A better test for Alzheimer’s disease
The tool revolves around examining mild cognitive impairment (MCI), which is often a precursor to Alzheimer’s disease.
“Someone with mild cognitive impairment is usually living normally in the community and able to look after oneself, but when tested with neurocognitive tests, performing below what would be expected for this age,” according to Professor Nicolas Cherbuin, head of the Centre for Research on Ageing, Health and Wellbeing at the Australian National University, and one of the developers of the tool.
“They might have some memory lapses or other difficulties in thinking, but generally speaking, they live normally.”
Around one in six people aged over 60 have MCI, but it’s not always an indicator of Alzheimer’s.
“Of those, about one in three progress to Alzheimer’s disease within 1.5 to five years,” says Cherbuin.
“But it leaves two in three who do not progress, who either remain stable or, for a small fraction, might even return to normal cognition.”
Currently, it’s difficult to predict an individual’s Alzheimer’s risk, because the diagnostic tools are difficult to access.
“Some are very invasive, and therefore they’re not recommended or suitable for frequent use in a clinical setting. Some might be very expensive, and many are technically demanding,” says Cherbuin.
Cherbuin, along with his medical student, Nicolas Darmanthéc, and colleague Dr Hossein Tabatabaei-Jafari, set out to rectify this by developing a simpler predictive tool.
“We used data from the Alzheimer’s Disease Neuroimaging Initiative [a US-based longitudinal study], and we focused only on people who had mild cognitive impairments, for many years, with multiple assessments. So it was possible to tell at what stage they progressed from this mild cognitive impairment stage to the clinical Alzheimer’s disease stage.”
The researchers examined the subjects’ scores from a test called the “mini-mental state examination”, which is commonly used in clinics, and a biomarker called plasma neurofilament light chains (pNFL), which can be found with a blood test.
“When neurons in the brain are damaged or die, they start breaking apart in small pieces, and part of the scaffolding of these cells breaks into little chains,” explains Cherbuin.
“One of these chains is called a neurofilament light chain. These fragments of neurons then make their way into the bloodstream where, if we take a blood draw, we can measure them.”
Combined, the mini-mental state examination and the pNFL test had good predictive power for Alzheimer’s.
“What we found is that when we combine these two measures together, we can predict with good accuracy who is at higher risk versus less risk of progressing towards Alzheimer’s disease within five years.
“That ‘within five years’ is the important factor,” adds Cherbuin. “Studies that have predicted conversion from mild cognitive impairment to Alzheimer’s disease have not really focused on a time frame, and for use in a clinical setting, this is what is really needed.”
People with MCI don’t just want to know their chances of developing Alzheimer’s – they want to know if it’s imminent, as well.
“If they have a better sense earlier as to whether they’re at higher risk or not, they can plan what they want to do and how they want to be treated,” says Cherbuin.
“They might want to write a will. And it might also provide an opportunity for the clinician to target the treatment.”
This predictive test could be clinically available in two to three years, assuming it succeeds at spotting Alzheimer’s in a few more clinical trials. But there are two major roadblocks at the moment for it to be widely used.
One is the data on the pNFL test. “It’s widely available for research, and it’s not particularly expensive, and it’s reliable, but it hasn’t been approved for clinical use,” says Cherbuin.
This is likely to change in the near future, however, as a large multi-centre study on the pNFL test wraps up; according to Cherbuin, early data from the study is promising. Once approved, the test can be done easily at large pathology labs.
The other roadblock is further trials of the predictive tool. “We’ve shown this in one very well-characterised population, but it needs to be repeated in in several other populations to confirm it behaves in the same way,” says Cherbuin.