Simple blood test maps hidden Alzheimer’s disease changes

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Dec 21 2025

A large community study shows that Alzheimer’s-related brain changes are far more widespread with age than symptoms alone suggest, highlighting both the promise and the complexity of blood-based screening.

Study: Prevalence of Alzheimer’s disease pathology in the community. Image Credit: Antonio Marca / Shutterstock

In a recent study published in the journal Nature, researchers estimated the age- and cognition-stratified prevalence of Alzheimer’s disease neuropathological changes (ADNCs) in adults aged 70 years and older using plasma phosphorylated tau at threonine 217 (pTau217) as a blood-based surrogate marker. The study also examined associations with demographics, apolipoprotein E (APOE) genotype, and kidney function.

Alzheimer’s Disease and the Role of Blood Biomarkers

Dementia represents a major global public health challenge, with Alzheimer’s disease (AD) as its most common cause. Until recently, verification of AD neuropathological changes, amyloid-β plaques, and tau tangles required cerebrospinal fluid analysis or positron emission tomography, which are impractical for population screening. Blood-based biomarkers now enable detection of pTau217, a marker tightly linked to downstream tau pathology following amyloid-β accumulation.

In this study, elevated plasma pTau217 was used to classify individuals as ADNC-positive, intermediate, or ADNC-negative, rather than to directly confirm brain pathology at the individual level. Understanding how ADNC prevalence varies by age, cognitive status, and APOE ε4 carrier status can inform eligibility for disease-modifying therapies, health service planning, and personal decision-making. Additional research is needed to guide follow-up strategies for intermediate biomarker results.

Population Cohort and Cognitive Classification

Investigators analyzed 11,486 plasma samples from the Trøndelag Health (HUNT) population studies in Norway. This included 2,537 participants aged 58–69.9 years from HUNT3 and 8,949 participants aged 70 years and older from HUNT4 70+.

Formal cognitive assessment was conducted only in the HUNT4 70+ cohort using extensive clinical evaluation and Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) consensus. Participants were classified as cognitively normal, having mild cognitive impairment (MCI), or dementia. Participants under 70 years of age were not evaluated for cognitive status.

Biomarker Measurement and Statistical Methods

Plasma pTau217 concentrations were measured using the Single Molecule Array (Simoa) HD-X platform with a validated commercial assay. A two-cut-off strategy categorized individuals as ADNC-negative (less than 0.40 pg/mL), intermediate (0.40 to less than 0.63 pg/mL), or ADNC-positive (0.63 pg/mL or higher), in accordance with Global CEO Initiative recommendations. APOE genotyping identified ε2, ε3, and ε4 alleles.

Inverse probability weighting accounted for participation bias, selection into the biomarker subset, and differential blood sampling. Associations were examined with respect to age, sex, education, cognitive status, and kidney function. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration 2021 equation. Exploratory analyses estimated positive and negative predictive values by integrating age-specific prevalence with sensitivity and specificity estimates from external studies. Eligibility for disease-modifying therapies followed current clinical recommendations.

Age and Cognitive Status Differences in ADNC Prevalence

ADNC prevalence increased markedly with age. Using the ADNC-positive threshold of 0.63 pg/mL or higher, the estimated prevalence was under 8 percent among adults aged 58–69.9 years and approached 65 percent in individuals older than 90 years.

Among adults aged 70 years and older, approximately 10 percent were classified as preclinical AD, defined as cognitively normal with ADNC positivity. An additional 10.4 percent were classified as prodromal AD, defined as MCI with ADNC positivity, and 9.8 percent met criteria for AD dementia.

Within cognitive strata, roughly 60 percent of individuals with dementia were ADNC-positive, compared with 32.6 percent of those with MCI and 23.5 percent of cognitively unimpaired adults. In all groups, increasing age was associated with higher ADNC prevalence.

Genetic, Educational, and Renal Associations

Sex differences in ADNC prevalence were small. Men aged 80–89 years showed slightly higher prevalence than women, driven by preclinical and prodromal stages, while no sex difference was observed for AD dementia.

Education level was inversely associated with ADNC prevalence, particularly at older ages. Individuals with tertiary education exhibited the lowest prevalence, those with primary education the highest, and those with secondary education intermediate levels.

APOE ε4 carrier status was strongly associated with ADNC positivity in a dose-dependent manner. Prevalence increased from 27.1 percent in non-carriers to 46.4 percent in individuals with one ε4 allele and 64.6 percent in those with two ε4 alleles.

Reduced kidney function was associated with higher pTau217 concentrations, particularly below an eGFR of approximately 51 mL/min/1.73 m². After adjustment for demographic and clinical factors, self-reported comorbidities such as cardiovascular disease, diabetes, cancer, and chronic obstructive pulmonary disease were not significantly associated with ADNC positivity.

Clinical Implications and Population-Level Conclusions

Based on current eligibility criteria, approximately 10 percent of the HUNT4 70+ cohort and about 11 percent of the age-weighted population aged 70 years and older may qualify for anti-amyloid disease-modifying therapies. Predictive value analyses showed that positive predictive value increased with age, while negative predictive value declined, reflecting underlying prevalence effects.

Overall, these findings demonstrate pronounced age-related increases in the estimated prevalence of Alzheimer’s disease neuropathological changes, with meaningful variation by cognitive status, genetic risk, education level, and kidney function. Because classification relied on a blood-based surrogate marker rather than direct neuropathological confirmation, results should be interpreted at the population level rather than as individual diagnoses.