Published on September 9, 2026 in The Lancet Neurologythis work carried out by Columbia shows in particular that the risk does not play out in the same way depending on the genetic profile. When p-tau217 increases, people with APOE4 could develop symptoms three to four years later, compared to five to six years for other profiles.
What if Alzheimer’s could be detected before the first forgetfulness occurs?
This is one of the great challenges of Alzheimer’s research: the disease can begin to evolve in the brain. well before memory problems become visible.
Researchers from Columbia University Irving Medical Centerled by Richard Mayeux, wanted to go further: not only identify people at risk, but try to determine when the first symptoms might appear.
They analyzed data from 8,582 adultsfrom seven cohorts in North America and the Caribbean, ranging from people without cognitive impairment to patients with dementia.
The researchers crossed two pieces of information: the blood level of
p-tau217biomarker associated with Alzheimer’s pathological processes, and gene status APOE – the main genetic risk factor for Alzheimer’s.
Taken alone, a high level of p-tau217 signals an increased risk, but does not accurately date the onset of symptoms. The addition of APOE4, however, changes the situation.
Three years or six years: the genetic profile makes the difference
When p-tau217 becomes elevated in a person who is still asymptomatic, researchers observe a clear difference according to the APOE profile. People who own one or two copies of APOE4are likely to develop symptoms in approximately three to four years. In people with other APOE variants, the window is more five to six years.
“The combination of tests really makes a difference in terms of predictive power. And the projections are the same for everyone, whatever their origin.”explains Richard Mayeux, head of the department of neurology at Columbia University and neurologist-in-chief at NewYork-Presbyterian/Columbia University Irving Medical Center.
This difference is important: the same biological signal therefore does not seem to tell exactly the same story depending on the person’s genetic profile. But we must be careful not to interpret it too literally. These figures do not give a date of onset of illness for each individual. They correspond to probabilities observed in a population studied.
A high level of p-tau217 associated with APOE4 therefore does not mean that a person will necessarily develop Alzheimer’s in four years.
For now, it’s better not to run for the test
The prospect is attractive. She also poses a dizzying question: What would we do with such information when there is still no preventive treatment for people without symptoms?
Richard Mayeux himself calls for caution.
“I do not recommend that people take these tests today if they are asymptomatic“he warns. Before continuing: “What would you do with this information?”
Today, these tests remain essentially confined to research and certain specialized centers. The currently authorized monoclonal antibodies against Alzheimer’s are intended for people already presenting symptoms and are not offered to asymptomatic people simply identified as at risk.
However, the situation could change quickly. An ongoing study is seeking to determine whether these treatments can slow the progression of the disease when administered
before symptoms appearin people with high levels of p-tau217.
This is precisely where the new method could take on its full value.
The real issue: knowing when to act
For researchers, the goal is not just to predict the future. It is above all to be able to intervene when preventive intervention would be most likely to be useful.
““What this will allow us to do is make better predictions about when symptoms will appear in people at risk, and, when preventative medications become available, prescribe them at the right time.”explains Richard Mayeux.
He summarizes the hypothesis that could guide future treatments as follows: “If you have a drug that could prevent disease, what would be the optimal time to give it to people? Our data indicates that for high-risk people with the APOE4 gene, the best time is when their p-tau levels rise, approximately three years before symptom onset.“. The discovery therefore does not yet make it possible to predict with certainty a person’s future. But it opens a new perspective: that of a medicine which could, tomorrow, no longer wait for the first omissions to try to act.
For families who live with the fear of Alzheimer’s, these few years gained could then take on another dimension. No longer just as a delay to be expected, but as a possible window to prevent, monitor and, perhaps, preserve autonomy for longer.