Sex-specific associations between astrocytic reactivity and cognitive decline in unimpaired elderly

Chelsea Reichert Plaska; Tovia Jacobs; Davide Bruno; Sang Han Lee; Bruno P Imbimbo; Ricardo S Osorio; Andréa L Benedet; Nicholas J Ashton; Henrik Zetterberg; Nunzio Pomara

doi:10.64898/2026.01.30.702902

This is a preprint.

It has not yet been peer reviewed by a journal.

The National Library of Medicine is running a pilot to include preprints that result from research funded by NIH in PMC and PubMed.

[Preprint]. 2026 Jul 1:2026.01.30.702902. [Version 2] doi: 10.64898/2026.01.30.702902

Sex-specific associations between astrocytic reactivity and cognitive decline in unimpaired elderly

Chelsea Reichert Plaska ^a,^b, Tovia Jacobs ^c, Davide Bruno ^d,^e, Sang Han Lee ^f,^g, Bruno P Imbimbo ^h, Ricardo S Osorio ^i,^j, Andréa L Benedet ^k, Nicholas J Ashton ^l,^m,ⁿ, Henrik Zetterberg ^o,^p,^q,^r,^s,^t,^u, Nunzio Pomara ^v,^x, for the Alzheimer’s Disease Neuroimaging Initiative

^aGeriatric Psychiatry Division, Nathan S. Kline Institute for Psychiatric Research, 140 Old Orangeburg Road, Orangeburg, NY 10962, USA

^bDepartment of Psychiatry, NYU Grossman School of Medicine, 145 E 32^nd Street, New York, NY 10016, USA

^cPenn State College of Medicine, 700 HMC Cres Rd, Hershey, PA 17033, USA

^dSchool of Psychology, Liverpool John Moores University, Tom Reilly building, L3 3AF, Liverpool, UK

^eDepartment of Communication Sciences and Disorders, University of Wisconsin, 1308 W. Dayton St., Madison, WI 53715, USA

^fCenter for Dementia Research, Nathan S. Kline Institute for Psychiatric Research, 140 Old Orangeburg Road, Orangeburg, NY 10962, USA

^gDepartment of Psychiatry, NYU Grossman School of Medicine, 145 E 32^nd Street, New York, NY 10016, USA

^hDepartment of Research & Development, Chiesi Farmaceutici, Largo Francesco Belloli 11/a, 43122 Parma, Italy

ⁱClinical Research Department, Nathan S. Kline Institute for Psychiatric Research, 140 Old Orangeburg Road, Orangeburg, NY 10962

^jDepartment of Psychiatry, NYU Grossman School of Medicine, 145 E 32^nd St\reet, New York, NY 10016, USA

^kDepartment of Psychiatry and Neurochemistry, Institute of Neuroscience & Physiology, the Sahlgrenska Academy at the University of Gothenburg, SE- Wallinsgatan 6, 43141 Mölndal, Sweden.

^lDepartment of Psychiatry and Neurochemistry, Institute of Neuroscience & Physiology, the Sahlgrenska Academy at the University of Gothenburg, SE- Wallinsgatan 6, 43141 Mölndal, Sweden.

^mBanner Alzheimer’s Institute and University of Arizona, 901 E. Willetta Street, Phoenix, AZ, 85718 USA

ⁿBanner Sun Health Research Institute, 10515 W. Santa Fe Drive, Sun City, AZ 85351, USA

^oDepartment of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, the Sahlgrenska Academy at the University of Gothenburg, SE- Wallinsgatan 6, 43141 Mölndal, Sweden

^pClinical Neurochemistry Laboratory, Sahlgrenska University Hospital, Klin Neurokemi Lab Hus V3, SE-43180Mölndal, Sweden

^qDepartment of Pathology and Laboratory Medicine, University of Wisconsin School of Medicine and Public Health, University of Wisconsin-Madison; 53792-2420 Madison, WI, USA

^rWisconsin Alzheimer’s Disease Research Center, University of Wisconsin School of Medicine and Public Health, University of Wisconsin-Madison, 600 Highland Ave J5/1 Mezzanine, Madison, WI, 53792 USA

^sDepartment of Neurodegenerative Disease, UCL Institute of Neurology, Queen Square, London, WC1N 3BG, UK

^tUK Dementia Research Institute at UCL, London, UK

^uCentre for Brain Research, Indian Institute of Science, .V. Raman Avenue, Bangalore, 560 012, India

^vGeriatric Psychiatry Division, Nathan S. Kline Institute for Psychiatric Research, 140 Old Orangeburg Road, Orangeburg, NY 10962, USA

^xDepartment of Pathology, NYU Grossman School of Medicine, 145 East 32^nd Street, New York, 10016 NY, USA

Data used in preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List

^✉

Corresponding Author: Nunzio Pomara, MD, nunzio.pomara@nki.rfmh.org

PMCID: PMC13345015 PMID: 42427583

Abstract

INTRODUCTION:

Astrocyte reactivity may contribute to increased Alzheimer’s disease (AD) risk in women. Although higher plasma glial fibrillary acidic protein (GFAP) levels have been reported in women, few studies have investigated sex-specific relationships with other plasma AD biomarkers and cognition.

METHODS:

Participants were enrolled in the Memory Education and Research Initiative, a longitudinal community-based cohort, and underwent evaluation including blood biomarker sampling. The Alzheimer’s Disease Neuroimaging Initiative (ADNI) was included as a replication cohort.

RESULTS:

Unimpaired women exhibited higher plasma GFAP levels than men (N=584). Higher baseline GFAP was associated with worse longitudinal episodic memory, exclusively in women. In addition, both higher baseline and increasing GFAP levels over time were associated with lower longitudinal plasma Aβ42/40 ratio. The key results were replicated in the ADNI cohort.

DISCUSSION:

These findings suggest that astrocytic reactivity is more pronounced in women and may contribute to sex-specific vulnerability to AD-related pathology and cognitive decline.

Keywords: GFAP, Amyloid-β, Astrocytes, Plasma Biomarkers, Global Cognition, Episodic Memory, Semantic Memory, AD risk

1. INTRODUCTION

Women account for two-thirds of all Alzheimer’s disease (AD) diagnoses¹. Longer life expectancy contributes to AD prevalence^2,3, but does not fully explain it, pointing to additional biological vulnerabilities. Additional risk factors include menopausal estrogen decline⁴, higher prevalence of affective disorders⁵, and a greater burden of inflammatory conditions⁶. Yet, despite increasing attention to sex differences in AD, the impact of sex on neuroinflammatory processes and their relationship to AD pathology remains poorly understood^7,8.

Astrocytes are one of the key mediators of neuroinflammation^9–17. In response to AD pathology, astrocytes become reactive, with increased expression of cytoskeletal glial fibrillary acidic protein (GFAP)^18,19. Several lines of evidence, including post-mortem studies, have established elevated GFAP level in mild cognitive impairment (MCI) and AD, and its associations with greater AD pathology and cognitive decline^20–22. Elevated GFAP levels have also been associated with increased future AD risk²³ and predict progression to AD up to 15 years before clinical diagnosis²⁴.

Astrocytes are essential for supporting cognitive processes^25,26. In cognitively healthy older adults, elevated GFAP levels are associated with poorer cognitive performance, particularly lower episodic memory scores^25,27. Higher GFAP levels are also linked to worse performance on executive function tasks^28,29 and reduced language abilities^27,29. However, findings are not consistent; in at least one study comparing AD, other dementias and CU, GFAP showed no relationship with cognition³⁰. Collectively, these findings highlight astrocytes as key contributors to cognitive function and suggest that increased astrocyte reactivity associated with early AD may contribute to cognitive decline and thus warranting further investigation.

Most studies, including those discussed above, have overlooked the impact of sex. Moreover, many of these studies have statistically adjusted for sex as a covariate, thereby minimizing or removing its influence, instead of examining it as a potentially meaningful biological variable^3,31. However, recent reports suggest potential sex differences in glial cell immune response highlighting the importance of considering this factor when evaluating AD risk^5,7. Recently, increased astrocyte reactivity, as measured by plasma GFAP, was found in CU females at risk for future AD as compared with males³². Increased astrocyte reactivity has also been reported in females in several community-based and large clinic cohorts that included a range of CU, MCI, and dementia participants^33–35, and in a much larger population-based sample consisting of CU and MCI/dementia individuals³⁶. However, a few studies have found no sex-related differences in GFAP levels^37–39, or reported stronger associations between Aβ and p-Tau in astrocyte-reactive males compared with females⁴⁰. None of the above studies that reported sex differences in GFAP examined how the elevations in GFAP are related to cognition, especially in individuals who are CU or without significant AD pathology.

Thus, in the present study we investigated sex differences in plasma GFAP and longitudinal associations with AD biomarkers in cognitively unimpaired older adults enrolled in the MERI program. We also replicated our analyses using a replication cohort (i.e. ADNI). Additionally, we assessed whether GFAP was associated with longitudinal global and domain-specific (e.g., episodic memory) cognitive decline. We conducted a hypothesis-driven examination of CU based on the prior biomarker studies that established that increased astrocytic reactivity may be in response to early AD pathology (Aβ+)⁴¹, including those who are Aβ−¹⁸, and may contribute to more rapid tau accumulation ⁴⁰. We hypothesized that baseline plasma GFAP levels would be higher in women than in men and that higher GFAP would be associated with greater amyloid burden (reflected by lower plasma Aβ42/40) and increased tau pathology (reflected by elevated plasma p-Tau₂₃₁). Furthermore, that higher baseline GFAP and/or steeper longitudinal GFAP increases would predict greater declines in global cognition and domain-specific performance.

2. METHODS

2.1. Population

2.1.1. MERI Cohort

Participants were recruited through the Memory Education and Research Initiative (MERI), an ongoing, longitudinal observational cohort. The MERI program is open to adults aged 18 years and older from Rockland County, New York, and surrounding areas (e.g., Westchester County, NY and Bergen County, NJ), but primarily targets older adults aged 50 and above. Participants are often referred by local physicians for assessment of cognitive impairment, while others are self-referred due to subjective cognitive concerns or a family history of AD or dementia. The MERI cohort includes more than 1,300 individuals. The present analysis focused on participants who provided blood samples during at least one MERI visit, and were classified as CU, using a broad cutoff of Mini-Mental State Examination (MMSE) ≥24⁴². Longitudinal analyses required at least one follow-up visit, the average number of follow-up visits was 3 (range 2–10). For certain analyses, participants with missing data were excluded as detailed below (Figure 1). Most individuals enrolled in the MERI program were cognitively unimpaired (CU), healthy adults with stable medical conditions, as determined during the medical history.

Figure 1. — The total number of MERI participants considered for this analysis is indicated in the top box. Each subsequent box describes the number of MERI participants, out of the total, which had plasma biomarkers, *APOE* genotype and MMSE ≥ 24. The number of MERI participants included in each of the cross-sectional analyses is described in the bottom boxes for the BBM analysis, global cognition analysis, and the cognitive domain analysis.

2.1.2. ADNI Cohort

Data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). For up-to-date information, see www.adni-info.org.

The ADNI recruited individuals between the ages of 55 to 90 years old across several waves [ADNI1, ADNI2, ADNI3, ADNI4]; the full inclusion/exclusion criteria can be found at ADNI website. The present analysis focused on ADNI participants who were from all waves who had APOE genotype, the biomarkers of interest, and were deemed CU based on the previously described criteria⁴³. In brief, participants were included if the MMSE ≥ 24 and clinical dementia rating global score of 0 (Supplemental Figure 1). This ADNI cohort was included as a replication cohort.

2.2. Study Design

The MERI is a longitudinal study described in detail elsewhere⁴⁴. In brief, participants complete the MERI program annually. Each visit consists of three components: a medical evaluation, a neuropsychological assessment, and a psychiatric evaluation. The medical evaluation includes a detailed history of past and current medical conditions and an optional blood draw. The neuropsychological assessment comprises 13 standardized tests covering several cognitive domains, including general intelligence and episodic verbal memory. For the present analysis, we selected tests sensitive to early and progressive changes along the AD spectrum⁴⁵. The cognitive domains included in this analysis were episodic memory (Auditory Verbal Learning Test [AVLT] Total Learning and AVLT Delayed Recall), attention (Trail Making Test A time), executive function (Trail Making Test B time and Verbal Fluency), and semantic knowledge (Category Fluency). The psychiatric evaluation consists of a clinician-conducted interview and the assessment of psychiatric disorders, including depression and anxiety.

2.3. Blood Biomarkers

2.3.1. MERI Blood Biomarkers

Blood samples are collected in standard 10 mL EDTA tubes (BD Vacutainer) and processed as previously described⁴⁶. In brief, blood was collected, kept at room temperature, and centrifuged within one hour of collection. The resulting plasma was transferred to pure polypropylene tubes and stored at −80 °F. The current study included MERI participants whose plasma samples had not been thawed since initial collection and storage. The first visit with an available blood-based biomarker (BBMs) sample, regardless of visit number, was designated as the baseline for subsequent AD biomarker analyses. Plasma concentrations of Aβ40, Aβ42, GFAP and NfL were measured using the Neuro-4-Plex-E Single molecule array (Simoa) assay (Quanterix). Plasma p-Tau231 concentration was measured using an in-house Simoa assay, as previously described⁴⁷.

2.3.1. ADNI Blood Biomarkers

Blood samples were collected during the ADNI study visits, placed on dry ice, and were shipped for analysis on the day a sample was collected. Methodology is described on the ADNI website (http://www.adni-info.org/). Plasma concentrations of Aβ40, Aβ42, p-tau217, GFAP and NfL were measured using the Fujirebio Lumipulse G1200 and Quanterix HD-X⁴⁸.

2.4. Statistical Analysis

All statistical analyses were performed in R version 4.3.3, with custom code. In the MERI cohort, GFAP, NfL, and p-Tau231 values were log10-transformed to reduce skewness and kurtosis, whereas Aβ values were not transformed because the Aβ42/40 ratio was normally distributed and showed no extreme outliers. In the ADNI cohort, all BBMs were log10-transformed to reduce skewness and kurtosis. Cognitive performance was assessed both globally (total MMSE score) and by cognitive domain (episodic memory, attention, executive function, and semantic knowledge). APOE genotype was categorized as E2 carriers (one or two E2 alleles), E3 homozygotes (two E3 alleles), or E4 copies (one or two E4 alleles). Baseline characteristics and group differences were compared using the table one package in R. For continuous variables, table one applies the parametric one way.test function under the assumption of equal variances, which is equivalent to an ANOVA for multigroup comparisons (e.g., genotypes) and a t-test for two-group comparisons (e.g., sex). For categorical variables, tableone uses Pearson’s chi-squared test with continuity correction. Crosssectional analyses were conducted at baseline using standard linear regression models (LMs), while longitudinal analyses were performed using linear mixed regression models (LMMs) implemented with the lmerTest package. Depending on the model, the primary exposure was defined either as a baseline measure (e.g., baseline GFAP) or as a time-varying longitudinal measure (e.g., longitudinal GFAP). LMMs included time as an independent variable to account for variability in follow-up duration across participants; consequently, baseline age was included as a covariate in place of time-varying age. Random intercepts were included in all LMMs to account for within-subject correlations across repeated measures. All cross-sectional and longitudinal models were adjusted for age (or baseline age and time since baseline for LMMs), sex, education, APOE4 status, and race (non-Hispanic white vs other). To assess whether associations between GFAP and other BBMs or cognitive outcomes differed by sex, all models were additionally run within sex-stratified subgroups, with complementary analyses including sex interaction terms. Lastly, analyses were performed to evaluate associations between within-individual rates of change (Δ per unit time) in GFAP and other BBMs, with models otherwise specified as above but adjusted for concurrent age and time between visits. Figure 1 was created using the DiagrammeR package, and all other figures were generated using the ggplot2 package. All biomarkers were visually inspected for outliers and assessed normality.

3. RESULTS

3.1. Baseline Characteristics

3.1.1. MERI Cohort

Our sample consisted of 584 cognitively unimpaired older adults enrolled in MERI. Baseline participant characteristics are presented in Table 1. Although females were more numerous than males, the two groups did not differ in demographic variables, including the proportion of APOE4 carriers. We first examined sex differences in BBMs levels (Figure 2 a-f). Females showed significantly lower plasma concentrations of Aβ40 (p ≤ 0.001), Aβ42 (p = 0.012), and p-Tau231 (p = 0.002), and significantly higher plasma GFAP levels (p ≤ 0.001) compared with males. No significant sex differences were observed in baseline plasma Aβ42/40 ratio or NfL levels. Supplemental Table 1 presents performance on the MERI cognitive tests, including global cognition measured by the MMSE and domain-specific assessments. At baseline, females performed better than males across all cognitive tests, except for the MMSE and the verbal fluency test.

Table 1.

Cohort characteristics at baseline are presented for the whole sample and stratified by sex (female, male) for the MERI cohort and ADNI cohort. Group comparisons by sex were conducted using independent-samples t-tests.

Demographics
	MERI Cohort				ADNI Cohort
	Whole Cohort	Female	Male	p	Whole Cohort	Female	Male	p
N	584	337	247		845 ^†	506	339
Age (mean (SD))	71.93 (9.59)	71.33 (9.76)	72.76 (9.30)	0.075	71.94 (8.17)	70.66 (8.26)	73.86 (7.65)	<0.001^*
Race (%)				0.684
Black or African-American	25 (4.3)	16 (4.7)	9 (3.6)		156 (18.5)	117 (23.1)	39 (11.5)
Caucasian	514 (88.0)	297 (88.1)	217 (87.9)		598 (70.8)	338 (66.8)	260 (76.7)
Asian					51 (6.0)	26 (5.1)	25 (7.4)
Other	45 (7.7)	24 (7.1)	21 (8.5)		15 (1.8)	10 (2.0)	5 (1.5)
More than one race					25 (3.0)	15 (3.0)	10 (2.9)
Education (years) mean (SD))	15.83 (2.87)	15.66 (2.74)	16.04 (3.03)	0.114	16.59 (2.47)	16.26 (2.44)	17.09 (2.42)	<0.001^*
MMSE (mean (SD)) ^~	28.53 (1.69)	28.56 (1.72)	28.50 (1.66)	0.657	28.93 (1.31)	29.10 (1.13)	28.69 (1.51)	<0.001^*
Plasma Biomarkers
Aβ40 (pg/mL) (mean (SD)) ^**	101 (38.3)	95.9 (31.7)	108 (45.1)	<0.001^*	2.47 (0.11)	2.47 (0.09)	2.45 (0.14)	0.014^*
Aβ42 (pg/mL) (mean (SD)) ^**	6.14 (2.20)	5.95 (1.87)	6.41 (2.55)	0.012^*	1.43 (0.12)	1.44 (0.10)	1.41 (0.14)	<0.001^*
Aβ42/40 Ratio (mean (SD)) ^**	0.0607 (0.0121)	0.0613 (0.0116)	0.0599 (0.0126)	0.177	3.74e-02 (9.13e-03)	3.77e02 (6.82e-03)	3.7e-02 (1.18e-02)	0.280
p-Tau231 (pg/mL) (mean (SD)) ^**	0.946 (0.221)	0.922 (0.226)	0.979 (0.209)	0.002^*
p-Tau217 (pg/mL) (mean (SD)) ^**					6.63e-02 (5.22e-02)	6.33e-02 (4.93e-02)	7.09e-02 (5.6e-02)	0.038^*
NfL (pg/mL) (mean (SD)) ^**	1.40 (0.255)	1.38 (0.252)	1.42 (0.260)	0.074	1.24 (0.20)	1.22 (0.20)	1.26 (0.19)	0.002^*
GFAP (pg/mL) (mean (SD)) ^**	2.11 (0.249)	2.15 (0.245)	2.06 (0.247)	<0.001^*	2.13 (0.23)	2.15 (0.23)	2.11 (0.22)	0.027^*
	APOE Genotype n (%)			0.546	APOE Genotype n (%)			0.818
E2/E2	17 (2.9)	11 (3.3)	6 (2.4)		3 (0.4)	1 (0.2)	2 (0.6)
E2/E3	40 (6.8)	28 (8.3)	12 (4.9)		86 (10.2)	48 (9.5)	38 (11.2)
E2/E4	10 (1.7)	5 (1.5)	5 (2.0)		15 (1.8)	8 (1.6)	7 (2.1)
E3/E3	329 (56.3)	188 (55.8)	141 (57.1)		459 (54.3)	275 (54.3)	184 (54.3)
E3/E4	161 (27.6)	88 (26.1)	73 (29.6)		245 (29.0)	152 (30.0)	93 (27.4)
E4/E4	27 (4.6)	17 (5.0)	10 (4.0)		37 (4.4)	22 (4.3)	15 (4.4)

Open in a new tab

Statistical significance was set at p < 0.05 and is indicated by an asterisk (*).

^†

The total N for the ADNI may vary due to missing data (e.g., MMSE total N = 840).

MMSE was only completed as part of the MERI visit that corresponded to biomarker collection.

^**

For the MERI cohort, log10 transformation was only applied to p-Tau231, NFL and GFAP. For the ADNI cohort, log10 transformation was applied to all plasma biomarkers.

Figure 2. — Boxplots of Baseline MERI Cohort. Comparisons were plotted by sex (female: red, male: blue): Aβ40 (A), Aβ42 (B), Aβ42/40 (C), GFAP (D), p-Tau231 (E), and NfL (F). Significant associations are indicated with an asterisks (p<0.05*, p<0.01**, p<0.001***) and p > 0.05 indicated with ns.

3.1.2. ADNI Cohort

The ADNI database was used as replication cohort. Baseline participant characteristics for the ADNI cohort are presented in Table 1. Similar to the MERI cohort, females were more numerous than males. Males were significantly older than females, while females had a higher average level of education than males. The two groups did not differ in the proportion of APOE4 carriers. We first examined sex differences in BBMs levels (Supplemental Figure 2 a-f). Females showed significantly higher plasma concentrations of Aβ40 (p = 0.014), Aβ42 p ≤ 0.001), p-Tau₂₁₇ (p = 0.038), GFAP (p = 0.027), and NfL levels (p = 0.002) compared with males. No significant sex differences were observed in the baseline plasma Aβ42/40 ratio. Of note, while the direction of sex differences for Aβ40, Aβ42, and plasma p-Tau differed between the two cohorts (women lower in MERI, women higher in ADNI), the elevation of plasma GFAP in women and the absence of sex differences in the plasma Aβ42/40 ratio were consistent across both cohorts.

3.2. Associations Between Baseline GFAP and BBMS

3.2.1. Baseline associations – MERI Cohort

We first performed standard linear regression analyses to examine the associations between baseline GFAP levels and baseline BBMs (Supplemental Figure 3a-b). In the whole cohort (Supplemental Table 2), baseline GFAP was positively associated with Aβ40 (p < 0.001), Aβ42 (p = 0.002), p-Tau231 (p < 0.001), and NfL (p < 0.001), and negatively associated with the Aβ42/40 ratio (p<0.001). When analyses were stratified by sex, all significant associations were maintained, except that in males the association with Aβ42/40 was no longer significant (Supplemental Table 3). In females, significant associations remained, including the negative association with the Aβ42/40 ratio, but interestingly, Aβ42 was no longer significant (Supplemental Table 4).

3.2.2. Baseline associations – ADNI Cohort

We first performed standard linear regression analyses to examine the associations between baseline GFAP levels and baseline BBMs. In the whole cohort (Supplemental Table 5) and females only (Supplemental Table 6), baseline GFAP was positively associated with Aβ40 (Whole cohort: p = 0.008; Females: p = 0.029), p-Tau₂₁₇ (p < 0.001), and NfL (p < 0.001). When analyses were stratified by sex, all significant associations were maintained, except that in males the association with Aβ40 was no longer significant (Supplemental Table 7).

3.2.3. Longitudinal outcomes – MERI Cohort

We next performed linear mixed-effects regression analyses to examine the associations between baseline GFAP levels and longitudinal changes in BBMs. On average, participants completed between three and four follow-up visits. In the whole cohort (Table 2a), which included both females and males, baseline GFAP levels were significantly associated with longitudinal increases in Aβ40 (p = 0.005), p-Tau₂₃₁ (p < 0.001), and NfL (p < 0.001). Additionally, baseline GFAP was associated with longitudinal decreases in Aβ42/40 ratio (Figure 3; p = 0.022), When analyses were stratified by sex, only the association with NfL remained significant in males (Table 2b), whereas in females all significant associations between baseline GFAP levels and longitudinal BBMs were maintained (Table 2c). Finally, there were no significant sex × GFAP interactions on any BBMs (Supplemental Table 8).

Table 2.

Associations between baseline GFAP and longitudinal BBMs in the whole MERI cohort (a), and stratified by males (b) and females (c). Results are derived from linear mixed effects models adjusted for baseline age, time since baseline, race, education, and APOE4 status. In the whole-cohort analysis, sex was additionally included as a covariate.

Whole MERI Cohort (N=210 participants across 605 total visits)

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

Aβ40	22.280 (7.819)	0.005^**	time since baseline, age at baseline
Aβ42	0.415 (0.520)	0.425	time since baseline, age at baseline, education, APOE4
Aβ42/40	-0.008 (0.003)	0.022^*	APOE4
p-Tau231	0.168 (0.049)	<0.001^***	time since baseline, age at baseline, APOE4, sex
NfL	0.357 (0.056)	<0.001^***	time since baseline, age at baseline, race

Males Only (N=86 participants across 256 total visits)

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

Aβ40	14.263 (12.703)	0.265	age at baseline
Aβ42	0.764 (0.825)	0.358	No significant covariates
Aβ42/40	−0.000 (0.005)	0.997	age at baseline, APOE4
p-Tau231	0.105 (0.071)	0.142	time since baseline, age at baseline
NfL	0.360 (0.087)	<0.001^***	time since baseline, age at baseline

Females Only (N=124 participants across 349 total visits)

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

Aβ40	30.102 (9.716)	0.002^**	time since baseline
Aβ42	−0.097 (0.686)	0.887	time since baseline, age at baseline, APOE4
Aβ42/40	−0.016 (0.005)	0.001^***	APOE4
p-Tau231	0.237 (0.069)	<0.001^***	time since baseline, APOE4
NfL	0.336 (0.077)	<0.001^***	time since baseline, age at baseline

Open in a new tab

Significant associations are indicated with an asterisks (p<0.05*, p<0.01**, p<0.001***).

Figure 3. — Unadjusted linear trendlines depicting longitudinal plasma Aβ42/40 ratio obtained throughout follow up visits, comparing participants who were in the lower GFAP median (solid line) vs. upper GFAP median (dotted line) at baseline. Plots A-C are the MERI Cohort. Comparisons were plotted in the full cohort (A), male participants only (B), and female participants only (C). Plots were constructed from the subset of 210 participants with 2+ visits where plasma Aβ42/40 values were measured. Y-axis is plasma Aβ42/40 ratio and X-axis is time (years). Plots D-F are the ADNI Cohort. Comparisons were plotted in the full ADNI cohort (A), male participants only (B), and female participants only (C). Plots were constructed from the subset of 294 participants with 2+ visits where plasma Aβ42/40 values were measured.

3.2.4. Longitudinal outcomes – ADNI Cohort

We next performed linear mixed-effects regression analyses to examine the associations between baseline GFAP levels and longitudinal changes in BBMs. On average, participants completed between two follow-up visits. In the whole cohort (Supplemental Table 9a), which included both females and males, baseline GFAP levels were significantly associated with longitudinal increases in Aβ40 (p = 0.045), p-Tau₂₁₇ (p < 0.001), and NfL (p < 0.001). When analyses were stratified by sex, both the association with p-Tau₂₁₇ and NfL remained significant in males (Supplemental Table 9b). Whereas in females all significant associations between baseline GFAP levels and longitudinal BBMs were maintained (Supplemental Table 9c) and Aβ_42/40 was significantly, negatively associated with GFAP (p = 0.025). Finally, there were no significant sex × GFAP interactions on any BBMs (Supplemental Table 10).

3.3. Associations Between Baseline GFAP and Cognition

3.3.1. Baseline associations – MERI Cohort

We first performed standard linear regression analyses to examine the associations between baseline GFAP levels and baseline cognitive performance (Supplemental Figure 3a-b). In the whole cohort (Supplemental Table 2), baseline GFAP was negatively associated with global cognition (p < 0.001) and several domain-specific tests, including AVLT Total Learning (p = 0.002), AVLT Delayed Recall (p < 0.001), verbal fluency (letters) (p = 0.046) and verbal fluency (category) (p < 0.001). Conversely, baseline GFAP was positively associated with attention and executive function measures, including the Trail Making A Test (p = 0.013) and Trail Making B Test (p = 0.038). In males (Supplemental Table 3), only MMSE (p = 0.049) and AVLT Total Learning (p = 0.026) were significantly and negatively associated with GFAP, indicating that higher GFAP levels were linked to lower MMSE scores and poorer Total Learning performance on the AVLT. In females (Supplemental Table 4), baseline GFAP was negatively associated with global cognition (p < 0.001), AVLT Total Learning (p = 0.001), AVLT Delayed Recall (p < 0.001), and verbal fluency (category) (p<0.001).

3.3.2. Baseline associations – ADNI Cohort

We performed standard linear regression analyses to examine the associations between baseline GFAP levels and baseline cognitive performance. There were no significant associations between baseline GFAP and global cognition or domain-specific assessments in the whole cohort (Supplemental Table 5) or stratified by males and females (Supplemental Table 6 and 7).

3.3.3. Longitudinal outcomes – MERI Cohort

We examined the associations between baseline GFAP levels and longitudinal changes in global cognition. In the full cohort, baseline GFAP was significantly associated with longitudinal declines in global cognition (Figure 4; p = 0.001). When analyses were stratified by sex, this association remained significant only in females (Table 3a). Next, we investigated the associations between baseline GFAP levels and domain-specific cognitive performance. In the full cohort, baseline GFAP was associated with poorer episodic memory performance (Table 4a), reflected by longitudinal decreases in AVLT Total Learning (p = 0.006) and AVLT Delayed Recall (p < 0.001). No significant associations were observed with measures of semantic memory, executive function, or language. When analyses were stratified by sex, no significant associations were found in males (Table 4b). In females (Table 4c), the same associations with episodic memory were observed, and baseline GFAP was additionally associated with poorer performance on a semantic memory measure, the verbal fluency (category) test (p = 0.042). Finally, the sex × GFAP interaction term was significant for AVLT Delayed Recall (p = 0.017), suggesting baseline GFAP was associated with steeper declines in episodic memory in females compared to males (Supplemental Table 11). A similar but non-significant pattern was observed for AVLT Total Learning (p = 0.137).

Figure 4. — Unadjusted linear trendlines depicting longitudinal MMSE scores obtained throughout follow up visits, comparing participants who were in the lower GFAP median (solid line) vs. upper GFAP median (dotted line) at baseline. Comparisons were plotted in the full MERI cohort (A), MERI cohort - male participants only (B), MERI cohort - female participants only (C), full ADNI cohort (D), ADNI cohort - male participants only (E), and ADNI cohort - female participants only (D),. Plots were constructed from the subset of 245 participants with 2+ visits where MMSE was completed. Y-axis is total MMSE score and X-axis is time (years).

Table 3.

Associations between baseline GFAP and longitudinal MMSE performance for a) the MERI cohort and b) the ADNI cohort. Results are derived from linear mixed-effects models with baseline GFAP as the primary exposure and total MMSE score across follow-up as the outcome variable. Each model was adjusted for baseline age, time since baseline, sex (unless stratified by sex), race, education, and APOE4 status.

Population	N participants (time points)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates
a) MERI Cohort

Full cohort	245 (765)	−1.964 (0.566)	0.001^***	time since baseline, age at baseline, sex, race, education
Males only	102 (325)	−1.466 (0.828)	0.080	time since baseline
Females only	143 (440)	−2.623 (0.816)	0.002^**	time since baseline

b) ADNI Cohort

Full cohort	447 (1,905)	−0.918 (0.330)	0.006^**	age, sex, education, time since baseline
Males only	195 (902)	−1.205 (0.518)	0.021^*	time since baseline
Females only	252 (1,003)	−0.713 (0.426)	0.095	education, time since baseline

Open in a new tab

Significant associations are indicated with an asterisks (p<0.05*, p<0.01**, p<0.001***).

Table 4.

Associations between baseline GFAP and longitudinal cognitive domain scores in the whole MERI cohort (a) and stratified by males (b) and c) females (c). Results are derived from linear mixed-effects models with baseline GFAP as the primary exposure and cognitive domain scores across follow-up as the outcome variables. Each model was adjusted for baseline age, time since baseline, race, education, and APOE4 status. In the whole-cohort analysis, sex was additionally included as a covariate.

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

Full Cohort (N=242 participants across 735 total visits)

AVLT, Total Learned	−9.279 (3.283)	0.005^**	time since baseline, age at baseline, sex, race, education
AVLT, Delayed Recall	−3.879 (1.185)	≤0.001^***	time since baseline, age at baseline, sex
Trail Making A (Time)	4.178 (4.381)	0.341	time since baseline, age at baseline, sex, race
Trail Making B (Time)	21.801 (16.612)	0.191	time since baseline, age at baseline, sex, race, education
Trail Making (B-A)	17.052 (13.534)	0.209	time since baseline, age at baseline, sex, race, education
Trail Making (B/A)	0.135 (0.285)	0.636	time since baseline, age at baseline, race, education
Verbal Fluency, Animals (Total)	−2.103 (1.488)	0.159	time since baseline, age at baseline, sex, race
Verbal Fluency, F+A+S (Total)	−2.197 (3.933)	0.577	time since baseline, sex, race, education
Males Only (N = 100 participants across 311 total visits)

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

AVLT, Total Learned	−5.112 (4.784)	0.288	time since baseline, age at baseline
AVLT, Delayed Recall	−1.255 (1.609)	0.437	No significant covariates
Trail Making A (Time)	8.019 (8.176)	0.330	time since baseline
Trail Making B (Time)	42.505 (30.784)	0.171	time since baseline, age at baseline
Trail Making (B-A)	33.837 (24.579)	0.173	time since baseline, age at baseline
Trail Making (B/A)	0.300 (0.449)	0.506	time since baseline, age at baseline
Verbal Fluency, Animals (Total)	0.314 (2.219)	0.888	time since baseline, age at baseline, race
Verbal Fluency, F+A+S (Total)	1.724 (5.810)	0.767	No significant covariates
Females Only (N = 142 participants across 424 total visits)

Outcome (longitudinal)	Baseline GFAP, β (SE)	Baseline GFAP, p-value	Significant Covariates

AVLT, Total Learned	−13.042 (4.667)	0.006^**	time since baseline, age at baseline
AVLT, Delayed Recall	−6.559 (1.746)	≤0.001^***	time since baseline
Trail Making A (Time)	1.883 (4.873)	0.700	time since baseline, age at baseline, race, APOE4
Trail Making B (Time)	6.853 (17.747)	0.700	time since baseline, age at baseline, race, education
Trail Making (B-A)	4.880 (14.769)	0.742	time since baseline, age at baseline, race, education
Trail Making (B/A)	−0.044 (0.369)	0.906	time since baseline, race, education
Verbal Fluency, Animals (Total)	−4.202 (2.045)	0.042^*	time since baseline, age at baseline, race
Verbal Fluency, F+A+S (Total)	−5.407 (5.409)	0.319	time since baseline, race, education

Open in a new tab

Significant associations are indicated with an asterisks (p<0.05*, p<0.01**, p<0.001***).

3.3.4. Longitudinal outcomes – ADNI Cohort

We examined the associations between baseline GFAP levels and longitudinal changes in global cognition (Table 3b). In the full cohort, baseline GFAP was significantly associated with longitudinal declines in global cognition (p = 0.006). When analyses were stratified by sex, this association remained significant only in males (p = 0.021 and was only marginally significant in females (p = 0.095). Next, we investigated the associations between baseline GFAP levels and domain-specific cognitive performance. In the full cohort, baseline GFAP was associated with poorer semantic memory performance (Supplemental Table 12a), reflected by longitudinal decreases in verbal fluency (category) test (p = 0.002) and poorer executive function, reflected by increased time to complete Trail Making B (p = 0.007). No significant associations were observed with measures of verbal memory or language. When analyses were stratified by sex, in males (Supplemental Table 12b) the same associations were found with semantic memory (p = 0.007) and additional a decrement in language was found (verbal fluency letter test p = 0.043). In females (Supplemental Table 12c), the same associations with executive function were observed (p = 0.002), and baseline GFAP was marginally associated with poorer performance on a semantic memory measure, the verbal fluency (category) test (p = 0.073). Finally, there were no significant between sex × GFAP interactions on either global or domain-specific cognition (Supplemental Table 13).

3.4. Longitudinal GFAP and Longitudinal BBM and Cognition analyses – MERI Cohort

We repeated the analyses for the MERI Cohort using longitudinal GFAP (i.e., a time-varying longitudinal measure of GFAP) as the primary exposure. First, we performed linear mixed-effects regressions to examine the associations between longitudinal GFAP levels and longitudinal BBMs. In the whole cohort, increases in GFAP were significantly associated with increases in Aβ40, Aβ42, p-Tau231, and NfL (p < 0.001; Supplemental Table 14a). These associations remained significant when analyses were repeated separately for females and males. Increases in GFAP were also associated with decreases in the Aβ42/40 ratio (p = 0.001); however, this association remained significant only in females (Supplemental Table 14c). Next, we conducted linear mixed-effects regressions to examine the associations between longitudinal GFAP levels and longitudinal global cognition. In both the whole cohort and in males only, increases in GFAP were associated with declines in MMSE performance (Supplemental Table 15). Finally, we assessed the associations between longitudinal GFAP levels and longitudinal domain-specific cognition. In the whole cohort, increases in GFAP were associated with declines in episodic memory performance (Supplemental Table 16a), reflected by longitudinal decreases in AVLT Total Learning (p < 0.001) and AVLT Delayed Recall (p = 0.003). Higher GFAP levels were also associated with poorer semantic memory performance on the verbal fluency (category) test (p = 0.013). In males (Supplemental Table 16b), in addition to significant associations with episodic memory, higher GFAP was related to poorer executive function performance (Trail Making Test B, p = 0.010). In females (Supplemental Table 16c), only associations with declines in AVLT Total Learning (p = 0.033) and verbal fluency (category) performance (p < 0.001) remained significant.

3.5. Rates of Change in GFAP and Longitudinal BBM analyses

3.5.1. MERI Cohort

We repeated the analyses for the MERI Cohort using Rate of Change in GFAP (i.e., a time-varying longitudinal measure of GFAP) as the primary exposure. Faster rates of GFAP increase were associated with faster rates of increase in Aβ40, Aβ42, p-Tau₂₃₁, and NfL, and faster rates of decrease in Aβ42/40 (all p≤0.012). In male-only analyses, associations persisted for all BBMs except p-tau₂₃₁ (Supplemental Table 17b). In female-only analyses, associations persisted for all BBMs except Aβ42 and Aβ42/40 which became marginal (p<0.100) (Supplemental Table 17c).

3.5.2. ADNI Cohort

We repeated the analyses using Rate of Change in GFAP for the ADNI cohort. Faster rates of GFAP increase were associated with faster rates of increase in all BBMs (all p<0.001). In male-only analyses, associations persisted for all BBMs except Aβ40 (Supplemental Table 18b). In female-only analyses, associations persisted across all BBMs (Supplemental Table 18c).

4. DISCUSSION

We examined associations between plasma GFAP and other BBMs in CU older adults from MERI and replicated the analyses in ADNI. In MERI, women had higher plasma GFAP and lower Aβ40, Aβ42, and p-Tau231 than men, with no sex differences in Aβ42/40. Higher baseline GFAP was associated with lower Aβ42/40 and higher p-Tau231 levels, consistent with greater AD-related pathology⁴⁹. Longitudinally, higher baseline GFAP and increasing GFAP were associated with decreasing Aβ42/40 and increasing p-Tau231, mainly in women. In ADNI, women also showed higher GFAP, and the absence of sex differences in Aβ42/40 was replicated. The inverse association between GFAP and Aβ42/40 and the association with longitudinal p-Tau217 increases were again evident in women. However, sex×GFAP interactions for BBM outcomes were not significant in either cohort. Thus, the female-predominant pattern should be interpreted as suggestive rather than definitive evidence of sex-specific BBM trajectories, and larger samples powered for interaction testing are needed.

Previous studies show that plasma GFAP increases across the AD continuum, discriminates preclinical AD from MCI and dementia, predicts cognitive and functional decline in Aβ+ individuals, and may rise years before clinical conversion^50–52. Postmortem studies similarly demonstrate that higher GFAP levels are associated with greater AD pathology⁵². Our findings extend this work by showing that GFAP relates to longitudinal amyloid- and tau-related plasma changes in CU individuals, particularly women, which may signal increased vulnerability to future cognitive decline and AD-related pathology. However, greater clinical progression was not observed and unlikely given the relatively short follow-up of approximately 3–4 years⁵². Nonetheless, elevated GFAP is not specific to AD. Chronic inflammatory conditions^53,54, major depressive disorder⁵⁵, bipolar depression⁵⁶, progressive multiple sclerosis⁵⁷, severe infections⁵⁸, and other systemic or neurological conditions may increase GFAP. Although MERI and ADNI participants were generally medically stable, comorbidities were not exclusionary and may have influenced GFAP levels.

While elevations in plasma GFAP have been consistently reported across the AD spectrum and in CU individuals^36,59, only recently have studies begun to examine the influence of sex. Across the AD spectrum, plasma GFAP levels are elevated in women^{22,32–35,37,60}, with only a few studies that have not observed sex differences in GFAP^41,61. Women exhibit faster increases in GFAP over time, but this may be depend on age³² as well as ε4 carrier status³⁴. Other reports suggest sex differences in associations with GFAP that are related to tau accumulation and vary by disease stage^36,62. CSF GFAP may also show different sex associations from plasma GFAP⁶³. This heterogeneity likely reflects differences in cohort composition, including age range, Aβ positivity, cognitive status, and assay platform. MERI was enriched for individuals with cognitive concerns but classified as CU, whereas ADNI CU participants were enrolled through standardized research criteria. Against these differences, the replication of higher GFAP in women and of no sex difference in Aβ42/40 supports the robustness of these signals.

Our study extends this literature in several ways. First, our primary analyses evaluated a strictly CU cohort recruited through physician- and self-referral for cognitive concerns, thereby addressing an earlier and less clinically confounded stage of the AD continuum. Second, we examined both sex differences in baseline and longitudinal GFAP in relation to cognitive trajectories within the same cohort. Third, by stratifying the analyses by sex, rather than only adjusting for sex as a covariate, we showed that the associations of GFAP with worsening amyloid-related and cognitive measures were concentrated in women, which helps with the understanding of heterogeneous findings across prior community-based and preclinical studies.

The absence of sex differences in the plasma Aβ42/40 in both cohorts suggests that greater amyloid burden does not fully explain higher GFAP in women^49,64. Sex differences in plasma tau were inconsistent, with higher p-Tau231 in men in MERI and higher p-Tau217 in women in ADNI, whereas larger community data reported no sex differences in p-Tau217⁶⁵. These discrepancies indicate that sex effects on plasma tau in CU populations remain unsettled. Prior work suggests that astrocyte reactivity may influence the relationship between amyloid and tau, and that diagnostic performance improves when GFAP is combined with p-Tau217^40,66. As our study did not include PET imaging or derive sex-specific cutoffs⁶³, conclusions about sex-specific-specific AD pathology remain speculative.

Higher GFAP was also associated with longitudinal cognitive decline. In MERI, baseline GFAP predicted global decline and episodic memory decline mainly in women, with a significant sex×GFAP interaction for AVLT Delayed Recall. In ADNI, GFAP predicted global decline mainly in men and was associated with executive and semantic memory outcomes, without significant sex interactions. Women had better baseline cognitive performance in both cohorts, consistent with prior work in older adults⁶⁷ and CU ADNI participants⁶⁸. Prior studies have linked higher GFAP to poorer global cognition^18,69–71, episodic memory^25,71, executive function^28,29 and language abilities^27,29. Therefore, the different domains implicated in MERI and ADNI likely reflect cohort characteristics, baseline performance, follow-up duration, and test sensitivity rather than divergent underlying biology.

Sex-specific glial and endocrine mechanisms may contribute to these findings⁷. Biological sex influences neuroinflammatory dynamics, and age-related endocrine changes affect astrocyte function^34,72. Estrogen modulates neuroimmune response⁷³, mitochondrial function⁷⁴, aromatase-dependent microglial activity⁷⁵, and neurotrophic signaling relevant to hippocampal activity⁷⁶. Evidence on hormone replacement therapy (HRT) during menopause is mixed⁷², with dementia risk depending on formulation and timing^77–79. Postmenopausal endocrine shifts may also alter microglial activation and astrocyte states⁸⁰. Because most women in our cohorts were postmenopausal, GFAP elevations may partly reflect these mechanisms⁸¹. Moreover, these findings raise the possibility that increased astrocytic reactivity and changes in astrocyte activation states may contribute to AD risk in women through mechanisms involving amyloid production or impaired clearance. However, menopause status, estrogen levels, and HRT use were not modeled, so this interpretation remains speculative.

The observation of higher GFAP in women despite no higher Aβ42/40 burden is clinically relevant. GFAP may help distinguish CU individuals with subjective complaints from Aβ– MCI⁸² and MERI was largely Aβ– based on plasma p-Tau231 cutoffs⁸³. Plasma GFAP may be more closely linked to brain amyloid pathology than CSF GFAP, whereas CSF GFAP may reflect broader neuroglial injury^22,84. Thus, plasma and CSF GFAP should not be treated as interchangeable. Additionally, astrocyte activation is not always a response to primary elevations in brain Aβ; thus, may reflect a response to blood–brain barrier (BBB) dysfunction or systemic inflammation⁸². Recent work also cautions that plasma GFAP cannot be interpreted as a pure measure of brain astrocyte reactivity because BBB dysfunction and non-astrocytic sources may contribute^22,84,85. Future studies using paired plasma and CSF sampling, PET, inflammatory markers, BBB measures, and estrogen-related measures are needed to clarify whether GFAP in CU women reflects early AD processes, systemic inflammation, BBB dysfunction, or overlapping mechanisms.

4.1. Limitations

Several limitations should be noted. First, MERI and ADNI were composed largely of highly educated, non-Hispanic White individuals, which limits generalizability. Second, women outnumbered men, although groups were broadly comparable and models adjusted for demographic variables and APOE. Previous studies suggest that plasma GFAP may not be strongly influenced by APOE-ε4 genotype⁸⁶, but APOE was included as a covariate. Third, GFAP increases with age, and age-related imbalance could influence results^12,87. However, MERI did not show a sex difference in the proportion aged ≥65 years (p=0.36), and analyses were adjusted for age. Fourth, direct measures of AD pathology were unavailable in MERI. Plasma AD biomarkers correlate with amyloid and tau pathology but remain indirect measures^47,88. Fifth, factors explaining GFAP variance including sex, only account for a small portion of variance in GFAP⁸⁹. Additional factors like body mass index, comorbidities, lifestyle⁸⁹, sleep quality and duration^90,91, and air pollution⁹² were not fully modeled. Finally, MERI participants were classified as CU but were enriched for subjective complaints or family history of dementia through physician- and self-referral recruitment. Findings should be interpreted in the context of this ascertainment and validated in more diverse population-based cohorts.

5. CONCLUSIONS

We found similar sex-based associations between GFAP and Aβ42/40 in our MERI cohort with the replication cohort extracted from the ADNI database. Based on prior biomarker and mechanistic studies, we propose that the observed elevations in plasma GFAP in women are likely downstream of initial Aβ accumulation and upstream of stronger tau-related and neurodegenerative changes. In this disease model, early amyloid-associated astrocyte reactivity may amplify microglial signaling, facilitate tau-related injury, and ultimately contribute to cognitive decline⁹³. These associations are likely influenced by sex-related immune and hormonal factors⁷² which is consistent with the female-predominant associations observed in our cohort and the ADNI. Moreover, this view is compatible with prior work showing that astrocyte biomarkers can mediate early AD progression and that astrocyte reactivity can influence the relationship between amyloid and tau pathology^40,94.

In conclusion, increased astrocytic reactivity in CU individuals may contribute to AD risk in women. In two independent cohorts, women exhibited higher plasma GFAP than men and, over time, showed evidence of increasing amyloid burden accompanied by declines in episodic memory and executive function. Future studies incorporating direct measures of brain AD pathology are needed to determine whether increased astrocytic reactivity in CU women arises in response to systemic inflammation or reflects early AD-related processes, including amyloid and tau pathology.

Supplementary Material

Supplement 1

media-1.zip^{(560.7KB, zip)}

ACKNOWLEDGEMENTS

We would like to acknowledge all the ADNI, the ADNI study sites and the ADNI participants for the use of the data analyzed as part of the replication analyses.

HZ is a Wallenberg Scholar and a Distinguished Professor at the Swedish Research Council supported by grants from the Swedish Research Council (#2023-00356, #2022-01018 and #2019-02397), the European Union’s Horizon Europe research and innovation programme under grant agreement No 101053962, Swedish State Support for Clinical Research (#ALFGBG-71320), the Alzheimer Drug Discovery Foundation (ADDF), USA (#201809-2016862), the AD Strategic Fund and the Alzheimer’s Association (#ADSF-21-831376-C, #ADSF-21-831381-C, #ADSF-21-831377-C, and #ADSF-24-1284328-C), the European Partnership on Metrology, co-financed from the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States (NEuroBioStand, #22HLT07), the Bluefield Project, Cure Alzheimer’s Fund, the Olav Thon Foundation, the Erling-Persson Family Foundation, Familjen Rönströms Stiftelse, Familjen Beiglers Stiftelse, Stiftelsen för Gamla Tjänarinnor, Hjärnfonden, Sweden (#FO2022-0270), the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 860197 (MIRIADE), the European Union Joint Programme – Neurodegenerative Disease Research (JPND2021-00694), the National Institute for Health and Care Research University College London Hospitals Biomedical Research Centre, the UK Dementia Research Institute at UCL (UKDRI-1003), and an anonymous donor.

FUDNING SOURCES

This MERI program has been partially supported by the Office of Mental Health of Rockland County, NY awarded to NP.

Data collection and sharing for the Alzheimer’s Disease Neuroimaging Initiative (ADNI) is funded by the National Institute on Aging (National Institutes of Health Grant U19 AG024904). The grantee organization is the Northern California Institute for Research and Education. In the past, ADNI has also received funding from the National Institute of Biomedical Imaging and Bioengineering, the Canadian Institutes of Health Research, and private sector contributions through the Foundation for the National Institutes of Health (FNIH) including generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research &Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics.

Footnotes

Declaration of generative AI use

During the preparation of this work the author(s) used ChatGPT 5.0 in order to check English grammar/style and improve readability. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.

CONSENT STATEMENT

The study protocol was approved by the Nathan S. Kline Institute/Rockland Psychiatric Center Institutional Review Board. All participants provided informed consent.

CONFLICTS OF INTEREST

CRP, TJ, DB, SHL, BPI, RO, ALB, NA, and NP have nothing to report. HZ has served at scientific advisory boards and/or as a consultant for Abbvie, Acumen, Alector, Alzinova, ALZpath, Amylyx, Annexon, Apellis, Artery Therapeutics, AZTherapies, Cognito Therapeutics, CogRx, Denali, Eisai, Enigma, LabCorp, Merck Sharp & Dohme, Merry Life, Nervgen, Novo Nordisk, Optoceutics, Passage Bio, Pinteon Therapeutics, Prothena, Quanterix, Red Abbey Labs, reMYND, Roche, Samumed, ScandiBio Therapeutics AB, Siemens Healthineers, Triplet Therapeutics, and Wave, has given lectures sponsored by Alzecure, BioArctic, Biogen, Cellectricon, Fujirebio, LabCorp, Lilly, Novo Nordisk, Oy Medix Biochemica AB, Roche, and WebMD, is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program, and is a shareholder of CERimmune Therapeutics (outside submitted work).

Data Sharing and Code Availability

The MERI data associated with this publication is not publicly available. The original study protocol did not include a data sharing provision; therefore, participants did not agree for their data to be shared publicly. The code generated for this analysis is available upon request. The ADNI data associated with this publication is publicly available to approved researchers (https://adni.loni.usc.edu/data-samples/adni-data/).

References

1.Budson A. Why are women more likely to develop Alzheimer’s disease. Harvard Health Publishing. 2022; [Google Scholar]
2.O’Neal MA. Women and the risk of Alzheimer’s disease. Frontiers in Global Women’s Health. 2024;4:1324522. doi: 10.3389/fgwh.2023.1324522 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Nebel RA, Aggarwal NT, Barnes LL, et al. Understanding the impact of sex and gender in Alzheimer’s disease: a call to action. Alzheimer’s & Dementia. 2018;14(9):1171–1183. doi: 10.1016/j.jalz.2018.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Nerattini M, Jett S, Andy C, et al. Systematic review and meta-analysis of the effects of menopause hormone therapy on risk of Alzheimer’s disease and dementia. Frontiers in aging neuroscience. 2023;15:1260427. doi: 10.3389/fnagi.2023.1260427 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Bonkhoff AK, Coughlan G, Perosa V, et al. Sex differences in age-associated neurological diseases—A roadmap for reliable and high-yield research. Science Advances. 2025;11(10):eadt9243. doi: 10.1126/sciadv.adt9243 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Heneka MT, van der Flier WM, Jessen F, et al. Neuroinflammation in Alzheimer disease. Nature Reviews Immunology. 2024/12/09 2024;doi: 10.1038/s41577-024-01104-7 [DOI] [Google Scholar]
7.Müller L, Di Benedetto S, Müller V. Influence of biological sex on neuroinflammatory dynamics in the aging brain. Frontiers in Aging Neuroscience. 2025;17:1670175. doi: 10.3389/fnagi.2025.1670175 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Bortolanza M, Gobbo D, Fang L, Scheller A, Bai X, Kirchhoff F. Sex-Specific Properties of Astrocytes: From Development to Evolutionary Insights. Neurochemical Research. 2025;50(4):267. doi: 10.1007/s11064-025-04512-w [DOI] [PMC free article] [PubMed] [Google Scholar]
9.DiSabato DJ, Quan N, Godbout JP. Neuroinflammation: the devil is in the details. Journal of neurochemistry. 2016;139:136–153. doi: 10.1111/jnc.13607 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Leng F, Edison P. Neuroinflammation and microglial activation in Alzheimer disease: where do we go from here? Nature Reviews Neurology. 2021;17(3):157–172. doi: 10.1038/s41582-020-00435-y [DOI] [PubMed] [Google Scholar]
11.Ulrich JD, Ulland TK, Colonna M, Holtzman DM. Elucidating the role of TREM2 in Alzheimer’s disease. Neuron. 2017;94(2):237–248. doi: 10.1016/j.neuron.2017.02.042 [DOI] [PubMed] [Google Scholar]
12.Frost GR, Li Y-M. The role of astrocytes in amyloid production and Alzheimer’s disease. Open biology. 2017;7(12):170228. doi: 10.1098/rsob.170228 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Fleeman RM, Proctor EA. Astrocytic propagation of tau in the context of Alzheimer’s disease. Frontiers in Cellular Neuroscience. 2021;15:645233. doi: 10.3389/fncel.2021.645233 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Jiwaji Z, Tiwari SS, Avilés-Reyes RX, et al. Reactive astrocytes acquire neuroprotective as well as deleterious signatures in response to Tau and Aß pathology. Nature communications. 2022;13(1):135. doi: 10.1038/s41467-021-27702-w [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Liddelow SA, Guttenplan KA, Clarke LE, et al. Neurotoxic reactive astrocytes are induced by activated microglia. Nature. 2017;541(7638):481–487. doi:http://www.nature.com/doifinder/10.1038/nature21029 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Oeckl P, Halbgebauer S, Anderl-Straub S, et al. Glial fibrillary acidic protein in serum is increased in Alzheimer’s disease and correlates with cognitive impairment. Journal of Alzheimer’s Disease. 2019;67(2):481–488. doi: 10.3233/JAD-180325 [DOI] [PubMed] [Google Scholar]
17.Garwood C, Pooler A, Atherton J, Hanger D, Noble W. Astrocytes are important mediators of Aβ-induced neurotoxicity and tau phosphorylation in primary culture. Cell death & disease. 2011;2(6):e167–e167. doi: 10.1038/cddis.2011.50 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Pereira JB, Janelidze S, Smith R, et al. Plasma GFAP is an early marker of amyloid-β but not tau pathology in Alzheimer’s disease. Brain. 2021;144(11):3505–3516. doi: 10.1093/brain/awab223 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Lista S, Imbimbo BP, Grasso M, et al. Tracking neuroinflammatory biomarkers in Alzheimer’s disease: a strategy for individualized therapeutic approaches? Journal of Neuroinflammation. 2024;21(1):187. doi: 10.1186/s12974-024-03163-y [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Phillips JM, Winfree RL, Seto M, et al. Pathologic and clinical correlates of region-specific brain GFAP in Alzheimer’s disease. Acta Neuropathologica. 2024;148(1):69. doi: 10.1007/s00401-024-02828-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Sánchez-Juan P, Valeriano-Lorenzo E, Ruiz-González A, et al. Serum GFAP levels correlate with astrocyte reactivity, post-mortem brain atrophy and neurofibrillary tangles. Brain. 2024;147(5):1667–1679. doi: 10.1093/brain/awae035 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Benedet AL, Milà-Alomà M, Vrillon A, et al. Differences between plasma and cerebrospinal fluid glial fibrillary acidic protein levels across the Alzheimer disease continuum. JAMA neurology. 2021;78(12):1471–1483. doi: 10.1001/jamaneurol.2021.3671 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Chatterjee P, Pedrini S, Stoops E, et al. Plasma glial fibrillary acidic protein is elevated in cognitively normal older adults at risk of Alzheimer’s disease. Translational psychiatry. 2021;11(1):27. doi: 10.1038/s41398-020-01137-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Guo Y, Yu J-T. Blood protein markers predict 15-year risk of dementia. Nature. 2024;4:173–174. doi: 10.1038/s43587-023-00566-z [DOI] [PubMed] [Google Scholar]
25.Bettcher BM, Olson KE, Carlson NE, et al. Astrogliosis and episodic memory in late life: higher GFAP is related to worse memory and white matter microstructure in healthy aging and Alzheimer’s disease. Neurobiology of aging. 2021;103:68–77. doi: 10.1016/j.neurobiolaging.2021.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Santello M, Toni N, Volterra A. Astrocyte function from information processing to cognition and cognitive impairment. Nature neuroscience. 2019;22(2):154–166. doi: 10.1038/s41593-018-0325-8 [DOI] [PubMed] [Google Scholar]
27.De Meyer S, Blujdea ER, Schaeverbeke J, et al. Longitudinal associations of serum biomarkers with early cognitive, amyloid and grey matter changes. Brain. 2024;147(3):936–948. doi: 10.1093/brain/awad330 [DOI] [PubMed] [Google Scholar]
28.Teitsdottir UD, Jonsdottir MK, Lund SH, Darreh-Shori T, Snaedal J, Petersen PH. Association of glial and neuronal degeneration markers with Alzheimer’s disease cerebrospinal fluid profile and cognitive functions. Alzheimer’s research & therapy. 2020;12:1–14. doi: 10.1186/s13195-020-00657-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Asken BM, VandeVrede L, Rojas JC, et al. Lower white matter volume and worse executive functioning reflected in higher levels of plasma GFAP among older adults with and without cognitive impairment. Journal of the International Neuropsychological Society. 2022;28(6):588–599. doi: 10.1017/S1355617721000813 [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Sun W, Ye S, Wang Y, et al. Plasma biomarkers for diagnosis and differentiation and their cognitive correlations in patients with Alzheimer’s disease. Brain Communications. 2025;7(2):fcaf094. doi: 10.1093/braincomms/fcaf094 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Mielke MM. Consideration of sex differences in the measurement and interpretation of Alzheimer disease-related biofluid-based biomarkers. The journal of applied laboratory medicine. 2020;5(1):158–169. doi: 10.1373/jalm.2019.030023 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Yakoub Y, Ashton NJ, Strikwerda Brown C, et al. Longitudinal blood biomarker trajectories in preclinical Alzheimer’s disease. Alzheimer’s & Dementia. 2023;19(12):5620–5631. doi: 10.1002/alz.13318 [DOI] [PubMed] [Google Scholar]
33.Boccalini C, Peretti DE, Scheffler M, et al. Sex differences in the association of Alzheimer’s disease biomarkers and cognition in a multicenter memory clinic study. Alzheimer’s research & therapy. 2025;17:46. doi: 10.1186/s13195-025-01684-z [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Stocker H, Beyer L, Trares K, et al. Association of nonmodifiable risk factors with Alzheimer disease blood biomarkers in community-dwelling adults in the ESTHER study. Neurology. 2025;104(9):e213500. doi: 10.1212/WNL.0000000000213500 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Milà Alomà M, Hausle I, Myoraku A, et al. Sex differences in Alzheimer’s disease plasma biomarker levels and clinical utility. Alzheimer’s & Dementia. 2026;22(2):e71244. doi: 10.1002/alz.71244 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Foley KE, Winder Z, Sudduth TL, et al. Alzheimer’s disease and inflammatory biomarkers positively correlate in plasma in the UK ADRC cohort. Alzheimer’s & Dementia. 2024;20(2):1374–1386. doi: 10.1002/alz.13485 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Vila-Castelar C, Akinci M, Palpatzis E, et al. Sex/gender effects of glial reactivity on preclinical Alzheimer’s disease pathology. Molecular Psychiatry. 2024:1–10. doi: 10.1038/s41380-024-02753-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Labonte J, Pettway EC, Sugarman M, et al. Sex Differences on Tau, Astrocytic and Neurodegeneration Plasma Biomarkers. Alzheimer’s & Dementia. 2023;19:e071807. doi: 10.1002/alz.071807 [DOI] [Google Scholar]
39.Duarte-Guterman P, Albert AY, Inkster AM, Barha CK, Galea LA, Initiative AsDN. Inflammation in Alzheimer’s disease: do sex and APOE matter? Journal of Alzheimer’s Disease. 2020;78(2):627–641. doi: 10.3233/JAD-200982 [DOI] [PubMed] [Google Scholar]
40.Bellaver B, Povala G, Ferreira PC, et al. Astrocyte reactivity influences amyloid-β effects on tau pathology in preclinical Alzheimer’s disease. Nature medicine. 2023;29(7):1775–1781. doi: 10.1038/s41591-023-02380-x [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Joynes CM, Bilgel M, An Y, et al. Sex differences in the trajectories of plasma biomarkers, brain atrophy, and cognitive decline relative to amyloid onset. Alzheimer’s & Dementia. 2024;doi: 10.1002/alz.14405 [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Rohden F, Ferreira PC, Bellaver B, et al. Glial reactivity correlates with synaptic dysfunction across aging and Alzheimer’s disease. Nature Communications. 2025;16(1):5653. doi: 10.1038/s41467-025-60806-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Beckett LA, Donohue MC, Wang C, et al. The Alzheimer’s Disease Neuroimaging Initiative phase 2: Increasing the length, breadth, and depth of our understanding. Alzheimer’s & Dementia. 2015;11(7):823–831. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Reichert C, Sidtis JJ, Pomara N. The Memory Education and Research Initiative. The Preservation of Memory. 2015;6 [Google Scholar]
45.Bäckman L, Jones S, Berger A-K, Laukka EJ, Small BJ. Cognitive impairment in preclinical Alzheimer’s disease: a meta-analysis. Neuropsychology. 2005;19(4):520. doi: 10.1037/0894-4105.19.4.520 [DOI] [PubMed] [Google Scholar]
46.Pomara N, Bruno D, Plaska CR, et al. Plasma Amyloid-β dynamics in late-life major depression: a longitudinal study. Translational Psychiatry. 2022;12(1):301. doi: 10.1038/s41398-022-02077-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Ashton NJ, Pascoal TA, Karikari TK, et al. Plasma p-tau231: a new biomarker for incipient Alzheimer’s disease pathology. Acta neuropathologica. 2021;141(5):709–724. doi: 10.1007/s00401-021-02275-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Shaw LM, Korecka M, Lee EB, et al. ADNI Biomarker Core: A review of progress since 2004 and future challenges. Alzheimer’s & Dementia. 2025;21(1):e14264. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Doecke JD, Pérez-Grijalba V, Fandos N, et al. Total Aβ42/Aβ40 ratio in plasma predicts amyloid-PET status, independent of clinical AD diagnosis. Neurology. 2020;94(15):e1580–e1591. doi: 10.1212/WNL.0000000000009240 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Wojdała AL, Bellomo G, Gaetani L, et al. Trajectories of CSF and plasma biomarkers across Alzheimer’s disease continuum: disease staging by NF-L, p-tau181, and GFAP. Neurobiology of Disease. 2023;189:106356. doi: 10.1016/j.nbd.2023.106356 [DOI] [PubMed] [Google Scholar]
51.Abbas S, Ferreira PC, Bellaver B, et al. Utility of plasma GFAP as a secondary endpoint for clinical trials in Alzheimer’s disease. The Journal of Prevention of Alzheimer’s Disease. 2025:100205. doi: 10.1016/j.tjpad.2025.100205 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Varma V, An Y, Kac PR, et al. Longitudinal progression of blood biomarkers reveals a key role of astrocyte reactivity in preclinical Alzheimer’s disease. medRxiv. 2024;doi: 10.1101/2024.01.25.24301779 [DOI] [PubMed] [Google Scholar]
53.Trabace L, Roviezzo F, Rossi A. Sex differences in Inflammatory diseases. Frontiers in Pharmacology. 2022;13:962869. doi: 10.3389/fphar.2022.962869 [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Ramos JJJ, Pupo NMG, Mena D, et al. Cognitive Decline in Chronic Inflammatory Conditions: Exploring Links Between Systemic Inflammation and Neurodegeneration. Cureus. 2025;17(7)doi: 10.7759/cureus.88397 [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Michel M, Fiebich BL, Kuzior H, et al. Increased GFAP concentrations in the cerebrospinal fluid of patients with unipolar depression. Translational psychiatry. 2021;11(1):308. doi: 10.1038/s41398-021-01423-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Kang MJ, Eratne D, Dean O, et al. Plasma Glial Fibrillary Acidic Protein and Neurofilament Light Are Elevated in Bipolar Depression: Evidence for Neuroprogression and Astrogliosis. Bipolar Disorders. 2025;doi: 10.1111/bdi.70029 [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Chitnis T, Magliozzi R, Abdelhak A, Kuhle J, Leppert D, Bielekova B. Blood and CSF biomarkers for multiple sclerosis: emerging clinical applications. The Lancet Neurology. 2025;doi: 10.1016/S1474-4422(25)00249-2 [DOI] [PubMed] [Google Scholar]
58.Passos FRS, Heimfarth L, Monteiro BS, et al. Oxidative stress and inflammatory markers in patients with COVID-19: Potential role of RAGE, HMGB1, GFAP and COX-2 in disease severity. International Immunopharmacology. 2022;104:108502. doi: 10.1016/j.intimp.2021.108502 [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Stevenson-Hoare J, Heslegrave A, Leonenko G, et al. Plasma biomarkers and genetics in the diagnosis and prediction of Alzheimer’s disease. Brain. 2023;146(2):690–699. doi: 10.1093/brain/awac128 [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Labonte J, Sugarman MA, Pettway E, et al. Sex differences on tau, astrocytic, and neurodegenerative plasma biomarkers. Journal of Alzheimer’s Disease. 2025;105(2):443–452. doi: 10.1177/13872877251329468 [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Mila-Aloma M, Salvado G, Gispert J, et al. Amyloid beta, tau, synaptic, neurodegeneration, and glial biomarkers in the preclinical stage of the Alzheimer’s continuum. Alzheimer’s Dementia 2020;16(10):1358–1371. doi: 10.1002/alz.12131 [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Wang X, Ye T, Dai B, Zhang J, Zhou W, Initiative AsDN. Sex-specific patterns in tau spreading throughout the Braak stages in the Alzheimer’s disease spectrum. Journal of Alzheimer’s Disease. 2026:13872877251406131. doi: 10.1177/13872877251406131 [DOI] [PubMed] [Google Scholar]
63.Milà-Alomà M, Van Hulle C, Brugulat-Serrat A, et al. Sex differences in Alzheimer’s disease CSF biomarkers and their association with Aβ pathology on PET in cognitively unimpaired individuals. Alzheimer’s Research & Therapy. 2025;17(1):235. doi: 10.1186/s13195-025-01844-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Graff-Radford NR, Crook JE, Lucas J, et al. Association of low plasma Aβ42/Aβ40 ratios with increased imminent risk for mild cognitive impairment and Alzheimer disease. Archives of neurology. 2007;64(3):354–362. doi: 10.1001/archneur.64.3.354 [DOI] [PubMed] [Google Scholar]
65.Aarsland D, Sunde AL, Tovar-Rios DA, et al. Prevalence of Alzheimer’s disease pathology in the community. Nature. 2025:1–5. doi: 10.1038/s41586-025-09841-y [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Wu CY, Chen L, Fatima H, et al. Combined use of plasma p tau217, NfL, and GFAP predicts domain specific cognitive decline in cognitively unimpaired and MCI individuals. Alzheimer’s & Dementia. 2025;21(12):e70934. doi: 10.1002/alz.70934 [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Julian R, Fröhlich S, Müller K, Dammhahn M, Voelcker-Rehage C. Sex differences in cognitive performance persist into your 80s. GeroScience. 2025:1–14. doi: 10.1007/s11357-025-01585-x [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Liu M, Bhatt K, Giaever MF, et al. Sex differences in cognitive performance in Alzheimer’s disease: Insights from the ADAS-Cog-13. Journal of Alzheimer’s Disease. 2025:13872877261440125. [DOI] [PubMed] [Google Scholar]
69.Gonzales MM, Wiedner C, Wang CP, et al. A population based meta analysis of circulating GFAP for cognition and dementia risk. Annals of Clinical and Translational Neurology. 2022;9(10):1574–1585. doi: 10.1002/acn3.51652 [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Rajan KB, Aggarwal NT, McAninch EA, et al. Remote blood biomarkers of longitudinal cognitive outcomes in a population study. Annals of neurology. 2020;88(6):1065–1076. doi: 10.1002/ana.25874 [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Gonzales MM, Short MI, Satizabal CL, et al. Blood biomarkers for dementia in Hispanic and non Hispanic White adults. Alzheimer’s & Dementia: Translational Research & Clinical Interventions. 2021;7(1):e12164. doi: 10.1002/trc2.12164 [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Price BR, Walker KA, Eissman JM, et al. Sex differences and the role of estrogens in the immunological underpinnings of Alzheimer’s disease. Alzheimer’s & Dementia: Translational Research & Clinical Interventions. 2025;11(3):e70139. doi: 10.1002/trc2.70139 [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Dimayuga FO, Reed JL, Carnero GA, et al. Estrogen and brain inflammation: effects on microglial expression of MHC, costimulatory molecules and cytokines. Journal of neuroimmunology. 2005;161(1–2):123–136. doi: 10.1016/j.jneuroim.2004.12.016 [DOI] [PubMed] [Google Scholar]
74.Rettberg JR, Yao J, Brinton RD. Estrogen: a master regulator of bioenergetic systems in the brain and body. Frontiers in neuroendocrinology. 2014;35(1):8–30. doi: 10.1016/j.yfrne.2013.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Medway C, Combarros O, Cortina-Borja M, et al. The sex-specific associations of the aromatase gene with Alzheimer’s disease and its interaction with IL10 in the Epistasis Project. European Journal of Human Genetics. 2014;22(2):216–220. doi: 10.1038/ejhg.2013.116 [DOI] [PMC free article] [PubMed] [Google Scholar]
76.Scharfman HE, MacLusky NJ. Estrogen and brain-derived neurotrophic factor (BDNF) in hippocampus: complexity of steroid hormone-growth factor interactions in the adult CNS. Frontiers in neuroendocrinology. 2006;27(4):415–435. doi: 10.1016/j.yfrne.2006.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
77.Coughlan GT, Betthauser TJ, Boyle R, et al. Association of age at menopause and hormone therapy use with tau and β-amyloid positron emission tomography. JAMA neurology. 2023;80(5):462–473. doi: 10.1001/jamaneurol.2023.0455 [DOI] [PMC free article] [PubMed] [Google Scholar]
78.Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. Jama. 2003;289(20):2651–2662. doi: 10.1001/jama.289.20.2651 [DOI] [PubMed] [Google Scholar]
79.Shadyab AH, Zhang B, LaCroix AZ, et al. Plasma Phosphorylated Tau 217 and Incident Mild Cognitive Impairment and Dementia in Older Women. JAMA Network Open. 2026;9(3):e261295. doi: 10.1001/jamanetworkopen.2026.1295 [DOI] [PMC free article] [PubMed] [Google Scholar]
80.Hanamsagar R, Bilbo SD. Sex differences in neurodevelopmental and neurodegenerative disorders: focus on microglial function and neuroinflammation during development. The Journal of steroid biochemistry and molecular biology. 2016;160:127–133. doi: 10.1016/j.jsbmb.2015.09.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
81.Zárate S, Stevnsner T, Gredilla R. Role of estrogen and other sex hormones in brain aging. Neuroprotection and DNA repair. Frontiers in aging neuroscience. 2017;9:322754. doi: 10.3389/fnagi.2017.00430 [DOI] [PMC free article] [PubMed] [Google Scholar]
82.Yoon B, Lee H-j, Chae H, et al. Plasma glial fibrillary acidic protein as a potential biomarker for differentiating amyloid-negative subjective cognitive decline and mild cognitive impairment. Journal of Alzheimer’s Disease. 2025:13872877261427806. doi: 10.1177/13872877261427806 [DOI] [PubMed] [Google Scholar]
83.Pais MV, Forlenza OV, Diniz BS. Plasma biomarkers of Alzheimer’s disease: a review of available assays, recent developments, and implications for clinical practice. Journal of Alzheimer’s disease reports. 2023;7(1):355–380. doi: 10.3233/ADR-230029 [DOI] [PMC free article] [PubMed] [Google Scholar]
84.Simrén J, Weninger H, Brum WS, et al. Differences between blood and cerebrospinal fluid glial fibrillary acidic protein levels: the effect of sample stability. Alzheimer’s & Dementia. 2022;18(10):1988–1992. doi: 10.1002/alz.12806 [DOI] [PMC free article] [PubMed] [Google Scholar]
85.Youn W, Yun M, Lee CJ, Schöll M. Cautions on utilizing plasma GFAP level as a biomarker for reactive astrocytes in neurodegenerative diseases. Molecular Neurodegeneration. 2025;20(1):54. doi: 10.1186/s13024-025-00846-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
86.Snellman A, Ekblad LL, Ashton NJ, et al. Head-to-head comparison of plasma p-tau181, p-tau231 and glial fibrillary acidic protein in clinically unimpaired elderly with three levels of APOE4-related risk for Alzheimer’s disease. Neurobiology of Disease. 2023;183:106175. doi: 10.1016/j.nbd.2023.106175 [DOI] [PubMed] [Google Scholar]
87.David J-P, Ghozali F, Fallet-Bianco C, et al. Glial reaction in the hippocampal formation is highly correlated with aging in human brain. Neuroscience letters. 1997;235(1–2):53–56. doi: 10.1016/S0304-3940(97)00708-8 [DOI] [PubMed] [Google Scholar]
88.Jack CR Jr, Wiste HJ, Algeciras-Schimnich A, et al. Predicting amyloid PET and tau PET stages with plasma biomarkers. Brain. 2023;146(5):2029–2044. doi: 10.1093/brain/awad042 [DOI] [PMC free article] [PubMed] [Google Scholar]
89.Bluma M, Chiotis K, Bucci M, et al. Disentangling relationships between Alzheimer’s disease plasma biomarkers and established biomarkers in patients of tertiary memory clinics. EBioMedicine. 2025;112doi: 10.1016/j.ebiom.2024.105504 [DOI] [PMC free article] [PubMed] [Google Scholar]
90.Young VM, Wiedner C, Baril AA, et al. Sex differences in the association between sleep duration and blood based biomarkers of Alzheimer’s disease and neurodegeneration: The Framingham Heart Study. Alzheimer’s & Dementia. 2025;21:e099558. doi: 10.1002/alz70856_099558 [DOI] [Google Scholar]
91.Van Egroo M, Beckers E, Ashton NJ, Blennow K, Zetterberg H, Jacobs HI. Sex differences in the relationships between 24-h rest-activity patterns and plasma markers of Alzheimer’s disease pathology. Alzheimer’s research & therapy. 2024;16(1):277. doi: 10.1186/s13195-024-01653-y [DOI] [PMC free article] [PubMed] [Google Scholar]
92.Balachandar R, Viramgami A, Singh DP, et al. Association between chronic PM2. 5 exposure and neurodegenerative biomarkers in adults from critically polluted area. BMC Public Health. 2025;25(1):1413. doi: 10.1186/s12889-025-22641-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
93.Ferrari-Souza JP, Povala G, Rahmouni N, et al. Microglia modulate Aβ-dependent astrocyte reactivity in Alzheimer’s disease. Nature Neuroscience. 2025:1–7. doi: 10.1038/s41593-025-02103-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
94.Pelkmans W, Shekari M, Brugulat Serrat A, et al. Astrocyte biomarkers GFAP and YKL 40 mediate early Alzheimer’s disease progression. Alzheimer’s & Dementia. 2024;20(1):483–493. doi: 10.1002/alz.13450 [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.

Supplementary Materials

Supplement 1

media-1.zip^{(560.7KB, zip)}

Data Availability Statement

[R1] 1.Budson A. Why are women more likely to develop Alzheimer’s disease. Harvard Health Publishing. 2022; [Google Scholar]

[R2] 2.O’Neal MA. Women and the risk of Alzheimer’s disease. Frontiers in Global Women’s Health. 2024;4:1324522. doi: 10.3389/fgwh.2023.1324522 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R3] 3.Nebel RA, Aggarwal NT, Barnes LL, et al. Understanding the impact of sex and gender in Alzheimer’s disease: a call to action. Alzheimer’s & Dementia. 2018;14(9):1171–1183. doi: 10.1016/j.jalz.2018.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] 4.Nerattini M, Jett S, Andy C, et al. Systematic review and meta-analysis of the effects of menopause hormone therapy on risk of Alzheimer’s disease and dementia. Frontiers in aging neuroscience. 2023;15:1260427. doi: 10.3389/fnagi.2023.1260427 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] 5.Bonkhoff AK, Coughlan G, Perosa V, et al. Sex differences in age-associated neurological diseases—A roadmap for reliable and high-yield research. Science Advances. 2025;11(10):eadt9243. doi: 10.1126/sciadv.adt9243 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Heneka MT, van der Flier WM, Jessen F, et al. Neuroinflammation in Alzheimer disease. Nature Reviews Immunology. 2024/12/09 2024;doi: 10.1038/s41577-024-01104-7 [DOI] [Google Scholar]

[R7] 7.Müller L, Di Benedetto S, Müller V. Influence of biological sex on neuroinflammatory dynamics in the aging brain. Frontiers in Aging Neuroscience. 2025;17:1670175. doi: 10.3389/fnagi.2025.1670175 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Bortolanza M, Gobbo D, Fang L, Scheller A, Bai X, Kirchhoff F. Sex-Specific Properties of Astrocytes: From Development to Evolutionary Insights. Neurochemical Research. 2025;50(4):267. doi: 10.1007/s11064-025-04512-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.DiSabato DJ, Quan N, Godbout JP. Neuroinflammation: the devil is in the details. Journal of neurochemistry. 2016;139:136–153. doi: 10.1111/jnc.13607 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Leng F, Edison P. Neuroinflammation and microglial activation in Alzheimer disease: where do we go from here? Nature Reviews Neurology. 2021;17(3):157–172. doi: 10.1038/s41582-020-00435-y [DOI] [PubMed] [Google Scholar]

[R11] 11.Ulrich JD, Ulland TK, Colonna M, Holtzman DM. Elucidating the role of TREM2 in Alzheimer’s disease. Neuron. 2017;94(2):237–248. doi: 10.1016/j.neuron.2017.02.042 [DOI] [PubMed] [Google Scholar]

[R12] 12.Frost GR, Li Y-M. The role of astrocytes in amyloid production and Alzheimer’s disease. Open biology. 2017;7(12):170228. doi: 10.1098/rsob.170228 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Fleeman RM, Proctor EA. Astrocytic propagation of tau in the context of Alzheimer’s disease. Frontiers in Cellular Neuroscience. 2021;15:645233. doi: 10.3389/fncel.2021.645233 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.Jiwaji Z, Tiwari SS, Avilés-Reyes RX, et al. Reactive astrocytes acquire neuroprotective as well as deleterious signatures in response to Tau and Aß pathology. Nature communications. 2022;13(1):135. doi: 10.1038/s41467-021-27702-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[R15] 15.Liddelow SA, Guttenplan KA, Clarke LE, et al. Neurotoxic reactive astrocytes are induced by activated microglia. Nature. 2017;541(7638):481–487. doi:http://www.nature.com/doifinder/10.1038/nature21029 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Oeckl P, Halbgebauer S, Anderl-Straub S, et al. Glial fibrillary acidic protein in serum is increased in Alzheimer’s disease and correlates with cognitive impairment. Journal of Alzheimer’s Disease. 2019;67(2):481–488. doi: 10.3233/JAD-180325 [DOI] [PubMed] [Google Scholar]

[R17] 17.Garwood C, Pooler A, Atherton J, Hanger D, Noble W. Astrocytes are important mediators of Aβ-induced neurotoxicity and tau phosphorylation in primary culture. Cell death & disease. 2011;2(6):e167–e167. doi: 10.1038/cddis.2011.50 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Pereira JB, Janelidze S, Smith R, et al. Plasma GFAP is an early marker of amyloid-β but not tau pathology in Alzheimer’s disease. Brain. 2021;144(11):3505–3516. doi: 10.1093/brain/awab223 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Lista S, Imbimbo BP, Grasso M, et al. Tracking neuroinflammatory biomarkers in Alzheimer’s disease: a strategy for individualized therapeutic approaches? Journal of Neuroinflammation. 2024;21(1):187. doi: 10.1186/s12974-024-03163-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Phillips JM, Winfree RL, Seto M, et al. Pathologic and clinical correlates of region-specific brain GFAP in Alzheimer’s disease. Acta Neuropathologica. 2024;148(1):69. doi: 10.1007/s00401-024-02828-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Sánchez-Juan P, Valeriano-Lorenzo E, Ruiz-González A, et al. Serum GFAP levels correlate with astrocyte reactivity, post-mortem brain atrophy and neurofibrillary tangles. Brain. 2024;147(5):1667–1679. doi: 10.1093/brain/awae035 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] 22.Benedet AL, Milà-Alomà M, Vrillon A, et al. Differences between plasma and cerebrospinal fluid glial fibrillary acidic protein levels across the Alzheimer disease continuum. JAMA neurology. 2021;78(12):1471–1483. doi: 10.1001/jamaneurol.2021.3671 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R23] 23.Chatterjee P, Pedrini S, Stoops E, et al. Plasma glial fibrillary acidic protein is elevated in cognitively normal older adults at risk of Alzheimer’s disease. Translational psychiatry. 2021;11(1):27. doi: 10.1038/s41398-020-01137-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] 24.Guo Y, Yu J-T. Blood protein markers predict 15-year risk of dementia. Nature. 2024;4:173–174. doi: 10.1038/s43587-023-00566-z [DOI] [PubMed] [Google Scholar]

[R25] 25.Bettcher BM, Olson KE, Carlson NE, et al. Astrogliosis and episodic memory in late life: higher GFAP is related to worse memory and white matter microstructure in healthy aging and Alzheimer’s disease. Neurobiology of aging. 2021;103:68–77. doi: 10.1016/j.neurobiolaging.2021.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Santello M, Toni N, Volterra A. Astrocyte function from information processing to cognition and cognitive impairment. Nature neuroscience. 2019;22(2):154–166. doi: 10.1038/s41593-018-0325-8 [DOI] [PubMed] [Google Scholar]

[R27] 27.De Meyer S, Blujdea ER, Schaeverbeke J, et al. Longitudinal associations of serum biomarkers with early cognitive, amyloid and grey matter changes. Brain. 2024;147(3):936–948. doi: 10.1093/brain/awad330 [DOI] [PubMed] [Google Scholar]

[R28] 28.Teitsdottir UD, Jonsdottir MK, Lund SH, Darreh-Shori T, Snaedal J, Petersen PH. Association of glial and neuronal degeneration markers with Alzheimer’s disease cerebrospinal fluid profile and cognitive functions. Alzheimer’s research & therapy. 2020;12:1–14. doi: 10.1186/s13195-020-00657-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Asken BM, VandeVrede L, Rojas JC, et al. Lower white matter volume and worse executive functioning reflected in higher levels of plasma GFAP among older adults with and without cognitive impairment. Journal of the International Neuropsychological Society. 2022;28(6):588–599. doi: 10.1017/S1355617721000813 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Sun W, Ye S, Wang Y, et al. Plasma biomarkers for diagnosis and differentiation and their cognitive correlations in patients with Alzheimer’s disease. Brain Communications. 2025;7(2):fcaf094. doi: 10.1093/braincomms/fcaf094 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Mielke MM. Consideration of sex differences in the measurement and interpretation of Alzheimer disease-related biofluid-based biomarkers. The journal of applied laboratory medicine. 2020;5(1):158–169. doi: 10.1373/jalm.2019.030023 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Yakoub Y, Ashton NJ, Strikwerda Brown C, et al. Longitudinal blood biomarker trajectories in preclinical Alzheimer’s disease. Alzheimer’s & Dementia. 2023;19(12):5620–5631. doi: 10.1002/alz.13318 [DOI] [PubMed] [Google Scholar]

[R33] 33.Boccalini C, Peretti DE, Scheffler M, et al. Sex differences in the association of Alzheimer’s disease biomarkers and cognition in a multicenter memory clinic study. Alzheimer’s research & therapy. 2025;17:46. doi: 10.1186/s13195-025-01684-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[R34] 34.Stocker H, Beyer L, Trares K, et al. Association of nonmodifiable risk factors with Alzheimer disease blood biomarkers in community-dwelling adults in the ESTHER study. Neurology. 2025;104(9):e213500. doi: 10.1212/WNL.0000000000213500 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] 35.Milà Alomà M, Hausle I, Myoraku A, et al. Sex differences in Alzheimer’s disease plasma biomarker levels and clinical utility. Alzheimer’s & Dementia. 2026;22(2):e71244. doi: 10.1002/alz.71244 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Foley KE, Winder Z, Sudduth TL, et al. Alzheimer’s disease and inflammatory biomarkers positively correlate in plasma in the UK ADRC cohort. Alzheimer’s & Dementia. 2024;20(2):1374–1386. doi: 10.1002/alz.13485 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Vila-Castelar C, Akinci M, Palpatzis E, et al. Sex/gender effects of glial reactivity on preclinical Alzheimer’s disease pathology. Molecular Psychiatry. 2024:1–10. doi: 10.1038/s41380-024-02753-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Labonte J, Pettway EC, Sugarman M, et al. Sex Differences on Tau, Astrocytic and Neurodegeneration Plasma Biomarkers. Alzheimer’s & Dementia. 2023;19:e071807. doi: 10.1002/alz.071807 [DOI] [Google Scholar]

[R39] 39.Duarte-Guterman P, Albert AY, Inkster AM, Barha CK, Galea LA, Initiative AsDN. Inflammation in Alzheimer’s disease: do sex and APOE matter? Journal of Alzheimer’s Disease. 2020;78(2):627–641. doi: 10.3233/JAD-200982 [DOI] [PubMed] [Google Scholar]

[R40] 40.Bellaver B, Povala G, Ferreira PC, et al. Astrocyte reactivity influences amyloid-β effects on tau pathology in preclinical Alzheimer’s disease. Nature medicine. 2023;29(7):1775–1781. doi: 10.1038/s41591-023-02380-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] 41.Joynes CM, Bilgel M, An Y, et al. Sex differences in the trajectories of plasma biomarkers, brain atrophy, and cognitive decline relative to amyloid onset. Alzheimer’s & Dementia. 2024;doi: 10.1002/alz.14405 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R42] 42.Rohden F, Ferreira PC, Bellaver B, et al. Glial reactivity correlates with synaptic dysfunction across aging and Alzheimer’s disease. Nature Communications. 2025;16(1):5653. doi: 10.1038/s41467-025-60806-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Beckett LA, Donohue MC, Wang C, et al. The Alzheimer’s Disease Neuroimaging Initiative phase 2: Increasing the length, breadth, and depth of our understanding. Alzheimer’s & Dementia. 2015;11(7):823–831. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Reichert C, Sidtis JJ, Pomara N. The Memory Education and Research Initiative. The Preservation of Memory. 2015;6 [Google Scholar]

[R45] 45.Bäckman L, Jones S, Berger A-K, Laukka EJ, Small BJ. Cognitive impairment in preclinical Alzheimer’s disease: a meta-analysis. Neuropsychology. 2005;19(4):520. doi: 10.1037/0894-4105.19.4.520 [DOI] [PubMed] [Google Scholar]

[R46] 46.Pomara N, Bruno D, Plaska CR, et al. Plasma Amyloid-β dynamics in late-life major depression: a longitudinal study. Translational Psychiatry. 2022;12(1):301. doi: 10.1038/s41398-022-02077-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Ashton NJ, Pascoal TA, Karikari TK, et al. Plasma p-tau231: a new biomarker for incipient Alzheimer’s disease pathology. Acta neuropathologica. 2021;141(5):709–724. doi: 10.1007/s00401-021-02275-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Shaw LM, Korecka M, Lee EB, et al. ADNI Biomarker Core: A review of progress since 2004 and future challenges. Alzheimer’s & Dementia. 2025;21(1):e14264. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Doecke JD, Pérez-Grijalba V, Fandos N, et al. Total Aβ42/Aβ40 ratio in plasma predicts amyloid-PET status, independent of clinical AD diagnosis. Neurology. 2020;94(15):e1580–e1591. doi: 10.1212/WNL.0000000000009240 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] 50.Wojdała AL, Bellomo G, Gaetani L, et al. Trajectories of CSF and plasma biomarkers across Alzheimer’s disease continuum: disease staging by NF-L, p-tau181, and GFAP. Neurobiology of Disease. 2023;189:106356. doi: 10.1016/j.nbd.2023.106356 [DOI] [PubMed] [Google Scholar]

[R51] 51.Abbas S, Ferreira PC, Bellaver B, et al. Utility of plasma GFAP as a secondary endpoint for clinical trials in Alzheimer’s disease. The Journal of Prevention of Alzheimer’s Disease. 2025:100205. doi: 10.1016/j.tjpad.2025.100205 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Varma V, An Y, Kac PR, et al. Longitudinal progression of blood biomarkers reveals a key role of astrocyte reactivity in preclinical Alzheimer’s disease. medRxiv. 2024;doi: 10.1101/2024.01.25.24301779 [DOI] [PubMed] [Google Scholar]

[R53] 53.Trabace L, Roviezzo F, Rossi A. Sex differences in Inflammatory diseases. Frontiers in Pharmacology. 2022;13:962869. doi: 10.3389/fphar.2022.962869 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Ramos JJJ, Pupo NMG, Mena D, et al. Cognitive Decline in Chronic Inflammatory Conditions: Exploring Links Between Systemic Inflammation and Neurodegeneration. Cureus. 2025;17(7)doi: 10.7759/cureus.88397 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] 55.Michel M, Fiebich BL, Kuzior H, et al. Increased GFAP concentrations in the cerebrospinal fluid of patients with unipolar depression. Translational psychiatry. 2021;11(1):308. doi: 10.1038/s41398-021-01423-6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R56] 56.Kang MJ, Eratne D, Dean O, et al. Plasma Glial Fibrillary Acidic Protein and Neurofilament Light Are Elevated in Bipolar Depression: Evidence for Neuroprogression and Astrogliosis. Bipolar Disorders. 2025;doi: 10.1111/bdi.70029 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] 57.Chitnis T, Magliozzi R, Abdelhak A, Kuhle J, Leppert D, Bielekova B. Blood and CSF biomarkers for multiple sclerosis: emerging clinical applications. The Lancet Neurology. 2025;doi: 10.1016/S1474-4422(25)00249-2 [DOI] [PubMed] [Google Scholar]

[R58] 58.Passos FRS, Heimfarth L, Monteiro BS, et al. Oxidative stress and inflammatory markers in patients with COVID-19: Potential role of RAGE, HMGB1, GFAP and COX-2 in disease severity. International Immunopharmacology. 2022;104:108502. doi: 10.1016/j.intimp.2021.108502 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R59] 59.Stevenson-Hoare J, Heslegrave A, Leonenko G, et al. Plasma biomarkers and genetics in the diagnosis and prediction of Alzheimer’s disease. Brain. 2023;146(2):690–699. doi: 10.1093/brain/awac128 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] 60.Labonte J, Sugarman MA, Pettway E, et al. Sex differences on tau, astrocytic, and neurodegenerative plasma biomarkers. Journal of Alzheimer’s Disease. 2025;105(2):443–452. doi: 10.1177/13872877251329468 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] 61.Mila-Aloma M, Salvado G, Gispert J, et al. Amyloid beta, tau, synaptic, neurodegeneration, and glial biomarkers in the preclinical stage of the Alzheimer’s continuum. Alzheimer’s Dementia 2020;16(10):1358–1371. doi: 10.1002/alz.12131 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] 62.Wang X, Ye T, Dai B, Zhang J, Zhou W, Initiative AsDN. Sex-specific patterns in tau spreading throughout the Braak stages in the Alzheimer’s disease spectrum. Journal of Alzheimer’s Disease. 2026:13872877251406131. doi: 10.1177/13872877251406131 [DOI] [PubMed] [Google Scholar]

[R63] 63.Milà-Alomà M, Van Hulle C, Brugulat-Serrat A, et al. Sex differences in Alzheimer’s disease CSF biomarkers and their association with Aβ pathology on PET in cognitively unimpaired individuals. Alzheimer’s Research & Therapy. 2025;17(1):235. doi: 10.1186/s13195-025-01844-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] 64.Graff-Radford NR, Crook JE, Lucas J, et al. Association of low plasma Aβ42/Aβ40 ratios with increased imminent risk for mild cognitive impairment and Alzheimer disease. Archives of neurology. 2007;64(3):354–362. doi: 10.1001/archneur.64.3.354 [DOI] [PubMed] [Google Scholar]

[R65] 65.Aarsland D, Sunde AL, Tovar-Rios DA, et al. Prevalence of Alzheimer’s disease pathology in the community. Nature. 2025:1–5. doi: 10.1038/s41586-025-09841-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[R66] 66.Wu CY, Chen L, Fatima H, et al. Combined use of plasma p tau217, NfL, and GFAP predicts domain specific cognitive decline in cognitively unimpaired and MCI individuals. Alzheimer’s & Dementia. 2025;21(12):e70934. doi: 10.1002/alz.70934 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] 67.Julian R, Fröhlich S, Müller K, Dammhahn M, Voelcker-Rehage C. Sex differences in cognitive performance persist into your 80s. GeroScience. 2025:1–14. doi: 10.1007/s11357-025-01585-x [DOI] [PMC free article] [PubMed] [Google Scholar]

[R68] 68.Liu M, Bhatt K, Giaever MF, et al. Sex differences in cognitive performance in Alzheimer’s disease: Insights from the ADAS-Cog-13. Journal of Alzheimer’s Disease. 2025:13872877261440125. [DOI] [PubMed] [Google Scholar]

[R69] 69.Gonzales MM, Wiedner C, Wang CP, et al. A population based meta analysis of circulating GFAP for cognition and dementia risk. Annals of Clinical and Translational Neurology. 2022;9(10):1574–1585. doi: 10.1002/acn3.51652 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R70] 70.Rajan KB, Aggarwal NT, McAninch EA, et al. Remote blood biomarkers of longitudinal cognitive outcomes in a population study. Annals of neurology. 2020;88(6):1065–1076. doi: 10.1002/ana.25874 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R71] 71.Gonzales MM, Short MI, Satizabal CL, et al. Blood biomarkers for dementia in Hispanic and non Hispanic White adults. Alzheimer’s & Dementia: Translational Research & Clinical Interventions. 2021;7(1):e12164. doi: 10.1002/trc2.12164 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R72] 72.Price BR, Walker KA, Eissman JM, et al. Sex differences and the role of estrogens in the immunological underpinnings of Alzheimer’s disease. Alzheimer’s & Dementia: Translational Research & Clinical Interventions. 2025;11(3):e70139. doi: 10.1002/trc2.70139 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R73] 73.Dimayuga FO, Reed JL, Carnero GA, et al. Estrogen and brain inflammation: effects on microglial expression of MHC, costimulatory molecules and cytokines. Journal of neuroimmunology. 2005;161(1–2):123–136. doi: 10.1016/j.jneuroim.2004.12.016 [DOI] [PubMed] [Google Scholar]

[R74] 74.Rettberg JR, Yao J, Brinton RD. Estrogen: a master regulator of bioenergetic systems in the brain and body. Frontiers in neuroendocrinology. 2014;35(1):8–30. doi: 10.1016/j.yfrne.2013.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R75] 75.Medway C, Combarros O, Cortina-Borja M, et al. The sex-specific associations of the aromatase gene with Alzheimer’s disease and its interaction with IL10 in the Epistasis Project. European Journal of Human Genetics. 2014;22(2):216–220. doi: 10.1038/ejhg.2013.116 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R76] 76.Scharfman HE, MacLusky NJ. Estrogen and brain-derived neurotrophic factor (BDNF) in hippocampus: complexity of steroid hormone-growth factor interactions in the adult CNS. Frontiers in neuroendocrinology. 2006;27(4):415–435. doi: 10.1016/j.yfrne.2006.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R77] 77.Coughlan GT, Betthauser TJ, Boyle R, et al. Association of age at menopause and hormone therapy use with tau and β-amyloid positron emission tomography. JAMA neurology. 2023;80(5):462–473. doi: 10.1001/jamaneurol.2023.0455 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R78] 78.Shumaker SA, Legault C, Rapp SR, et al. Estrogen plus progestin and the incidence of dementia and mild cognitive impairment in postmenopausal women: the Women’s Health Initiative Memory Study: a randomized controlled trial. Jama. 2003;289(20):2651–2662. doi: 10.1001/jama.289.20.2651 [DOI] [PubMed] [Google Scholar]

[R79] 79.Shadyab AH, Zhang B, LaCroix AZ, et al. Plasma Phosphorylated Tau 217 and Incident Mild Cognitive Impairment and Dementia in Older Women. JAMA Network Open. 2026;9(3):e261295. doi: 10.1001/jamanetworkopen.2026.1295 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R80] 80.Hanamsagar R, Bilbo SD. Sex differences in neurodevelopmental and neurodegenerative disorders: focus on microglial function and neuroinflammation during development. The Journal of steroid biochemistry and molecular biology. 2016;160:127–133. doi: 10.1016/j.jsbmb.2015.09.039 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R81] 81.Zárate S, Stevnsner T, Gredilla R. Role of estrogen and other sex hormones in brain aging. Neuroprotection and DNA repair. Frontiers in aging neuroscience. 2017;9:322754. doi: 10.3389/fnagi.2017.00430 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R82] 82.Yoon B, Lee H-j, Chae H, et al. Plasma glial fibrillary acidic protein as a potential biomarker for differentiating amyloid-negative subjective cognitive decline and mild cognitive impairment. Journal of Alzheimer’s Disease. 2025:13872877261427806. doi: 10.1177/13872877261427806 [DOI] [PubMed] [Google Scholar]

[R83] 83.Pais MV, Forlenza OV, Diniz BS. Plasma biomarkers of Alzheimer’s disease: a review of available assays, recent developments, and implications for clinical practice. Journal of Alzheimer’s disease reports. 2023;7(1):355–380. doi: 10.3233/ADR-230029 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R84] 84.Simrén J, Weninger H, Brum WS, et al. Differences between blood and cerebrospinal fluid glial fibrillary acidic protein levels: the effect of sample stability. Alzheimer’s & Dementia. 2022;18(10):1988–1992. doi: 10.1002/alz.12806 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R85] 85.Youn W, Yun M, Lee CJ, Schöll M. Cautions on utilizing plasma GFAP level as a biomarker for reactive astrocytes in neurodegenerative diseases. Molecular Neurodegeneration. 2025;20(1):54. doi: 10.1186/s13024-025-00846-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R86] 86.Snellman A, Ekblad LL, Ashton NJ, et al. Head-to-head comparison of plasma p-tau181, p-tau231 and glial fibrillary acidic protein in clinically unimpaired elderly with three levels of APOE4-related risk for Alzheimer’s disease. Neurobiology of Disease. 2023;183:106175. doi: 10.1016/j.nbd.2023.106175 [DOI] [PubMed] [Google Scholar]

[R87] 87.David J-P, Ghozali F, Fallet-Bianco C, et al. Glial reaction in the hippocampal formation is highly correlated with aging in human brain. Neuroscience letters. 1997;235(1–2):53–56. doi: 10.1016/S0304-3940(97)00708-8 [DOI] [PubMed] [Google Scholar]

[R88] 88.Jack CR Jr, Wiste HJ, Algeciras-Schimnich A, et al. Predicting amyloid PET and tau PET stages with plasma biomarkers. Brain. 2023;146(5):2029–2044. doi: 10.1093/brain/awad042 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R89] 89.Bluma M, Chiotis K, Bucci M, et al. Disentangling relationships between Alzheimer’s disease plasma biomarkers and established biomarkers in patients of tertiary memory clinics. EBioMedicine. 2025;112doi: 10.1016/j.ebiom.2024.105504 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R90] 90.Young VM, Wiedner C, Baril AA, et al. Sex differences in the association between sleep duration and blood based biomarkers of Alzheimer’s disease and neurodegeneration: The Framingham Heart Study. Alzheimer’s & Dementia. 2025;21:e099558. doi: 10.1002/alz70856_099558 [DOI] [Google Scholar]

[R91] 91.Van Egroo M, Beckers E, Ashton NJ, Blennow K, Zetterberg H, Jacobs HI. Sex differences in the relationships between 24-h rest-activity patterns and plasma markers of Alzheimer’s disease pathology. Alzheimer’s research & therapy. 2024;16(1):277. doi: 10.1186/s13195-024-01653-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[R92] 92.Balachandar R, Viramgami A, Singh DP, et al. Association between chronic PM2. 5 exposure and neurodegenerative biomarkers in adults from critically polluted area. BMC Public Health. 2025;25(1):1413. doi: 10.1186/s12889-025-22641-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R93] 93.Ferrari-Souza JP, Povala G, Rahmouni N, et al. Microglia modulate Aβ-dependent astrocyte reactivity in Alzheimer’s disease. Nature Neuroscience. 2025:1–7. doi: 10.1038/s41593-025-02103-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

[R94] 94.Pelkmans W, Shekari M, Brugulat Serrat A, et al. Astrocyte biomarkers GFAP and YKL 40 mediate early Alzheimer’s disease progression. Alzheimer’s & Dementia. 2024;20(1):483–493. doi: 10.1002/alz.13450 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

This is a preprint.

Sex-specific associations between astrocytic reactivity and cognitive decline in unimpaired elderly

Chelsea Reichert Plaska

Tovia Jacobs

Davide Bruno

Sang Han Lee

Bruno P Imbimbo

Ricardo S Osorio

Andréa L Benedet

Nicholas J Ashton

Henrik Zetterberg

Nunzio Pomara

Abstract

INTRODUCTION:

METHODS:

RESULTS:

DISCUSSION:

1. INTRODUCTION

2. METHODS

2.1. Population

2.1.1. MERI Cohort

Figure 1. Participant flow diagram.

2.1.2. ADNI Cohort

2.2. Study Design

2.3. Blood Biomarkers

2.3.1. MERI Blood Biomarkers

2.3.1. ADNI Blood Biomarkers

2.4. Statistical Analysis

3. RESULTS

3.1. Baseline Characteristics

3.1.1. MERI Cohort

Table 1.

Figure 2. Boxplots of Blood-based Biomarkers (BBMs).

3.1.2. ADNI Cohort

3.2. Associations Between Baseline GFAP and BBMS

3.2.1. Baseline associations – MERI Cohort

3.2.2. Baseline associations – ADNI Cohort

3.2.3. Longitudinal outcomes – MERI Cohort

Table 2.

Figure 3. Longitudinal Aβ42/40 trajectories of participants in the lower vs. upper median GFAP at baseline.

3.2.4. Longitudinal outcomes – ADNI Cohort

3.3. Associations Between Baseline GFAP and Cognition

3.3.1. Baseline associations – MERI Cohort

3.3.2. Baseline associations – ADNI Cohort

3.3.3. Longitudinal outcomes – MERI Cohort

Figure 4. Longitudinal MMSE score trajectories of female participants in the lower vs. upper median GFAP at baseline for the MERI Cohort.

Table 3.

Table 4.

3.3.4. Longitudinal outcomes – ADNI Cohort

3.4. Longitudinal GFAP and Longitudinal BBM and Cognition analyses – MERI Cohort

3.5. Rates of Change in GFAP and Longitudinal BBM analyses

3.5.1. MERI Cohort

3.5.2. ADNI Cohort

4. DISCUSSION

4.1. Limitations

5. CONCLUSIONS

Supplementary Material

ACKNOWLEDGEMENTS

FUDNING SOURCES

Footnotes

Data Sharing and Code Availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases