An analysis of approximately 1,500 blood proteins identified biomarkers that could predict the risk of developing dementia up to 15 years before diagnosis.
The results, reported today in Natural aging1are a step toward blood tests that can detect Alzheimer’s disease and other forms of dementia at a very early, pre-symptomatic stage – something scientists have been seeking for decades.
Researchers analyzed blood samples from more than 50,000 healthy adults in the UK Biobank, 1,417 of whom developed dementia over a 14-year period.
They found that high blood levels of four proteins – GFAP, NEFL, GDF15 and LTBP2 – were strongly associated with dementia.
“Studies like this are needed if we are to intervene with disease-modifying therapies in the early stages of dementia,” Amanda Heslegrave, a neuroscientist at University College London, said in a statement to UK Science Media Central London.
According to the World Health Organization, more than 55 million people worldwide currently suffer from dementia.
People are often diagnosed only when they notice memory problems or other symptoms. At this point, the disease could have progressed for years. “Once we diagnose it, it’s almost too late,” says Jian-Feng Feng, a study co-author and a computational biologist also at Fudan University in Shanghai, China. “And it’s impossible to reverse the situation.”
Risk of dementia linked to blood protein imbalance in middle age
By examining 1,463 proteins in blood samples from 52,645 people, the authors found that increased levels of GFAP, NEFL, GDF15 and LTBP2 were associated with dementia and Alzheimer’s disease. Blood levels of these proteins in participants who developed dementia were outside normal limits more than ten years before symptoms appeared.
GFAP, a protein that provides structural support to nerve cells called astrocytes, has previously been proposed as a biomarker to diagnose Alzheimer’s disease.2just like GDF15.
Latest study finds that people with high levels of GFAP in their blood are more than twice as likely as people with normal levels to develop dementia, and are nearly three times as likely to develop Alzheimer’s disease. .
The authors used machine learning to design predictive algorithms, combining the four protein biomarkers with demographic factors such as age, gender, education and family history. They trained the model on data from two-thirds of the study participants and tested its performance using data from the remaining 17,549 people.
The model predicted the incidence of three subtypes of dementia, including Alzheimer’s disease, with approximately 90% accuracy, using data from more than a decade before the participants were officially diagnosed.
The authors say their results could be used to develop blood tests to identify people at risk of developing dementia. Other researchers caution that new biomarkers require further validation before being used as clinical screening tools.
The study “needs to be replicated and biomarkers that allow us to not only screen for disease risk but also differentiate between diseases should be a priority,” Heslegrave said.
Gn En gealth