Cognitive tests may not tell the whole story about memory concerns

By Dr. Liji Thomas, MDReviewed by Lauren HardakerOct 1 2026

People with subjective cognitive decline showed a high burden of modifiable dementia risks, while elevated p-Tau217 was linked to early signs of neurodegeneration, and sharing biomarker results may encourage healthier lifestyle changes.

Study: Dementia risk factors and biomarkers in subjective cognitive decline: real world evidence from the monza brain health service. Image credit: pathdoc/Shutterstock.com

A recent paper published in the journal Neurological Sciences reported a high prevalence of modifiable risk factors among individuals with subjective cognitive decline (SCD). It also revealed signs of neurodegeneration among SCD individuals with suprathreshold levels of the dementia biomarker phosphorylated Tau217 (p-Tau217). Finally, structured disclosure of biomarker results and risk estimates was well accepted, and about 60% of participants who underwent disclosure reported greater motivation for lifestyle changes.

Why subjective cognitive decline matters for dementia prevention

Up to 45% of dementia cases could potentially be prevented or delayed by addressing 14 modifiable risk factors identified by the Lancet Commission, putting growing emphasis on prevention among people who may be at increased risk of cognitive decline.

One group attracting particular attention is people with subjective cognitive decline (SCD), who notice changes in their cognition even though they continue to perform normally on standard cognitive tests. SCD is far more common than mild cognitive impairment (MCI) and can, in some people, represent an early stage on the path toward dementia. Depending on the setting, previous research suggests that up to one-third of people with SCD may harbor preclinical Alzheimer’s disease, although SCD is highly heterogeneous and can also have other neurodegenerative or non-neurodegenerative causes.

Advances in blood-based biomarkers are now raising difficult questions about how to identify and communicate that risk. Under the 2024 Alzheimer’s Association (AA) criteria, Alzheimer’s disease can be defined biologically using abnormal biomarkers, including plasma markers such as p-Tau217, irrespective of cognitive status, provided they perform adequately in the population being tested. But whether Alzheimer’s should be diagnosed at the SCD stage, and what people should be told about their biomarker results, remains debated.

The AA Workgroup currently recommends against diagnostic biomarker testing in cognitively unimpaired people outside observational or therapeutic research. The International Working Group (IWG) takes a more conservative approach, generally regarding biomarker-positive, cognitively unimpaired people as being at risk of Alzheimer’s rather than as having the disease. However, it allows biomarker testing in people with SCD through non-diagnostic pathways aimed at estimating, communicating, and reducing future risk.

That changing approach has helped drive the emergence of Brain Health Services (BHS) across Europe. Unlike conventional memory clinics, which primarily diagnose and treat cognitive disorders, BHS are designed around dementia prevention, with risk assessment, communication, risk reduction, and cognitive enhancement at their core.

People can enter the BHS through self-referral, primary care, or a memory clinic. They undergo a broad assessment that can include cognitive screening, blood tests for biomarkers such as p-Tau217 and GFAP, APOE genotyping, and brain imaging. Those found to have MCI can be referred back to the memory clinic, while the assessments help clinicians distinguish between SCD, MCI, and functional cognitive decline.

In the current study, the researchers examined 186 consecutive attendees at a fully functioning BHS in Monza, Italy, including 132 people with SCD. They set out to characterize the clinical, biological, and neuroimaging features of this real-world population and explore what happened when participants were given information about their biomarkers and estimated dementia risk.

Blood biomarkers divide SCD participants by Alzheimer’s-related risk

The study included 186 consecutive BHS attendees, of which 132 had SCD. Among SCD participants with available plasma biomarkers, individuals were classified as A+ and A-, depending on whether their plasma p-Tau217 levels were at or above, or below, respectively, a threshold of 0.15 pg/mL.

Cumulative dementia risk was calculated using the Lancet Commission total relative risk (RR), applied as appropriate for the age group.

At a second visit, p-Tau217 and APOE genotype results were communicated, along with the estimated dementia risk. Targeted preventive measures were recommended to help reduce dementia risk. The individualized risk estimate combined the Lancet relative-risk multiplier with baseline risk calculated using the Brookmeyer method and Italian mortality data. One week later, they were assessed online for the impact of the disclosure using self-reported questionnaires.

Cognitive profiles differ across brain health service attendees

The investigators found that SCD, MCI, functional cognitive decline (FCD), and worried-well (WW) groups each had their own cognitive and clinical characteristics, though this is only an exploratory analysis due to the small sample size.

FCD subjects were younger, more often female, showed depressive symptoms, and reported social isolation.

MCI was present in older people with higher levels of amyloid biomarkers, neurodegeneration, and neuroinflammation, with lower cognition.

WW represents people worried about their risk but who had less cognitive impairment and lower biomarker levels.

Among people with SCD, memory was the most commonly reported area of cognitive decline, and up to a third were A+, suggesting an increased probability of underlying amyloid pathology and the possibility of preclinical AD. However, an A+ result in this study should not be interpreted as confirmation of amyloid pathology.

Modifiable dementia risks are common among SCD participants

The investigators found that the most common potentially modifiable risk factors among people with SCD in this cohort were air pollution, impaired sleep, smoking, physical inactivity, and hypertension. Air-pollution exposure was assigned to all participants based on the geographic setting and therefore did not distinguish between study groups, while sleep impairment was an additional putative risk factor. The Lancet RR was correlated with the number of modifiable dementia risk factors.

A+ SCD individuals had a higher median age and performed less well on cognitive screening, although their qMCI scores remained within the normal range. Their Lancet RR was higher than A- subjects, but this difference vanished after adjusting for age and multiple testing.

They also had higher levels of biomarkers, including GFAP and p-Tau181, though this does not confirm amyloid deposition. They also demonstrated more signs of brain atrophy on magnetic resonance imaging (MRI), with differences in global and parietal cortical atrophy remaining significant after adjustment for age. The difference in NfL, a marker of neurodegeneration, did not remain significant after age adjustment and was marginal after correction for multiple comparisons.

These findings support the relevance of SCD as a transitional stage between normal cognition and MCI in the AD continuum.

A notable finding was that almost half of people with A- SCD thought their memory was impaired compared with their peers. However, this was the case with only a fifth of A+ SCD subjects, which the authors suggest could reflect an early erosion of insight in the A+ group, compared with heightened self-awareness in A- individuals.

Multiple biomarkers reveal two distinct SCD risk profiles

The researchers identified two participant clusters, one lower-risk group with moderate stability, characterized by high cognition and relatively lower modifiable and biological risk levels, and the other more stable group with lower cognitive performance and high risk (driven mainly by GFAP, NfL, and p-Tau217).

Most participants welcome information about their dementia risk

Among the 42 SCD participants who underwent disclosure, almost all considered their biomarker and risk disclosure informative and useful and were satisfied with how it was conducted. About 60% reported greater motivation to change their lifestyle.

There were no significant differences between A+ and A- participants in changes in anxiety or depression scores after disclosure. A+ participants showed somewhat greater event-related distress, but the differences did not survive correction for multiple comparisons, and median scores remained well below the threshold for post-traumatic stress disorder. The authors consider this to indicate the potential utility of communicating biological risk to concerned individuals.

However, plasma p-Tau217, APOE genotype, and the individualized dementia-risk estimate were communicated together, so the impact of disclosing each component could not be assessed separately.

Limitations

Formal neuropsychological testing was not systematically performed in all SCD participants. Memory clinic assessment test data were not available for this study, which may have led to misclassification of SCD.

Low sample size limits the strength of the conclusions. The cross-sectional design prevented conclusions about longitudinal progression of subjects in each cluster.

Risk disclosure involved multiple pieces of information, precluding assessment of the impact of each one separately. Disclosure data were available for only 42 of the 84 SCD participants with biomarker data, limiting the validity of the conclusions and making longer-term studies necessary to establish the psychological and behavioral effects.

Blood biomarkers may strengthen personalized dementia prevention

The findings show that SCD is associated with a high prevalence of modifiable risk factors. A+ individuals show signs of early neurodegeneration in brain structure and in biomarker tests.

The study also suggests that SCD individuals with an above-threshold p-Tau217 result may have reduced insight into their memory difficulties. Using multiple markers, the researchers identified subjects with a multidimensional profile of brain vulnerability.

The study also provides preliminary evidence that disclosure of risk biomarker status is feasible and well tolerated in the short term, and may increase motivation for lifestyle change.

“These findings support incorporating plasma biomarkers and structured disclosure protocols in BHS, complementing multimodal prevention strategies to delay dementia onset.”