A Blood Test for Alzheimer's Disease?

Looking ahead. A blood-based test predicting the onset of Alzheimer's disease could help families plan future care.

Looking ahead. A blood-based test predicting the onset of Alzheimer's disease could help families plan future care.


Alzheimer’s disease is expected to triple in prevalence by 2050, affecting 115 million people worldwide. There’s no cure or treatment yet for the fatal neurodegenerative condition, but many physicians and scientists suggest that drugs that have failed so far will work if given much earlier, a strategy that requires diagnosing the disease before symptoms develop. Now, a research team has discovered a group of molecules in the blood that they say can predict with 90% accuracy whether older people will develop the disease over the course of 2 to 3 years. Although such a test is not ready for general use, and may never be, the technique could still help recruit people most at risk of developing Alzheimer’s into clinical trials of possible treatments.

Beyond an autopsy analysis of a person’s brain, two accepted methods of diagnosing Alzheimer’s disease exist at present, says Douglas Galasko, a neuroscientist at the University of California (UC), San Diego, School of Medicine. One technique uses brain imaging to detect the hallmark protein found in plaques in brain tissue which marks the disorder. The other measures levels of these proteins by extracting fluid from the spinal cord. Few people wish to undergo that painful procedure, however, and because both techniques are expensive and not terribly accurate, particularly at the earliest stages of Alzheimer’s, researchers have spent decades looking for a less invasive, more affordable blood-based test. So far, however, these efforts have produced “no success,” Galasko says.

To ferret out blood molecules that might signal the onset of Alzheimer’s disease in elderly people, a team from Georgetown University and several other institutions recruited several hundred senior citizens age 70 and up from retirement communities in New York and California. They took blood samples and shipped them on ice to a lab with a mass spectrometer in order to precisely quantify the blood samples’ chemical makeup.

Over the next 3 years, the researchers tracked the seniors’ mental health, and identified 53 people with mild cognitive impairment or Alzheimer’s disease, 18 of whom had not displayed any symptoms at the beginning of the study. At the 3-year mark, they returned to the analyses of the blood samples and compared those of the people who had developed symptoms of Alzheimer’s disease with 53 of the elderly people from the group who remained healthy. In the group whose mental health had declined, there were significant alterations in the blood levels of 10 different chemicals, including fatty molecules called phospholipids, which help keep cell membranes in the brain and body intact, the team reports online today in Nature Medicine.

To check that its observations weren’t just a random event, the team tested whether the same altered pattern could predict whether 41 other elderly people from the same retirement communities had developed Alzheimer’s disease, and found that it hit the mark 90% of the time. Despite these encouraging findings, senior author Howard Federoff, a neuroscientist at Georgetown University in Washington, D.C., says the test needs much more validation. “This is a new observation, and it’s imperative that it be extended and replicated on an independent group of individuals.”

Others are equally cautious. “We won’t know if it’s going to be a big deal or not” until other groups replicate the study, agrees Michael Weiner, a neuroscientist at UC San Francisco. The population of people with Alzheimer’s disease is so diverse, and is fraught with so many health problems, that it may turn out that although the test is good at detecting people who are at risk for cognitive decline, it will pick up too many other conditions in the process to be useful as a diagnostic tool. If that’s the case, the test could still be used to screen people for preventative clinical trials of Alzheimer’s drugs for the disease, he says.

That Federoff and colleagues validated their findings in an independent group of elderly people is “impressive,” says Robert Green, a medical geneticist at Harvard University. Many such studies “have turned out to be a flash in the pan,” he says, but the new study “is more sophisticated than most.”

*Correction, 17 March, 4:06 p.m.: This story has been corrected to reflect that only one of Alzheimer's two hallmark protein deposits can currently be imaged in the brain for diagnostic purposes.

Follow News from Science