A Blood Test May Reveal Who Responds Best to Alzheimer’s Drug Lecanemab

TWENTY MILLILITRES of blood, drawn from the arm every few months, is turning out to be enough to tell whether an Alzheimer’s drug is doing its job. A blood test for a protein called p-tau217 can flag, within three to six months of starting the infusion drug lecanemab, which patients are pulling ahead biologically and which are not, months before a cognitive test would show the difference. Researchers at Korea University Guro Hospital in Seoul tracked the marker in 153 people with early Alzheimer’s disease who had just begun the drug, then followed their memory and thinking for a year afterward. Patients whose p-tau217 fell fastest in that first stretch also lost less ground on standard cognitive scores over the following months, and the study’s biggest clue about why sits in blood pressure, not brain scans.

Lecanemab works by clearing amyloid, the sticky protein that clumps in the brains of people with Alzheimer’s, and its pivotal trial found that patients on the drug declined moderately more slowly than those on placebo. It is expensive, it requires an infusion every two weeks, and, per the Alzheimer’s Association’s clinical guidance for prescribers, it carries real side-effect risk that calls for structured monitoring, so knowing early whether it is working matters. Until now, the main way to check has been repeat brain imaging, which is slow, costly, and not always available.

A Biomarker That Moves Fast, Then Settles

P-tau217 is a fragment of the tau protein that shows up in blood roughly in step with the amyloid and tangle changes happening in the brain. Doctors already use it to help diagnose Alzheimer’s, with accuracy rivaling a spinal-fluid test or a brain scan; the Korea University team wanted to know whether it could also track treatment response, the way a lower cholesterol number tracks a statin working. 81 of the 153 patients had complete blood draws at the start of treatment and again at three and six months, enough to chart an actual trajectory rather than a single before-and-after snapshot. Levels fell soon after treatment began, kept falling faster between three and six months, and then largely flattened out through the twelve-month mark, a curve the researchers describe as an early, fast-moving signal that settles once the drug’s initial effect on amyloid has played out.

Not every patient’s curve looked the same, though. Splitting the 81 patients by how sharply their p-tau217 fell produced two clearly separated groups: 29 whose levels dropped hard and kept dropping, and 52 whose levels barely budged or even ticked up briefly before a modest decline. That gap turned out to matter for more than the blood test alone.

Where Blood Pressure Enters the Picture

People in the steep-decline group had lower amyloid burden going in and slightly better memory scores at baseline, which is not entirely surprising. What is surprising is how strongly a single common condition sorted the two groups: hypertension. Only three of the 29 steep decliners had high blood pressure, against twenty-two of the 52 slow decliners, and the pattern held even after the researchers adjusted for age, sex, amyloid levels, and cognitive status at the start. After that adjustment, a patient with hypertension carried an adjusted odds ratio of just 0.09 for landing in the steep-decline group, a stark separation for a single yes-or-no risk factor. The authors suspect the blood vessels themselves may be part of the story, since high blood pressure is already linked to a leakier blood-brain barrier and impaired clearance of amyloid from brain tissue, both of which could blunt how efficiently amyloid, and downstream tau, actually get cleaned out.

“We found that plasma p-tau217 levels decreased significantly from 3 months after treatment initiation,” says Sung Hoon Kang, an associate professor of neurology at Korea University College of Medicine who led the study. “The greatest decline occurred between 3 and 6 months, followed by a plateau.” His team is now watching whether that plateau, and the split between steep and shallow decliners, holds up as more patients complete a full 18 months of follow-up.

Why the Split Matters for the Clinic

The clinical payoff shows up in the cognitive numbers. Patients in the steep-decline group held roughly steady on the Mini-Mental State Examination over the year, while the slow decliners drifted downward. On the Clinical Dementia Rating scale, a measure clinicians use to track how much dementia interferes with daily function, the slow decliners worsened noticeably faster than the steep decliners, a gap that stayed significant even after the researchers controlled for how each patient started out. Adding a patient’s blood-based trajectory group into the statistical model, on top of everything already known about them at baseline, sharpened the model’s ability to predict how their thinking would change over the following year, on both cognitive measures the team examined. That is the part that turns a lab curiosity into something closer to a clinical tool: a number that predicts more than it already knew.

None of this proves the biomarker is driving the outcome rather than simply riding alongside it. The study had no untreated comparison group, so there is no way to separate how much of the biomarker’s fall reflects the drug specifically versus the natural course of early treatment in a real clinic. It is also a single hospital’s patients, just 81 of them in the trajectory analysis, split further into groups of 29 and 52, numbers too small to nail down an exact cutoff a doctor could act on.

The team could not directly test the vascular explanation either. Confirming that hypertension blunts p-tau217 response by damaging clearance pathways, rather than simply marking patients who started with more underlying tau pathology, would need brain imaging of tau itself or a matched cerebrospinal fluid measurement, neither of which was collected here.

Even with those caveats, the direction of the finding lines up with a broader pattern researchers have been circling for a few years now: vascular health and Alzheimer’s biology are tangled together more than the old amyloid-only story suggested. A patient’s blood pressure was, oddly enough, more informative here than their starting amyloid load. If that holds in larger, multicenter follow-up work, treating blood pressure aggressively before or during anti-amyloid therapy could become part of getting the drug to work as intended, not just a separate item on the chart.

“As blood-based Alzheimer’s biomarkers become widely available,” Kang says, “repeated p-tau217 testing may allow clinicians to monitor biological treatment response in a manner that is less invasive, less costly, and more accessible than repeated PET imaging.” Whether that promise survives contact with a bigger, more varied patient population is the next thing to find out, and Kang’s hospital registry is still enrolling patients to answer it.

See also:

• Alzheimer’s Brain Damage May Start In The Neck
• Two Proteins In Ordinary Blood Tests Track Who Loses Independence
• Youth Protein TIMP2 Restores Cleanup Crews In The Aging Mouse Brain

FAQ