plasma phosphorylated Tau218 predicted dementia in women

 A recent study found that women with high levels of plasma phosphorylated tau217 (p-tau217) had a higher risk of subsequently developing mild cognitive impairment (MCI) or dementia (see dementia tau217 inc risk women JAMA2026 in dropbox, or doi:10.1001/jamanetworkopen.2026.1295)

Details

-- 2766 women were recruited from 39 US clinical sites between 1996 and 1999 into the Women’s Health Initiative Memory Study (WHIMS), a study of community-dwelling cognitively unimpaired women who were post-menopausal, aged 65-79yo, then randomized to either oral conjugated equine estrogens 0.625mg/d alone (with addition of medroxyprogesterone acetate 2.5mg/d if they had intact uterus) vs placebo

-- baseline plasma p-tau217 was measured in 2024

-- mean age 69.9, White 74%/Black 18%/Asian 5%/7% Latina

-- hormone therapy treatment group:

    -- estrogen alone, the placebo group: 569 women (21%)

    -- estrogen alone, the intervention group: 570 women (21%)

    -- estrogen plus progestin, the placebo group: 803 women (29%)

    -- estrogen plus progestin, the intervention group: 824 women (30%)

-- BMI 29, smoking never 57%/past smoker 38%/current smoker 5%

-- education level: high school 9%/high school diploma or GED 22%/some college 37%/college grad 31%

-- diabetes 7%, cardiovasc disease 5%; hypertension 71%

-- physical activity 11.5 MET h/wk, 

-- total cholesterol 234 mg/dL, HDL 54 mg/dL; eGFR (creatinine-based) 84

-- APOE e4 genotype (only measured in White women), based on the rs429358 and rs7412 single nucleotide polymorphisms: carrier in 27%

-- the WHIMS trials were stopped in 2002 and 2004, but annual in-person follow-up continued through 2007, then transitioned to annual telephone-administered cognitive assessment

-- included in the study: all 1334 women with incident MCI or probable dementia plus a subset of 1502 controls without these cognitive issues

-- primary outcome was the combined end point of incident mild cognitive impairment (MCI), based on the Petersen criteria, or probable dementia (based on the standard psych criteria in DSM)

-- secondary outcomes included MCI and dementia examined separately

-- sensitivity analysis was done excluding those with eGFR <60, since chronic kidney disease is associated with elevated plasma Alzheimer's disease (AD) biomarkers, even in the absence of AD neuropathology

-- up to 25 years of follow-up through 2021, median of 14.1 years

Results

-- 1311 participants developed the combined endpoint of MCI or dementia

-- overall, women with higher levels of baseline p-tau217 were more likely to be older, White (lower in Black women), have lower BMI, have never smoked, have lower eGFR, and be APOE ε4 carriers

-- in the fully adjusted model (adjusted for hormone therapy trial group, age, race, ethnicity, education, BMI, smoking status, diabetes, cardiovascular disease, hypertension, physical activity, eGFR, total cholesterol and HDL):

association of p-tau217 with incident mild cognitive impairment:

association of p-tau217 with incident dementia: 

--The association of p-tau217 with dementia and MCI:

    -- larger in magnitude for estrogen plus progestin for incident dementia, HR 4.18 (3.41-5.13) vs placebo, HR 3.07, (2.41-3.91) (P for interaction = .04)

        -- in contrast, the interaction between p-tau217 and estrogen plus progestin was not significant for the combined MCI or dementia outcome or MCI by itself

    -- p-tau217 had larger magnitudes of association with MCI or dementia among White women, HR 2.68 (2.38-3.01) vs Black women, HR 1.33 (1.04-1.70) (P for interaction < .001), as well as dementia in Black women: HR 1.90 (1.24-2.92) vs  White women HR 3.42 (2.99- 3.92), P for interaction = .01

    -- p-tau217 was not associated with MCI among Black women, HR 1.19 (0.94-1.52) vs White women, HR 2.16 (1.90-2.47), P for interaction < .001

    -- associations of p-tau217 with the cognitive outcomes varied by APOE ε4, with larger magnitudes of an association in carriers than noncarriers, an interaction that was significant for the combined MCI or dementia outcome (P for interaction = .02) and, separately, for MCI (P for interaction = .02, but not for dementia when examined separately (P for interaction = .13)

    -- there were also subgroup differences by age, with larger magnitudes of association for those aged older than 70 years vs 70 years or younger for MCI or dementia (P for interaction = .04)

-- sensitivity analysis of excluding women with eGFR<60 revealed similar results

Commentary:

-- the researchers chose to follow plasma p-tau217 since it seemed to be best marker in other studies in detecting Alzheimer's disease (AD) pathology, equivalent to measuring CSF p-tau217

    -- community samples in prior studies have found the plasma p-tau217 association with incident dementia

-- this current study assessed minimal cognitive impairment as well, different from most prior studies, and also assessed the role of hormone replacement therapy, and it had longer follow-up than prior studies. this study also included a more diverse group of women than on the prior studies

--  i am generally hesitant to highlight genetic differences by race, since race is complex (and many people have genetic input from several racial groups) and for many evaluations race is a poor surrogate marker of socioeconomic, lifestyle, stress/discrimination experienced by different racial groups. However, i thought it was appropriate given the role of the APOE e4 genetic polymorphisms, and the quite large cognitive issues comparing White and Black women found. It should be noted in prior studies that directly addressed this racial question "the ε4 allele is associated with a faster rate of decline in working memory for Whites compared to Whites without the ε4 allele, but Blacks with and without ε4 have the same rate of cognitive decline in working memory. The same pattern was observed for semantic memory – the ε4 was related to a faster rate of decline in Whites but not in Blacks.": https://pubmed.ncbi.nlm.nih.gov/23364031/, or doi: 10.1159/000342778

-- overall findings in this study in cognitively unimpaired women:

     -- higher levels of p-tau217 was associated with incident mild cognitive impairment as well as dementia for up to 25 years of follow-up.

     -- higher levels of p-tau217 was associated with the higher the incidence of subsequent dementia in women randomized to estrogen plus progestin (done in women with intact uteri) vs placebo, though the associations did not vary by estrogen alone (in those without uteri) vs placebo

     -- p-tau217 had a larger association with MCI and dementia in White women vs Black women, though there was similar discriminative accuracy for dementia in both racial groups when combining p-tau217 with age

    -- comparing the highest vs lowest quartile of p-tau217, there was a 7-fold higher risk of dementia

-- other studies in the literature that supports the findings in this study:

    --prior observational studies have also found that hormone replacement therapy (HT) seems to be associated with increased tau accumulation, amyloid deposition, and dementia. for example, a study that included 193 cognitively unimpaired women and 99 unimpaired men found that female sex, earlier age at menopause and HT use were associated with higher regional deposition of tau in PET scans in individuals with elevated Amyloid-β compared with male sex, later age at menopause, and HT nonuse. Affected brain regions included the medial and lateral regions of the temporal and occipital lobes. also, late initiation of HT (>5 years following age at menopause) was associated with higher tau PET findings when compared with early initiation: https://jamanetwork.com/journals/jamaneurology/fullarticle/2802791

    -- a study of 198 community-dwelling, dementia-free participants found that plasma p-tau217 found was associated with a 2-fold higher risk of Alzheimer's Disease (AD) in a 4-year study: doi:10.1212/CPJ.0000000000200291

    -- the BioFINDER study of 435 cognitively unimpaired adults similarly found a 2-fold higher risk of all-cause dementia after 4.8 years of follow-up:doi:10.1038/s41467-021-23746-0

    -- A study of 2148 dementia-free adults (including those with MCI) found a 3-fold higher risk of all-cause dementia during the 16-year follow-up in those in the highest vs lowest quartile of baseline plasma p-tau217, with greater risk in women vs men: doi:10.1038/s41591-025-03605-x

    -- overall, these studies support a potential dose-response association of p-tau217 with incident dementia, with markedly higher incidence at the highest quartile.

-- there are a few issues stimulated by this current study, reinforcing the use of p-tau217 determinations:

    -- is there benefit in checking p-tau217 levels in men and women who may have age-related cognitive impairment? would that knowledge help determine the likelihood of AD if p-tau217 were elevated vs the slower increase in cognitive dysfunction with the age-related cognitive decline if the p-tau217 were low? for a recent review of p-tau217 as a reliable blood-based marker of Alzheimer's Disease in men and women: https://www.mdpi.com/2227-9059/12/8/1836

    -- are there interventions that slow the cognitive decline in people, especially since p-tau217 levels rise very early in the course of cognitive decline, either through pharmacologic or nonpharmacologic means? for example, it has been shown that a Mediterranean diet does seem to improve cognitive function in patients at high vascular risk of cognitive decline, especially by the use of extra-virgin olive oil. In addition, high adherence to a Mediterranean diet was associated with higher total brain volume in both gray and white matter, especially if there was higher fish and lower meat consumption: https://gmodestmedblogs.blogspot.com/2015/11/mediterranean-diet-and-brain-volume.html

    -- but, is testing for p-tau217 too much information for many patients to handle? will it lead to depression if high? perhaps leading to faster decline? loss of self-esteem? decreased quality of life?

Limitations:

-- there are a large array of risk factors associated with chronic inflammation, and chronic inflammation is associated with essentially all of the chronic diseases we see (including hypertension, visceral obesity, diabetes, cardiovascular disease, cancer, dyslipidemia... ), as well as stress (https://gmodestmedblogs.blogspot.com/2023/04/stress-decreases-cognitive-function.html), discrimination, depression, air pollution, microplastics, sleep deprivation.... https://gmodestmedblogs.blogspot.com/2026/02/racial-disparities-stress-and-mortality.html

-- and there was no granular information on diet, important since there are pretty clear benefits from the Mediterranean diet, as noted above; but ultraprocessed foods, red and processed meats (https://pmc.ncbi.nlm.nih.gov/articles/PMC9194089/), and sugar or some artificial sweeteners (especially sucralose and aspartame) are associated with chronic inflammation. And a sedentary life is associated with chronic inflammation/exercise decreases inflammation

-- in this study, some of the cognitive information on individuals was based on telephone-based assessment, which may be inaccurate

-- they excluded women with creatinine-based eGFR<60. this raises a few issues:

    -- as mentioned in many of my study critiques, cystatin C-based calculations of eGFR are more predictive of bad clinical outcomes than creatinine-based ones: https://gmodestmedblogs.blogspot.com/2023/12/cystatin-c-better-predictor-of-bad.html

    -- though there is an association of lower eGFR with elevated plasma Alzheimer's disease (AD) biomarkers, their sensitivity analysis above did not find different outcomes. but should we be assessing p-tau217 plasma levels in younger individuals who are less likely to have elevated eGFR levels?? there is some information that high p-tau217 in younger adults is predictive of later AD: https://pmc.ncbi.nlm.nih.gov/articles/PMC12782097/

so, there is a large and increasing medical literature highlighting the validity of plasma p-tau217 as a surrogate marker for the subsequent development of mild cognitive impairment and Alzheimer's dementia.

-- although a high result of the p-tau217 might lead people to enjoy their current life more in anticipation of increasing dementia, it might also lead to decreased quality of life and faster decline from feeling hopeless/depressed

-- of course, knowing a high p-tau217 would be particularly advantageous if we had a clear remedy to minimize the risk of dementia. But we can work with patients to improve their diet, exercise, and improve their chronic diseases

    -- meds that decrease systemic inflammation include an array of hypertension meds (spironolactone, ARBs/ACE-i, calcium channel blockers, torsemide) and some meds increase it (thiazides, chlorthalidone, furosemide: https://gmodestmedblogs.blogspot.com/2025/09/resistant-hypertension-are-diuretics.html ). perhaps we should be avoiding the latter?? And torsemide 5mg has similar diuretic and blood pressure effects vs thiazides/chlorthalidone; and torsemide is likely a better med overall than furosemide: https://gmodestmedblogs.blogspot.com/2025/09/resistant-hypertension-are-diuretics.html

-- the Mediterranean studies have shown some potential to help with dementia, but that would need to be tested in a group of patients with early mild cognitive impairment and followed for several years (from some of the studies cited above, perhaps even a 4-year study would be beneficial)

    -- alcohol is a clear inducer of Alzheimer's dementia (https://www.nature.com/articles/s41398-025-03619-6), and there is an argument in cohort, case-control studies and Mendelian randomization that even small amounts of alcohol is too much

    -- which means that the red wine in the Mediterranean diet should be revisited....

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

to get access to the dropbox: https://www.dropbox.com/scl/fo/pcf4vymtpo77f49wd2ljr/ANyJqB40CuBd-DMEr67XDoI?rlkey=5dsalmuhf8ekd0svsgvn93xm1&dl=0

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them