Blood biomarkers like tau217, NfL, and GFAP may predict dementia up to 10 years early, but aren’t yet reliable for widespread screening, a Karolinska Institutet study finds.
A new study published in Nature Medicine has examined the potential of specific biomarkers, tau217, Neurofilament Light (NfL), and Glial Fibrillary Acidic Protein (GFAP), to predict the onset of dementia, including Alzheimer’s disease, up to ten years before clinical diagnosis in cognitively healthy older adults living in the community.
While earlier research has indicated that these biomarkers may aid in early detection of dementia, most prior studies have focused on individuals already experiencing cognitive symptoms or seeking medical attention for memory concerns.
To better understand the predictive value of these biomarkers in the general population, a larger, community-based study was needed and this new research aims to fill that gap.
Large-Scale Study on Cognitively Healthy Adults
Led by researchers from the Aging Research Center of Karolinska Institutet in collaboration with SciLifeLab and KTH Royal Institute of Technology in Stockholm, the study analyzed blood biomarkers in more than 2,100 adults aged 60+, who were followed over time to determine if they developed dementia.
At a follow-up ten years later, 17 percent of participants had developed dementia. The accuracy of the biomarkers used in the study was found to be up to 83 percent.
“This is an encouraging result, especially considering the 10-year predictive window between testing and diagnosis. It shows that it is possible to reliably identify individuals who develop dementia and those who will remain healthy,” says Giulia Grande, assistant professor at the Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, and first author of the study.
Insights Into Risk and Reassurance
“Our findings imply that if an individual has low levels of these biomarkers, their risk of developing dementia over the next decade is minimal,” explains Davide Vetrano, associate professor at the same department and the study’s senior author. “This information could offer reassurance to individuals worried about their cognitive health, as it potentially rules out the future development of dementia.”
However, the researchers also observed that these biomarkers had low positive predictive values, meaning elevated biomarker levels alone could not reliably identify individuals who would surely develop dementia within the next ten years. Therefore, the study authors advise against the widespread use of these biomarkers as screening tools in the population at this stage.
“These biomarkers are promising, but they are currently not suitable as standalone screening tests to identify dementia risk in the general population,” says Davide Vetrano.
The researchers also noted that a combination of the three most relevant biomarkers – p-tau217 with NfL or GFAP – could improve predictive accuracy.
“Further research is needed to determine how these biomarkers can be effectively used in real-world settings, especially for elderly living in the community or in primary health care services,” says Grande.
“We need to move a step further and see whether the combination of these biomarkers with other clinical, biological, or functional information could improve the possibility of these biomarkers to be used as screening tools for the general population,” Grande continues.
Reference: “Blood-based biomarkers of Alzheimer’s disease and incident dementia in the community” by Giulia Grande, Martina Valletta, Debora Rizzuto, Xin Xia, Chengxuan Qiu, Nicola Orsini, Matilda Dale, Sarah Andersson, Claudia Fredolini, Bengt Winblad, Erika J. Laukka, Laura Fratiglioni and Davide L. Vetrano, 26 March 2025, Nature Medicine.
DOI: 10.1038/s41591-025-03605-x
The study was mainly funded by the Swedish Research Council, The Swedish Brain Foundation and The Strategic Research Area in Epidemiology and Biostatistics at Karolinska Institutet. The researchers declare that there are no conflicts of interest.
Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.