• Increased cortisol tends to be associated with Alzheimer’s disease (AD)-related markers. • Findings depend on cortisol measurement methodology and study population. • Few studies have examined links between novel AD-related biomarkers and cortisol. • Sex differences in these relationships remain underexplored. • Longitudinal research and standardization of cortisol measurement methods are needed. Chronic stress triggers the dysregulation of hypothalamic–pituitary–adrenal (HPA) axis which is characterized by altered diurnal cortisol patterns and increased basal cortisol levels. This dysregulation has been found in Alzheimer’s Disease (AD), but its link with AD-related biomarkers remains underexplored. This systematic review aims to synthesize the current knowledge regarding the association between cortisol and biomarkers related to AD. Keywords and phrases related to cortisol and AD-related biomarkers were used to search the MEDLINE database via PubMed and OVID interfaces up to January 2024. Studies focused on the association between cortisol levels and genetic, neuroimaging, and fluid biomarkers related to AD. Additionally, sex differences in the associations between cortisol and AD-related markers were explored. A total of 2465 studies were identified, 54 of which were included in the systematic review. Most included studies were cross-sectional (91%), and measured blood (51.9%) or salivary (44%) cortisol. Only 6.8% of studies had reported on sex differences. Studies assessing the relationship between diurnal cortisol patterns and novel AD-related biomarkers remain scarce. Elevated cortisol levels tended to be associated with AD-related biomarkers, although associations were dependent on the cortisol measurement strategy and study populations. Findings have been mixed, partly due to differences in cortisol measurement methodologies, including the timing and source of cortisol assessment. More longitudinal studies are needed to determine whether cortisol dysregulation preceded pathological changes related to AD.
Holleman et al. (2026) studied this question.