Columbia Study Links Blood Test and Gene Marker to Timing of Alzheimer's Symptoms

Researchers at Columbia University say they've found a way to estimate not just whether someone is at risk for Alzheimer's disease, but roughly when symptoms might start showing up.

The study, published in The Lancet Neurology on September 9, combined two existing tests: a blood test that measures phosphorylated-tau 217 (p-tau217), a protein tied to Alzheimer's-related brain changes, and a genetic test for the APOE4 gene variant, the strongest known inherited risk factor for the disease.

Newsweek reported the researchers reviewed clinical data from almost 9,000 people and found that those with both elevated p-tau217 and the APOE4 gene had a 24 percent higher risk of progressing to Alzheimer's, compared to 13 percent for people without the gene. Trade publication CLP reported the same Columbia University Irving Medical Center research drawing on data from approximately 8,500 participants, framing the finding differently: once p-tau217 levels rise, APOE4 carriers typically develop symptoms within three to four years, while non-carriers take five to six years. Both figures trace back to the same Columbia study, and neither outlet flagged the discrepancy in headcount.

Richard Mayeux, chair of neurology at Columbia University Vagelos College of Physicians and Surgeons, said in a Columbia release carried by CLP that pairing the two tests sharpens the prediction significantly. "The combination of the tests really makes a difference in predictive power. And the projections are the same for everyone regardless of their background," Mayeux said.

Why Doctors Still Say Wait

Mayeux raised a significant limitation: he doesn't want asymptomatic people running out to get tested. "I don't recommend that people get these tests right now if they're asymptomatic," he said, according to CLP. "What would you do with that information?"

The monoclonal antibody drugs currently approved for Alzheimer's are only indicated for people who already have symptoms. Someone who tests positive for both markers but feels fine has no approved drug to take yet. CLP reported a clinical trial is underway testing whether those antibody drugs can slow the disease if given to asymptomatic people with elevated p-tau217, but that trial hasn't produced results establishing the approach works.

Ebrahim Zandi, a professor of immunology and immune therapeutics at the University of Southern California who was not involved in the study, told Newsweek the research is meaningful but shouldn't be oversold. "This is an important step forward because it shows that a blood marker of Alzheimer's-related pathology, p-tau217, contains information not only about risk but potentially about how quickly cognitive impairment may develop," Zandi said. He added a specific caution: "I would be cautious about calling this an Alzheimer's 'clock' — the study predicts clinical cognitive impairment, not the exact timing of Alzheimer's disease symptoms in an individual."

Other Biomarker Research Points the Same Direction

Columbia's work isn't happening in isolation. Researchers at the University of Miami Miller School of Medicine, in a study reported by the school's InventUM outlet, found a related blood marker, p-tau181, was linked to worse memory performance six years later in a study of 1,170 older adults, even in people who showed no memory problems at the time their blood was drawn. Lead researchers James Galvin and Deirdre O'Shea were careful to note this is a population-level pattern, not a test that can tell an individual person whether they'll develop dementia.

Separately, a team at Harvard Medical School published a study August 26, 2026, in the journal Neurology examining early-onset Alzheimer's, which strikes people before age 65. Lead author Alexandra Touroutoglou developed an MRI-based biomarker that measures gray-matter shrinkage in eight brain regions tied to memory and reasoning. Among 130 people living with mild cognitive impairment, tracked for an average of two years, about 65 percent progressed to dementia, and the study found each one-standard-deviation increase in brain shrinkage raised the risk of that progression by 1.24 times, according to News Medical's report on the findings. Touroutoglou said the tool functions "as a timer for predicting when dementia will start and how quickly someone progressed from MCI to dementia."

What's Actually Provable Right Now

These four pieces of research show scientists getting better at forecasting the biological timeline of Alzheimer's, not just the odds of getting it. None of them are diagnostic tests the average person should ask their doctor for today.

The question is whether the ongoing trial testing antibody drugs in asymptomatic, high-risk patients pans out. If it does, Mayeux's concern that there's currently nothing to do with an early positive result goes away, and predictive tests like Columbia's could move from research tool to something doctors actually order.