previous similar reviews were also checked to ensure that relevant trials were not missed. Of the five trials reported in Somboonporn et al. (2005), four did not meet the eligibility criteria for this review because they used active hormonal comparators, and one study (Sherwin 1988) was eligible but could only be analysed narratively because extractable data were unavailable. Of the three cognition trials reported in Islam et al. (2019), two were excluded for the same eligibility reasons, and one study (Davis et al. (2014) was included.
A further limitation is that the objective test selection process (selecting a single test per cognitive domain for studies that reported multiple measures) may have introduced variation in test sensitivity. The selected test’s ability to detect cognitive domain-specific changes may have been influenced both positively and negatively. Furthermore, neuropsychological tests may represent more than one cognitive domain.
Another limitation of this review was the lack of studies exploring subjective cognition. One study was excluded during the screening stages of this review. Glynne et al. (2025) conducted a retrospective cohort study of 510 perimenopausal and postmenopausal women using MHT, assessing self-reported mood and cognitive symptoms. However, without a comparison or control group, the reliability of these findings is limited, and the reported outcomes could be attributable to expectancy or placebo effects.
Variations in testosterone characteristics were also a limiting factor in this review. There were notable differences in the delivery methods and formulations of testosterone. These variations can affect absorption, drug stability and metabolic processes. Therefore, differences in cognitive outcomes may not be fully attributable to testosterone therapy alone, given its pharmacokinetic properties. The duration of testosterone therapy ranged from 4 weeks to 24–26 weeks (approximately 6 months). Testosterone may take 3–6 months to achieve full dose effects (Panay 2022), which raises the question of whether a 4-week treatment duration would be sufficient to detect cognitive changes. No studies explored long-term follow-up, limiting any conclusions about the persistence or decline of effects over time.
The ages of participants across the included studies ranged from 41 to 65 years (mean average ≈ 57 years), spanning MT to postmenopause (Santoro 2016, McCarthy & Raval 2020). Although all participants in this review were confirmed to be postmenopausal, no standardised categorisation was used (e.g., STRAW+10; Harlow et al. 2012), nor was the time since menopause at the start of the intervention recorded. It is also important to consider that any classification may mask the effects of biological ageing. Cognitive outcomes in midlife are thought to be influenced by both chronological ageing (e.g. structural and functional changes in the nervous system (Salthouse 2010)) and hormonal status. Therefore, age may be a confounding variable, limiting the ability to attribute cognitive changes to testosterone therapy.
Finally, there are limitations to comparing surgical menopause (i.e. a sudden and total switch off in hormones) with physiological menopause (i.e. a gradual decline in hormone levels). Furthermore, if a bilateral oophorectomy was performed for reasons of malignancy, other factors, such as medication (e.g. chemotherapy) or psychological trauma, may also impact cognition.
Implications for future research and practice
Based on this review, it remains unclear whether testosterone improves cognition in MT and postmenopausal women, owing to the low statistical power analysis and limited number of included studies. Many factors that require consideration in further research before conclusions regarding the effectiveness of testosterone therapy on cognition can be drawn.
Future trials should include a non-active comparison group and clearly detail the drugs’ characteristics (e.g. dose, formulation, and delivery method) with clear rationales for their selection. This would allow the pharmacokinetic properties of testosterone to be carefully considered. Furthermore, there is a need to explore testosterone therapy in MT. Cognitive complaints appear to peak during the menopause transition (Unkenstein et al. 2016) when hormone fluctuations are greatest. Early discussions from Greendale et al. (2009) and ideas alongside the critical window hypothesis of hormone therapy and cognitive function suggest the importance of exploring testosterone therapy initiation during MT.
Differences in findings between objective and subjective cognition may reflect underlying epistemological approaches. Objective cognition is often assessed within a biomedical framework, whereas subjective cognition aligns more closely with a phenomenological epistemology, emphasising lived experience. An integrative approach combining medicine and psychology may be required to explore the possible effects of testosterone on subjective cognition.
This review reports underpowered and insufficient evidence to prescribe testosterone therapy for cognitive benefit, raising considerations for clinical practice. Although off-label and individualised care can be considered at times in menopause symptom management, women must understand the current lack of robust evidence for testosterone therapy and cognition. This review does highlight the continued need for a biopsychosocial approach to the management of cognitive complaints and integration between medical prescribing and psychological support.
Conclusion
This review concludes that the evidence is underpowered and insufficient, rather than evidence of no effect, to explore the effect of testosterone therapy on cognition for women in MT and postmenopause. Cognitive difficulties during MT and postmenopause can impact quality of life, including reduced self-confidence, difficulties at work, and anxiety. Evidence for other forms of MHT remains mixed, and the role of testosterone for objective or subjective cognitive complaints remains underexplored and further research is required.
Supplementary materials
Declaration of interest
There is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
Author contribution statement
AF, AS and RD participated in the conception and design of the study. AF conducted the searches and data extraction and performed the statistical analyses. SAE conducted a secondary review of screening papers. RD conducted secondary reviews of all data extraction and analyses. NK provided medical supervision during the interpretation of study outcomes and implications. AS and RD provided supervision across all stages of the study. All authors approved the final version of the manuscript.
Acknowledgements
The authors thank Dr Rebecca Webster for her medical guidance and Prof. Vaughan Bell for his neuropsychological guidance.
References
- Andy C, Nerattini M, Jett S, et al. 2024. Systematic review and meta-analysis of the effects of menopause hormone therapy on cognition. Front Endocrinol 15 1350318. ( 10.3389/fendo.2024.1350318) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Australian Menopause Society 2019. Testosterone and women. Retrieved 13/11/2025. (https://menopause.org.au/health-info/resources/testosterone-and-women)
- Barrett-Connor E & Goodman-Gruen D. 1999. Cognitive function and endogenous sex hormones in older women. J Am Geriatr Soc 47 1289–1293. ( 10.1111/j.1532-5415.1999.tb07427.x) [DOI] [PubMed] [Google Scholar]
- Bazeley A, Marren C & Shepherd A. 2022. Menopause and the workplace. (https://www.fawcettsociety.org.uk/Handlers/Download.ashx?IDMF=9672cf45-5f13-4b69-8882-1e5e643ac8a6)
- British Menopause Society 2026. BMS tool for clinicians testosterone replacement in menopause. (https://thebms.org.uk/wp-content/uploads/2026/05/08-NEW-BMS-TfC-Testosterone-replacement-in-menopause-MAY2026-A.pdf) [DOI] [PubMed]
- Bromberger JT, Matthews KA, Schott LL, et al. 2007. Depressive symptoms during the menopausal transition: the study of women’s health across the nation (SWAN). J Affect Disord 103 267–272. ( 10.1016/j.jad.2007.01.034) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bromberger JT, Chang Y, Randolph JF Jr, et al. 2013. Does risk for anxiety increase during the menopausal transition? Study of women’s health across the nation. Menopause 20 488–495. ( 10.1097/gme.0b013e3182730599) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Campbell Collaboration 2023. Effect Size Calculator. Retrieved 18/1/2026. (https://www.campbellcollaboration.org/calculator/d-means-sds) [Google Scholar]
- Critical Appraisal Skills Programme 2023. CASP randomised control trial checklist. Retrieved 09/09/2025. (https://casp-uk.net/casp-tools-checklists/randomised-controlled-trial-rct-checklist/)
- Davis SR & Wahlin-Jacobsen S. 2015. Testosterone in women the clinical significance. Lancet Diabetes Endocrinol 3 980–992. ( 10.1016/S2213-8587(15)00284-3) [DOI] [PubMed] [Google Scholar]
- Davis SR, Jane F, Robinson PJ, et al. 2014. Transdermal testosterone improves verbal learning and memory in postmenopausal women not on oestrogen therapy. Clin Endocrinol 81 621–628. ( 10.1111/cen.12459) [DOI] [PubMed] [Google Scholar]
- Davison SL, Bell RJ, Gavrilescu M, et al. 2011. Testosterone improves verbal learning and memory in postmenopausal women: results from a pilot study. Maturitas 70 307–311. ( 10.1016/j.maturitas.2011.08.006) [DOI] [PubMed] [Google Scholar]
- Dong X, Jiang H, Li S, et al. 2021. Low serum testosterone concentrations are associated with poor cognitive performance in older men but not women [original research]. Front Aging Neurosci 13 712237. ( 10.3389/fnagi.2021.712237) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dratva MA, Banks SJ, Panizzon MS, et al. 2024. Low testosterone levels relate to poorer cognitive function in women in an APOE-ε4-dependant manner. Biol Sex Differ 15 45. ( 10.1186/s13293-024-00620-4) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Glynne S, Kamal A, Kamel AM, et al. 2025. Effect of transdermal testosterone therapy on mood and cognitive symptoms in peri- and postmenopausal women: a pilot study. Arch Wom Ment Health 28 541–550. ( 10.1007/s00737-024-01513-6) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gold EB, Colvin A, Avis N, et al. 2006. Longitudinal analysis of the association between vasomotor symptoms and race/ethnicity across the menopausal transition: study of women’s health across the nation. Am J Public Health 96 1226–1235. ( 10.2105/AJPH.2005.066936) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Greendale GA, Huang MH, Wight RG, et al. 2009. Effects of the menopause transition and hormone use on cognitive performance in midlife women. Neurology 71 1850–1857. ( 10.1212/WNL.0b013e3181a71193) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gribble KD, Bewley S, Bartick MC, et al. 2022. Effective communication about pregnancy, birth, lactation, breastfeeding and newborn care: the importance of sexed language. Front Glob Womens Health 7 818856. ( 10.3389/fgwh.2022.818856) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harlow S D, Gass M, Hall J E, et al. 2012. Executive summary of the Stages of Reproductive Aging Workshop +10: addressing the unfinished agenda of staging reproductive aging. Climacteric 15 105–114. ( 10.3109/13697137.2011.650656) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Higgins J, Thomas J, Chandler J, et al. 2024. Cochrane Handbook for Systematic Reviews of Interventions. (www.cochrane.org/handbook) [Google Scholar]
- Huang G, Wharton W, Travison TG, et al. 2015. Effects of testosterone administration on cognitive function in hysterectomized women with low testosterone levels: a dose–response randomized trial. J Endocrinol Investig 38 455–461. ( 10.1007/s40618-014-0213-3) [DOI] [PMC free article] [PubMed] [Google Scholar]
- International Menopause Society [IMS] 2026. Recommendations and key messages on women’s midlife health and menopause. (https://www.imsociety.org/wp-content/uploads/2026/03/IMS-Recommendations-Full-Document.pdf) [DOI] [PubMed]
- Islam RM, Bell RJ, Green S, et al. 2019. Safety and efficacy of testosterone for women: a systematic review and meta-analysis of randomised controlled trial data. Lancet Diabetes Endocrinol 7 754–766. ( 10.1016/S2213-8587(19)30189-5) [DOI] [PubMed] [Google Scholar]
- Kilpi F, Soares ALG, Fraser A, et al. 2020. Changes in six domains of cognitive function with reproductive and chronological ageing and sex hormones: a longitudinal study in 2411 UK mid-life women. BMC Womens Health 20 177. ( 10.1186/s12905-020-01040-3) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kocoska-Maras L, Zethraeus N, Rådestad AF, et al. 2011. A randomized trial of the effect of testosterone and estrogen on verbal fluency, verbal memory, and spatial ability in healthy postmenopausal women. Fertil Steril 95 152–157. ( 10.1016/j.fertnstert.2010.05.062) [DOI] [PubMed] [Google Scholar]
- Kravitz HM, Ganz PA, Bromberger J, et al. 2003. Sleep difficulty in women at midlife: a community survey of sleep and the menopausal transition. Menopause 10 19–28. (https://journals.lww.com/menopausejournal/fulltext/2003/10010/sleep_difficulty_in_women_at_midlife__a_community.5.aspx) [DOI] [PubMed] [Google Scholar]
- Maki PM & Weber MT. 2021. A research primer for studies of cognitive changes across the menopause transition. Climacteric 24 382–388. ( 10.1080/13697137.2021.1905625) [DOI] [PubMed] [Google Scholar]
- Maki PM 2013. Critical window hypothesis of hormone therapy and cognition: a scientific update on clinical studies. Menopause 20 695–709. ( 10.1097/GME.0b013e3182960cf8) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Maki PM 2015. Verbal memory and menopause. Maturitas 82 288–290. ( 10.1016/j.maturitas.2015.07.023) [DOI] [PubMed] [Google Scholar]
- McCarthy M & Raval AP. 2020. The peri-menopause in a woman’s life: a systemic inflammatory phase that enables later neurodegenerative disease. J Neuroinflammation 17 317. ( 10.1186/s12974-020-01998-9) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miller KK, Grieco KA & Klibanski A. 2005. Testosterone administration in women with anorexia nervosa. J Clin Endocrinol Metab 90 1428–1433. ( 10.1210/jc.2004-1181) [DOI] [PubMed] [Google Scholar]
- National Institute for Health and Care Excellence [NICE] 2024. Menopause: identification and management (NICE guideline NG23). (https://www.nice.org.uk/guidance/ng23) [PubMed]
- Naysmith LF, Ward H, Elliott P, et al. 2026. Cognition and the menopause transition: cross-sectional evidence from a large community cohort. NPJ Womens Health 4 14. ( 10.1038/s44294-026-00132-z) [DOI] [Google Scholar]
- Page MJ, Moher D, Bossuyt PM, et al. 2021. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews. BMJ 372 n160. ( 10.1136/bmj.n160) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Panay N 2022. British menopause society tool for clinicians: testosterone replacement in menopause. Post Reprod Health 28 158–160. ( 10.1177/20533691221104266) [DOI] [PubMed] [Google Scholar]
- Rosario ER, Chang L, Head EH, et al. 2011. Brain levels of sex steroid hormones in men and women during normal aging and in Alzheimer’s disease. Neurobiol Aging 32 604–613. ( 10.1016/j.neurobiolaging.2009.04.008) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Salthouse T 2010. Major Issues in Cognitive Aging, vol 1: Oxford University Press. [Google Scholar]
- Santoro N 2016. Perimenopause: from research to practice. J Womens Health 25 332–339. ( 10.1089/jwh.2015.5556) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schaafsma M, Homewood J & Taylor A. 2010. Subjective cognitive complaints at menopause associated with declines in performance of verbal memory and attentional processes. Climacteric 13 84–98. ( 10.3109/13697130903009187) [DOI] [PubMed] [Google Scholar]
- Schmidt I, Wilson M & Alam M. 2025. The Health and Economic Impacts of Menopause. Policies and Investments to Advance Care, Opportunity, and Equity. F. Analytics. (https://impactsofmenopause.com/wp-content/uploads/sites/12/2025/05/FPA-Bayer-Menopause-PDF.pdf) [Google Scholar]
- Sherwin BB 1988. Estrogen and/or androgen replacement therapy and cognitive functioning in surgically menopausal women. Psychoneuroendocrinology 13 345–357. ( 10.1016/0306-4530(88)90060-1) [DOI] [PubMed] [Google Scholar]
- Somboonporn W, Bell RJ & Davis SR. 2005. Testosterone for peri and postmenopausal women. Cochrane Database Syst Rev 4 CD004509. ( 10.1002/14651858.CD004509.pub2) [DOI] [PubMed] [Google Scholar]
- Strauss E, Sherman EMS & Spreen O. 2006. A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary, 3rd edn: Oxford University Press. [Google Scholar]
- Sullivan M & Woods F. 2001. Midlife women’s attributions about perceived memory changes: observations from the Seattle Midlife Women’s Health Study. J Wom Health Gend Base Med 10 351–362. ( 10.1089/152460901750269670) [DOI] [PubMed] [Google Scholar]
- Talaulikar V 2022. Menopause transition: physiology and symptoms. Best Pract Res Clin Obstet Gynaecol 81 3–7. ( 10.1016/j.bpobgyn.2022.03.003) [DOI] [PubMed] [Google Scholar]
- Tepper PG, Randolph JF Jr, McConnell DS, et al. 2012. Trajectory clustering of estradiol and follicle-stimulating hormone during the menopausal transition among women in the study of women’s health across the nation (SWAN). J Clin Endocrinol Metab 97 2872–2880. ( 10.1210/jc.2012-1422) [DOI] [PMC free article] [PubMed] [Google Scholar]
- Unkenstein AE, Bryant CA, Judd FK, et al. 2016. Understanding women’s experience of memory over the menopausal transition: subjective and objective memory in pre-peri-and postmenopausal women. Menopause 23 1319–1329. ( 10.1097/GME.0000000000000705) [DOI] [PubMed] [Google Scholar]
- Viechtbauer W 2010. Conducting meta-analyses in R with the metafor package. J Stat Software 36 1–48. ( 10.18637/jss.v036.i03) [DOI] [Google Scholar]
- Wåhlin-Jacobsen S, Pedersen AT, Kristensen E, et al. 2015. Is there a correlation between androgens and sexual desire in women? J Sex Med 12 358–373. ( 10.1111/jsm.12774) [DOI] [PubMed] [Google Scholar]
- Wan X, Wang W, Liu J, et al. 2014. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 14 135. ( 10.1186/1471-2288-14-135) [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.

This work is licensed under a