A blood test might detect the earliest molecular signs of Alzheimer’s disease more than a decade before memory lapses become noticeable, according to a study published in The Lancet on May 28 as part of a special neurology issue.
Researchers from the University of California, San Francisco and collaborating institutions measured Alzheimer’s-associated proteins in the blood of 1,350 dementia-free adults with an average age of 61 — a midlife demographic where detecting the disease’s earliest seeds has proven especially challenging. They found that 6% of participants already harbored biomarker evidence of Alzheimer’s neuropathology, and those individuals were already showing subtle cognitive differences compared with their biomarker-negative peers.
The cohort
The study drew from the Coronary Artery Risk Development in Young Adults (CARDIA) study, a multi-site, community-based longitudinal cohort that began in 1983. Of 2,248 participants who completed the year 35 visit (2020–2022), 1,500 were randomly selected for plasma biomarker measurement. After exclusions, the final sample included 1,350 adults — 58% women, 45% Black, 55% White — all dementia-free and with a mean age of 61 years (SD 3.6).
This design matters. Most Alzheimer’s biomarker research has been conducted in older adults (70+) or in memory clinic populations. A community-based midlife cohort captures the disease process much earlier — and in people who look healthy on the surface.
What the biomarkers showed
The team used the FDA-approved Fujirebio Lumipulse G1200 platform to measure plasma p-tau217, Aβ42, and Aβ40, calculating the p-tau217/Aβ42 ratio — a marker that closely tracks amyloid PET positivity. They found:
| Biomarker Definition | Prevalence in Cohort |
|---|---|
| p-tau217/Aβ42 ratio positive | 86 participants (6%) |
| Aβ42/40 ratio positive | 196 participants (15%) |
| p-tau217 alone positive | 48 participants (4%) |
The 6% amyloid positivity rate at mean age 61 aligns with what population-based amyloid PET studies have estimated for this age range. It confirms that Alzheimer’s pathology begins its long march through the brain decades before the first clinical symptom.
Cognitive differences
The lucid middle-aged participants in the study underwent a battery of five standardized cognitive tests at their year 35 visit, with longitudinal data spanning about five years from the year 30 visit.
At baseline, biomarker-positive individuals scored significantly lower on:
- Processing speed: standardized difference −0.54 to −0.25 across biomarker definitions (p = 0.0001 to 0.0048)
- Executive function: standardized difference −0.42 to −0.19 (p = 0.0070 to 0.049)
No significant association was found for global cognition, verbal memory, or fluency at baseline — the cognitive domains most affected in later-stage Alzheimer’s.
Over the 5-year follow-up, however, the picture sharpened:
- Participants with abnormal Aβ42/40 ratios had 4.31 times higher odds of accelerated verbal memory decline (95% CI 1.71–10.9)
- Those with elevated p-tau217 had 3.98 times higher odds of accelerated processing speed decline (95% CI 1.71–9.3)
- The p-tau217/Aβ42 ratio predicted 3.35 times higher odds of accelerated processing speed decline (95% CI 1.77–6.35)
“These are people in their 50s and 60s who are cognitively normal by any clinical standard,” said Xiaqing Jiang, PhD, the study’s first author, a postdoctoral scholar at UCSF’s Department of Psychiatry and Behavioral Sciences. “But the blood test is picking up something happening in their brains that is already beginning to affect cognitive trajectories.”
The ‘midlife’ advance
What makes this study distinct is the age window. Prior blood biomarker work — notably by Ashton et al. (2024, Nature Medicine) and Palmqvist et al. (2020, JAMA) — demonstrated that p-tau217 can detect amyloid PET pathology with AUCs of 0.92–0.98, but those studies enrolled adults with an average age of 70 or older. The CARDIA study pushes the detection window approximately 15–20 years earlier.
“This extends our understanding of the preclinical phase of Alzheimer’s disease,” said senior author Kristine Yaffe, MD, vice chair of psychiatry and professor of neurology and epidemiology at UCSF. “We’re showing it’s possible to detect the pathology in midlife, in a community-based setting, using a blood test.”
The study also found stronger associations among women, Black participants, and APOE ε4 carriers — subgroups that may benefit most from early screening.
The sober bottom line
The study is not a screening recommendation. In an accompanying commentary, Anna Rosenberg, MD, PhD, and Tiia Ngandu, MD, PhD, of the Finnish Institute for Health and Welfare cautioned that “in populations with low pre-test probability… positive predictive value decreases,” adding that plasma biomarkers are “not suitable for large-scale untargeted screening” at this stage. False positives remain a concern when the condition being screened is rare — and Alzheimer’s pathology at age 60 is still relatively uncommon.
The study also had limitations: biomarkers were measured at a single time point, not longitudinally; CARDIA survivors may differ from the general population; and the test detects Alzheimer’s pathology specifically, not other forms of dementia, which account for 60–70% of all dementia cases.
Still, the implications are clear. The infrastructure for midlife blood-based Alzheimer’s screening — an FDA-approved platform, validated biomarkers, and community-based normative data — is now substantially closer to clinical reality than it was before this study.
Source: Jiang, X. et al. “Alzheimer’s disease neuropathology plasma biomarkers and cognition in midlife: a community-based cohort study.” The Lancet 407(10544):2208–2216 (2026). DOI: 10.1016/S0140-6736(26)00515-5
>
Funding: National Heart, Lung, and Blood Institute; National Institute on Aging (grants R01AG063887, R01AG091431, R35AG071916, K99AG083211); Alzheimer’s Association (AARFD-23-1150636)