A Blood Test Predicted Dementia Risk in Women Up to 25 Years Before Symptoms Appeared

University of California San Diego

Twenty-five years before a woman notices she cannot remember where she parked the car, a protein in her blood may already be signaling trouble. That is the central finding of a new study from UC San Diego, which shows that a single blood test could identify women at elevated risk for dementia decades before symptoms appear.

The biomarker: phosphorylated tau 217

The protein in question is phosphorylated tau 217 (p-tau217), a modified form of the tau protein that reflects early brain changes associated with Alzheimer's disease. Tau proteins normally stabilize the internal scaffolding of neurons. When they become excessively phosphorylated, they detach, clump together, and form the neurofibrillary tangles that are a hallmark of Alzheimer's pathology. The phosphorylated form leaks into the bloodstream, where it can be measured with a simple blood draw.

What makes p-tau217 particularly interesting is that it rises early in the disease process, potentially years or decades before tangles become dense enough to cause cognitive symptoms. That long lead time is both the promise and the challenge of this biomarker.

A 25-year window into future risk

The study, published in JAMA Network Open, drew on data from 2,766 participants in the Women's Health Initiative Memory Study, a large national study that enrolled women ages 65 to 79 in the late 1990s. All participants were cognitively unimpaired at enrollment. Blood samples collected at baseline were stored and analyzed years later for p-tau217 levels.

Over up to 25 years of follow-up, researchers tracked who developed mild cognitive impairment or dementia. The results showed a clear dose-response relationship: as baseline p-tau217 levels increased, so did the likelihood of later dementia. Women in the highest p-tau217 group faced the greatest long-term risk.

"That kind of long lead time opens the door to earlier prevention strategies and more targeted monitoring," said first author Aladdin Shadyab, an associate professor of public health and medicine at UC San Diego.

Risk differed across subgroups

The prediction was not equally strong for everyone. Higher p-tau217 levels were more strongly associated with cognitive decline among women over age 70 at baseline compared to those younger than 70. The association was also stronger in women carrying the APOE e4 genetic variant, the best-known genetic risk factor for Alzheimer's disease.

An intriguing finding emerged regarding hormone therapy. Women who had been randomized to estrogen plus progestin therapy in the original Women's Health Initiative trial showed a stronger predictive association between p-tau217 and dementia compared to those on placebo. The mechanism behind this interaction remains unclear and will require further investigation.

The strength of the association also differed between white and Black women. However, when researchers combined p-tau217 with age, the prediction accuracy improved similarly across both racial groups, suggesting that a composite approach could provide more equitable screening.

Blood tests versus brain scans and spinal taps

Currently, the most reliable ways to detect Alzheimer's-related brain changes before symptoms appear involve either PET brain imaging, which can visualize amyloid plaques and tau tangles directly, or cerebrospinal fluid analysis obtained through a lumbar puncture. Both are invasive, expensive, and difficult to scale to population-level screening.

A blood test is fundamentally different in accessibility. "Blood-based biomarkers like p-tau217 are especially promising because they are far less invasive and potentially more accessible than brain imaging or spinal fluid tests," said senior author Linda McEvoy, professor emeritus at UC San Diego and senior investigator at Kaiser Permanente Washington Health Research Institute.

If validated for clinical use, a p-tau217 blood test could theoretically be incorporated into routine health screenings, flagging individuals for closer monitoring or enrollment in prevention trials years before they would otherwise come to clinical attention.

What this test cannot yet do

Currently, blood-based biomarkers like p-tau217 are not recommended for clinical use in people without cognitive symptoms. Several important gaps remain before that changes.

First, the study was conducted exclusively in women, reflecting the Women's Health Initiative's enrollment criteria. Whether p-tau217 has the same predictive value in men is an open question. Second, the study is observational and associational. A high p-tau217 level predicted increased risk, but it did not determine whether early identification can actually change outcomes. Without effective prevention or early treatment strategies, knowing someone's risk 25 years in advance may cause anxiety without offering actionable benefit.

Third, the practical clinical utility of any cutoff value remains undefined. What level of p-tau217 should prompt action, and what action should be taken? These questions require prospective interventional studies that test whether early identification followed by specific interventions, whether pharmacological, behavioral, or both, actually delays or prevents dementia.

The study's demographic composition, while diverse by the standards of Alzheimer's research, may not fully represent the broader population. And the blood samples were analyzed with contemporary assay technology applied to stored specimens, introducing potential variability compared to fresh sample analysis.

The goal beyond prediction

"Ultimately, the goal is not just prediction, but using that knowledge to delay or prevent dementia altogether," Shadyab said. That goal remains aspirational. But a validated, accessible blood biomarker with a 25-year lead time would represent a substantial tool for researchers designing prevention trials and, eventually, for clinicians managing patients' long-term cognitive health.