Dementia is characterized by a decline in cognition involving one or more of the following cognitive domains: memory and learning, language, executive function, complex attention, or social cognition (McKhann et al., 2011). The number of people with dementia worldwide was estimated at 47.47 million in 2015 and is expected to reach 135.46 million in 2050. Due to the increasing number of older adults, dementia and age-related cognitive disorders present a significant challenge for health services and an increasing global problem (Ferri et al., 2005; Livingston et al., 2020; Tam et al., 2021).

This trend seems more rapid in South Korea (hereafter Korea). Based on the government-led data, it is estimated that the number of people aged ≥65 years will reach 8.53 million in 2021 and is expected to exceed 12.98 million by 2030 and 19 million (accounting for 39.8% of the older adult population) by 2050 (Jang et al., 2021; Kim J.H. et al., 2022). Additionally, the burden of dementia in Korea is an important issue and is expected to increase considerably, and the annual management cost per patient was approximately 20.72 million won in 2019 (Kong and Park, 2022). The cost of managing dementia at the national level is estimated to be 16.5 trillion won (approximately 0.86% of GDP), which is expected to increase to 63 trillion won by 2040 (Park et al., 2013; Kong and Park, 2022).

Most studies have shown that women have a higher prevalence of dementia than men. This trend, which is generally explained by differences in sex hormones, education, and life expectancy, becomes more evident as age increases, because the prevalence of Alzheimer’s disease (AD) is higher in women aged >80 years than in men (Ferri et al., 2005; McKhann et al., 2011). These demographic characteristics suggest an anticipated increase in the burden of dementia in women. Identifying women at a higher risk of dementia and correcting modifiable risk factors are essential (Derreberry and Holroyd, 2019).

Menopausal hormone therapy (MHT), which is prescribed for postmenopausal women with moderate to severe menopausal vasomotor symptoms, is not recommended to prevent dementia (Lee et al., 2020; Pinkerton, 2020). Although some epidemiological and observational studies on MHT in aging women support the positive effects of estrogen on cognitive function (Imtiaz et al., 2017), data from the Women’s Health Initiative and extensive research have shown that estrogen replacement does not protect against dementia and may increase its risk (Shumaker et al., 2004; McCarrey and Resnick, 2015; Pinkerton, 2020). Tibolone (TIB) treatment, as an alternative to conventional hormone replacement therapy, has estrogenic, progestogenic, and androgenic effects (Huang et al., 2010). Although there are many protective actions for the nervous system, TIB is contraindicated in women with a history of breast cancer and may increase the risk of stroke (Modelska and Cummings, 2002; Formoso et al., 2016). Although the number of participants was small, a recent nationwide case–control study showed that TIB treatment increased the risk of AD in women older than 60 years (Savolainen-Peltonen et al., 2019). Another nationwide study using Korean claims data suggested that TIB treatment did not significantly affect dementia risk aged in patients aged 50–80 years (Han et al., 2021). However, this study did not include possible confounders, such as the duration and initiation age of MHT. Furthermore, it did not adjust for various essential risk factors for dementia and gynecological factors, such as menstrual history, excessive alcohol intake, and physical activity.

Although recent nationwide studies have been conducted on the relationship between MHT and the risk of dementia, findings on this issue are limited. This study aimed to explore the risk of dementia and MHT using the Health Insurance Database of South Korea.

We conducted a large cohort study of menopausal women undergoing MHT to determine the association between MHT use and the risk of dementia. Our study revealed that TIB and oral estrogen alone were associated with an increased risk of AD dementia. In contrast, CEPM, CEPP, and topical estrogen were not associated with an increased risk of AD dementia. In this analysis, the use of TIB, CEPM, and oral estrogen alone was found to be related to an increased risk of non-AD dementia and TD. However, topical estrogen was not associated with an increased risk of AD or non-AD dementia.

A few studies have reported neither the benefit nor the risk of MHT concerning AD, which begins later than 65 or more than 5 years after menopause. A recent study that assessed the risks of developing TD and AD in women exposed to different types of MHT for different durations also reported no increased risks of developing dementia overall (Vinogradova et al., 2021). Furthermore, a retrospective study using a 10-year claims dataset found that MHT was associated with a reduced risk of all neurodegenerative diseases, including TD and AD (Kim et al., 2021). However, most studies have suggested the potential and overall risks of MHT later in life concerning dementia, including AD. The most influential study, the Women’s Health Initiative Memory Study, reported a significantly increased TD risk for using conjugated equine estrogen (CEE) + medroxyprogesterone acetate and CEE alone (Shumaker et al., 2003). Similar results were noted in the Kronos Early Estrogen Prevention Cognitive and Affective Ancillary Study (Gleason et al., 2015). Early versus Late Intervention Trial with Estradiol-Cognitive Endpoints indicated no beneficial or harmful effects of MHT on dementia risk (Henderson, 2011; Henderson et al., 2016). Until recently, prescribing MHT purely for dementia prevention was not recommended in the absence of other clinical indications for MHT, such as vasomotor symptoms (hot flashes and night sweats) and sleep disturbances (Baber et al., 2016; Marjoribanks et al., 2017; North American Menopause Society, 2017). Our study, which adjusted for other dementia risk factors, also showed that hormone therapy increased the overall risk of dementia in elderly menopausal women. Moreover, our results (Supplementary Table S2) are consistent with the content of the current hormone therapy guidelines, explicitly emphasizing the potential benefits of hormone therapy for younger postmenopausal patients and those who commence treatment within 10 years of menopause onset.

The most important finding of this study was that MHT had different effects on the incidence of AD and non-AD dementia. In this study, TIB, CEPM, and estrogen-only use were associated with an increased risk of non-AD dementia, consequently increasing overall TD incidence. Excluding CEPP, most oral hormone therapies (TIB, CEPM, and estrogen) showed an increased tendency towards non-AD dementia occurrence. This is believed to be due to the accumulation of prothrombotic metabolites that may occur during the metabolic process of oral hormone therapy, leading to atherosclerotic effects, ischemia, and vascular cognitive decline (Henderson and Popat, 2011; Scarabin-Carre et al., 2012). Although TIB is known to show a cardiovascular protection effect through its metabolite (Del Rio et al., 2020), long-term use of TIB may also be associated with increased cardiovascular risks, such as CRP elevation and hyperthermia, which could lead to brain vessel wall injury (Formoso et al., 2016; Zhu et al., 2021). A similar mechanism can explain this as the analysis result that topical estrogen is not associated with non-AD dementia occurrence. In other words, topical estrogen, which has a hepatic first-pass effect, is not metabolized in the liver and does not produce various metabolites that can show adverse vascular effects (Simon, 2012; Beck et al., 2017; Oliver-Williams et al., 2019).

Previous studies have reported the neuroprotective effects of estrogen on the brain, cognition, and dementia (Yaffe et al., 2000; Carlson et al., 2001; Lui et al., 2003). Estrogen improves synapse formation in the hippocampus and increases choline acetyltransferase activity in the basal forebrain of rats (Aubele and Kritzer, 2011). Moreover, estrogen improves prefrontal lobe-dependent executive functioning in women (Krug et al., 2006). In addition to its cognitive effects, estrogen has neuroprotective effects, such as inhibiting amyloid-beta depositions in the brain, primarily forming neuropathological lesions in patients with AD (Sohrabji, 2005). However, an increased level of free estradiol may indicate atherothrombogenic stroke risk in older postmenopausal women (Henderson and Popat, 2011). Insulin resistance, dyslipidemia, and inflammation are the potential mediators of this association (Scarabin-Carre et al., 2012). Although the risk of ischemic stroke differs by age, recent studies of MHT on stroke incidence suggest that the risk of ischemic stroke increases with oral MHT use in women aged >60 years (Bushnell et al., 2014). Considering the incidence of all-cause dementia, most non-AD cases appear to be vascular dementia cases (O'Brien and Thomas, 2015). However, the precise mechanism by which MHT affects brain vessels and influences the incidence of ischemic stroke has not been established (Miller et al., 2003; Khalil, 2013). Considering vascular dementia and the increased risk of ischemic stroke due to MHT, our findings indicate that MHT specifically increases the risk of vascular dementia rather than AD dementia. Consequently, the incidence of overall dementia increased among postmenopausal women. Although there are several studies regarding the risk of AD and vascular dementia after using MHT (Savolainen-Peltonen et al., 2019; Vinogradova et al., 2021), the present study’s findings are notable because they considered vascular dementia, which may be anticipated to occur at elevated rates in the presence of increased cerebral thrombosis risk in MHT users. This result is consistent with the finding of an earlier study that showed an overall increase of 20–40% in the risk of ischemic stroke.

The strength of our study was that it was a large-scale Asian population retrospective cohort study with a sufficient number of participants (N = 2,506,272) compared with the current observational study. It included many TIB users, and TIB is widely used in Europe, Asia, and Korea (Huang et al., 2010). Although we conducted a large-scale study of MHT effects on dementia, several issues limit the generalizability of our findings: the patients in our data were only Korean. Therefore, the generalization of the results to other ethnic or racial groups should be made with caution. There may have been detection bias in this study. Moreover, we did not have additional biomarkers indicating AD pathology, such as cerebrospinal fluid biomarkers, genetic testing, molecular imaging or neuropathological data from the participants. Also, although age is regarded to be an important parameter in estimating the risk of AD, our study has a limitation in that it did not reconcile younger skew of age and older age for development of AD. We assumed that non-AD dementia represents overall vascular dementia; however, other etiologies, such as Parkinson’s disease, may occur alongside non-AD dementia. Therefore, it is impossible to confirm the diagnosis code for any dementia, and our analysis may have led to a more inaccurate diagnosis. We assumed that most patients adhered to their physicians’ instructions as this study’s information on drug prescriptions was derived from claims data. However, menopausal women may not have adhered to their MHT prescriptions, which could have potentially introduced biases. It is also important to note that many Koreans prefer phytoestrogens over MHT (Kim Y. et al., 2022). Finally, although we analyzed as much data as possible, more information was needed regarding laboratory measurements, including serum hormone levels and blood pressure, which could be potential confounders.

This investigation showed that MHT had different effects on the incidence of AD and non-AD dementia, suggesting that it specifically increases the risk of non-AD dementia but that transdermal estrogen is not associated with dementia risk. The different effects of MHT should be considered in appropriately selected patients according to individual situations, such as underlying cardiovascular risk status.

The data analyzed in this study was obtained from The National Health Information Database of the National Health Insurance Service in South Korea. The datasets are not readily available because, due to NHIS’s privacy policy, only researchers have access to the data for a limited period of time. Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by Institutional Review Board of the Sanggye Paik Hospital (approval number: SGPAIK-2020-08-002). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

J-SY and JP: conception and design of the study, acquisition of data, and final approval of the manuscript. J-SY, JL, and JP: analysis and interpretation of the data, drafting and revising the manuscript for content. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.