Hormonal contraception and risk for cognitive impairment or Alzheimer's disease and related dementias in young women: a scoping review of the evidence

Introduction Women are significantly more likely to develop Alzheimer's disease and related dementias (ADRD) than men. Suggestions to explain the sex differences in dementia incidence have included the influence of sex hormones with little attention paid to date as to the effect of hormonal contraception on brain health. The aim of this scoping review is to evaluate the current evidence base for associations between hormonal contraceptive use by women and non-binary people in early adulthood and brain health outcomes. Methods A literature search was conducted using EMBASE, Medline and Google Scholar, using the keywords “hormonal contraception” OR “contraception” OR “contraceptive” AND “Alzheimer*” OR “Brain Health” OR “Dementia”. Results Eleven papers were identified for inclusion in the narrative synthesis. Studies recruited participants from the UK, USA, China, South Korea and Indonesia. Studies included data from women who were post-menopausal with retrospective data collection, with only one study contemporaneously collecting data from participants during the period of hormonal contraceptive use. Studies reported associations between hormonal contraceptive use and a lower risk of ADRD, particularly Alzheimer's disease (AD), better cognition and larger grey matter volume. Some studies reported stronger associations with longer duration of hormonal contraceptive use, however, results were inconsistent. Four studies reported no significant associations between hormonal contraceptive use and measures of brain health, including brain age on MRI scans and risk of AD diagnosis. Discussion Further research is needed on young adults taking hormonal contraceptives, on different types of hormonal contraceptives (other than oral) and to explore intersections between sex, gender, race and ethnicity. Systematic Review Registration https://doi.org/10.17605/OSF.IO/MVX63, identifier: OSF.io: 10.17605/OSF.IO/MVX63


Introduction
Sex and gender have long been recognized as important influencing factors for Alzheimer's disease and related dementias (ADRD).The lifetime risk for AD at age 45 is estimated at 1-in-5 for women and 1-in-10 for men (1).Female sex is associated with faster hippocampal atrophy (2) and greater pathological phosphorylated tau burden, key hallmarks of AD (3)(4)(5).The APOEε4 gene also confers a greater risk of AD in women compared to men (at least in White populations) (6,7).Sex hormones may explain some of the differences in risk for ADRD.Indeed, testosterone levels are a potential modifier of tau that may contribute to lower disease burden in males (3).Life-course evidence suggests pregnancy, adverse pregnancy outcomes, age of menarche, cumulative oestrogen exposure and menopause may all have implications for ADRD risk (8,9).The potential for these biologically driven mechanisms to explain the difference in dementia prevalence by sex justifies the application of a women's health lens to the study of brain health (10).
A number of studies have investigated associations between the use of hormone replacement therapy (HRT) and brain health (9,(11)(12)(13)(14)(15). Comparatively less attention has been paid to the associations with hormonal contraception (HC) use.HCs act to simultaneously reduce endogenous sex hormones whilst supplementing synthetic oestrogen and/or progestin.Understanding the links between HC use and cognitive function is of considerable interest given the widespread and long-term use.Globally, over 60% of reproductive-age women use contraceptives, most of which are hormonal methods (16).The average length of time for HC use is five years, although many individuals stop and restart use across their lifespan (17).In addition to avoiding pregnancy, women use HCs for other reasons such as managing a medical condition, such as endometriosis-related pain and menorrhagia (18,19).A recent study of young transgender individuals assigned-female-at-birth found that 80% were current or previous HC users (20, 21), highlighting the need for studies to be inclusive of this population.
The first use of HCs often occurs in young adulthood (22), a period increasingly acknowledged as a neglected stage in ADRD life-course research (23).Exposure to, and accumulation of, many modifiable risk factors (e.g., alcohol use and mental disorders 24, 25) begins during this life stage.HC use in young adulthood has been linked to changes in functional connectivity, profiled by increases in prefrontal connectivity and decreases in parietal connectivity (26).Studies have also reported changes in areas involved in affective and cognitive processing [e.g., amygdala, hippocampus, and cingulate gyrus (27)].However, most studies were reported to have major methodological limitations regarding internal validity (27).In addition, most studies focused on short-term exposure to HCs in samples with large age ranges (i.e., 18-45 years 28).Thus, there is a need for further investigation of long-term use, especially regarding young women.Behaviourally, effects of hormone use have been reported in cognitive tasks in women (e.g., mental rotation 29).HC use is also correlated with a first diagnosis of depression, which is a known risk factor for ADRD (30).
The aim of this scoping review was to evaluate the current evidence base for associations between HC use by women, nonbinary and transgender people in early adulthood and brain health outcomes.

Study design
A scoping review methodology was adopted to answer the research question, with a need to identify gaps in the current evidence base (31).A multi-stage approach was taken in line with scoping review methodology; define the research question, apply the PCC framework [as per Joanna Briggs Institute recommendations, the PCC (population, concept, context) framework was used to design the parameters of the scoping review (31)]; identify the databases and search terms and run the search; screen the papers; extract the data; synthesise the findings.

Population, concept, context
The concept was exposure to HCs.Method of action and mode of delivery included any contraceptive method classified as hormonal and targeted female reproductive systems.This included oral contraceptive (OC) pills, hormonal intrauterine devices (IUD), implant, injection, vaginal rings, and skin patches.Methods of HC were included if they contained oestrogen and/or progestin.The population was identified as female participants (where sex is reported) or women, nonbinary, and transgender participants (where gender is reported) who had provided data on use of HCs.The context was selected as cross-sectional or cohort studies.Only studies that reported on the direct associations between HC use and one of the outcomes of interest (risk for dementia, cognitive impairment, other brain health outcomes related to neurodegeneration) were included.No studies that reported on indirect associations (i.e., HC use to depression to neurodegeneration) were included.

Databases and search terms
A literature search was conducted using EMBASE, Medline and Google Scholar, using the keywords "hormonal contraception" OR "contraception" OR "contraceptive" AND "Alzheimer*" OR "Brain Health" OR "Dementia".Additionally, the following search parameters were added to identify any papers that additionally considered the role of gender in this topic: "women" OR "female" OR "transgender" OR "nonbinary".Papers were included if they were written in English or Spanish.No limitations were placed on the year of publication.

Eligibility criteria and selection
Articles were selected for inclusion in the scoping review if they reported on associations between HC use and brain health outcomes associated with ADRD or risk for ADRD.Although originally designed to only include studies reporting HC use between the ages of 18 to 39, no studies provided sufficient detail to determine this.As the majority of HC use is known to be during this age period (16), reported HC use is assumed to have been during this life stage in the included papers.A single reviewer (SG) assessed eligibility for inclusion, with 10% of papers cross-checked by a second author (KW), as recommended by Mak and Thomas (2022) (32).

Data extraction
A data extraction tool was created and piloted prior to use.Data extracted included (where provided) paper title, authors, year of publication, number of participants included, sex/gender breakdown, mean age of participants, HC type, duration of HC use, age started/stopped, brain health outcome measure used, and study results.

Narrative synthesis
A narrative synthesis was used to collate aims, methods and results across the included studies (33).The analysis involved synthesizing and summarizing findings for each outcome identified in the literature.Although we had planned to additionally synthesize results by HC type, most studies reported solely on OC use with the remaining studies providing insufficient detail to determine HC type.We reported effect estimates from studies where available (e.g., hazard ratios with 95% confidence intervals).As a final step, we outlined the broader implications for ADRD risk reduction and prevention, as well as suggestions for future studies.The scoping review was pre-registered on OSF.io (34).

Results
A total of 392 papers were identified in the initial search, 381 were not suitable after title and abstract screening with 11 papers included in the narrative synthesis.There was 100% concordance between the lead and secondary reviewer at both screening stages.Studies recruited participants from the UK (35)(36)(37)(38), USA (39, 40), Italy (41), Indonesia (42), Singapore (43) and South Korea (13,44).All studies except one included retrospectively collected data from women who were post-menopausal at the time of study enrolment, with only one study recruiting participants reporting on use during at early adulthood (see Table 1 for further information).No studies reported on the inclusion of participants who were non-binary or transgender, and as such the results reported relate only to papers that reported on "women" or "female participants".The sample sizes ranged from 99 (40) to 4,696,633 participants (44).Seven studies included in the narrative synthesis reported positive associations between HC use and better brain health (13,35,36,39,40,43,44), whereas four studies reported no significant associations between HC contraceptive use and brain health (37, 38, 41, 42).

Associations between hormonal contraception and ADRD risk
Four studies reported associations between HC use and a lower risk of ADRD (35,43,44), particularly AD (13), whilst three studies found no significant associations between risk of AD diagnosis (38,41) and subjective memory complaints (42).
In a study of UK Biobank participants, 81% of women reported using OCs.In this population, OC use was associated with a reduced risk of dementia with no evidence of an association between age of first use of OCs and risk of dementia (35).Subgroup analysis identified this association was only seen in women below 65 years of age at study baseline, suggesting OC use may only confer a benefit until a particular stage of life.A study of women from the Singaporean Chinese population similarly found 81% of participants had previously used OCs, with the majority reporting less than 5 years of use.Those who used OCs for less than 5 years were found to have a reduced risk of dementia compared to those who had never used OC, but interestingly, this association was not seen in those with more than 5 years of use (43).In contrast, a study utilising data from the South Korean NHIS found most women (80.6%) had never used OCs.Despite this, the analysis found OC use was significantly associated with approximately 10% lower risk of dementia compared to those who had never used OCs, with no differences between less than or more than a year's use (44).
A second study using both the South Korean NHIS and the National Cancer Screening Programme investigated subtypes of dementia.This study reported similarly low use of OCs (16.4% documented use) and found similar reductions in risk for allcause dementia.Considering subtypes of dementia there remained a significant association between OC use and decreased risk of AD, however, no significant association with risk for vascular dementia (13).
This specific association between OC use and risk for AD has not been replicated in other studies.In a British cohort study with low OC use rates (26% previous or current users), there was no significant association with the diagnosis of AD (38).This study included a comparatively small sample size with low rates of OC use compared to rates reported in the UK Biobank cohort which limits confidence in interpreting these results.Participants included in this study were recruited between the ages of 70-100 years, therefore many participants included would not have had access to HCs during their early adulthood explaining the low usage rates reported.Another case-control study recruiting   (41).Finally, a study in Indonesia recruiting women aged 25 and above (mean ∼47 years) with significantly higher HC usage (72% reported use) found no significant association with subjective memory complaint cases (42).

Associations between hormonal contraception and cognition
Two studies reported associations between HC use and better performance on cognitive tasks.In a study using the UK Biobank (current HC: 2%, previous HC: 78%), both past and current HC use was associated with significantly higher test scores on tasks measuring processing speed, executive    (36).A second study recruiting participants from the USA, with similar high HC usage rates (87% current or previous users), found significantly higher performance on tasks of visual-spatial ability and speed.There were no significant differences by HC use on tasks of verbal ability, working memory and verbal learning and memory (39).

Associations between hormonal contraception and MRI measures
Only two studies have reported on associations between HCs and MRI measures, with mixed results.A study recruiting participants in the USA (9% current users, 53% past users) reported significant associations between use and larger grey matter volume in the precuneus, fusiform gyrus, superior parietal lobule, angular gyrus and the inferior frontal gyrus of the left hemisphere and in fusiform gyrus of the right hemisphere (40).
A UK Biobank data analysis (86% HC users) found no association between usage and brain age (37).No other studies were identified that reported MRI outcomes and HC use in the context of ADRDs.

Discussion
Of the eleven papers included in this narrative review, the majority investigated associations between HC use and risk for dementia.Studies that were typically larger and with higher rates of OC use reported significant associations with decreased risk for dementia, including AD, however smaller studies including older women with lower rates of previous OC use reported no associations with AD risk.One study recruiting women from young adulthood did not find associations between HCs and subjective memory complaints.Two studies reported significant associations with better performance on cognitive tasks with HC use, and whilst one study reported higher grey matter volume amongst HC users, a second MRI study found no association with brain age.There is clearly a need for more research on this topic, with a particular need to focus on data collection within the age group most likely to be using HCs (young adults), more detailed investigation by HC type and an expansion of outcomes of interest to include more specific research around associations with cognitive performance and brain MRI measures relevant to ADRDs.
Studies conducted to date focus exclusively on associations between OCs and brain health or did not define what was meant by HC within their database used, with no studies explicitly considering other HCs such as the implant, injection or hormonal IUD.Understanding the associations between HCs and brain health is critical, as these methods continue to grow in popularity and may exert more localised effects compared to oral contraception.For example, one hormonal IUD has been associated with increased stress reactivity (45), demonstrating the potential for side effects of the contraception outside of the localised effects, and may raise important implications for ADRD given known associations with cortisol (46).
Another potential limitation that must be addressed is the reliance on retrospective data collection, which may be less accurate due to self-reporting events in the past that may not be recalled accurately, leading to potential recall bias (47).Future studies should collect data from participants in early adulthood to understand whether the potential benefits for brain health and ADRD risk reduction can be seen throughout the lifespan.There are also inherent limitations in all observational studies in establishing a direct cause-effect relationship which should be acknowledged.It is also important to consider confounding by indication, such as psychological factors, sexual debut or abstinence and personality that lead to decision to use HCs which may themselves be determined by brain function or structure (27).In addition, the menopausal status of participants included in the individual studies may have influenced the results independent of previous HC use, given known associations between menopause and brain health (48) (including importantly the potentially reversible nature of brain fog experienced in perimenopause identified in the SWAN studies that may have otherwise influenced cognitive performance in some of these studies 49).Understanding whether exposure to exogenous and synthetic hormones (both as contraceptives and HRT) can modify the risk for future dementia will be important, as evidenced by a recent paper from the UK Biobank which reported significant associations between more prolonged exposure to endogenous hormones and smaller burden of cerebral small vessel disease independent of HC and HRT use (50).Conversely, a study in Sweden found that a longer reproductive period was associated with a higher risk for future dementia, again independent of HC use (51).Slight differences in reproductive periods between the two studies discussed here (37 vs. 34 years on average) as well as different outcomes (small vessel disease on MRI vs. dementia diagnosis) may explain these findings, and highlights the need to develop a more comprehensive literature base in this area to understand the role of HCs, HRT, menopause and lifetime exposure to endogenous sex hormones in relation to brain health and risk for neurodegeneration.Future research could also focus on exploring associations between HC use and more novel and sensitive markers of neurodegeneration, including the use of amyloid and tau positron emission tomography (PET) scans, cerebrospinal fluid (CSF) analysis and blood-based biomarkers (52, 53).As previous research has focused on global cognitive assessments, or the diagnosis of dementia, this would provide evidence on any associations between HCs and the earliest stages of AD in particular, which could help to delineate whether there is an optimal time window where exogenous hormones may confer brain health benefits.
All studies focused exclusively on the dichotomy of sex and did not consider whether gender is relevant to this discussion.Even where studies report on "sex", there are often no documented

TABLE 1 Table
of papers included in narrative synthesis.

TABLE 1 Continued
, and visual and working memory.Longer duration of HC use was also associated with better performance on most cognitive tests, whilst an older age of starting HC use was associated with lower performance across all cognitive domains functioning