Blood Tests to Screen for Alzheimer's Disease

Masud Husain


Brain. 2021;144(2):355-356. 

Every now and then, you think you get a glimpse of the future. Sometimes it is a fleeting vision, a shimmering phantom. Too ephemeral to distinguish its form clearly, it promises much but ultimately fades into obscurity, failing to deliver. Very occasionally though it is a longer lasting apparition, one that reveals itself for sufficient time to consider its significance, to ponder what it might mean. The last few months have witnessed such a revelation. For the first time, we can seriously consider the possibility of a blood test to screen for Alzheimer's disease. The ramifications are profound.

In recent years, phenotyping of Alzheimer's disease has seen a major change with the advent of CSF amyloid and tau biomarkers, amyloid-PET and then tau-PET imaging. Together with structural MRI and CSF measures of neurodegeneration—atrophy and neurofilament light chain levels, respectively—this has led to the ATN (amyloid, tau and neurodegeneration) classification system. Characterization of patients into subgroups according to whether they are amyloid positive or negative (A+ or A−), tau positive or negative (T+ or T−) and have evidence of neurodegeneration or not (N+ or N−), has helped to stratify individuals in observational studies and for recruitment into clinical trials. This has been a step change in Alzheimer's research. But what has been equally evident is that such high level phenotyping is possible in only a few centres around the world. Both the costs and the invasive nature of tests have precluded widespread use of such 'staging' of the disease.

Over the past 12 months a series of papers has led to a sharp reconsideration of what might be possible, even in non-research clinical settings. The key development has been the establishment of plasma biomarkers as reliable measures of Alzheimer's disease. First, plasma tau phosphorylated at threonine 181 (p-tau181) has been shown to be present in significantly higher levels in people who have preclinical Alzheimer's disease, defined as being cognitively unimpaired but A+. It is higher still in those with mild cognitive impairment (MCI) or clinical Alzheimer's disease, and correlates significantly with CSF p-tau181, tau and amyloid PET findings (Janelidze et al., 2020; Karikari et al., 2020).

Second, p-tau181 levels distinguished between pathologically confirmed Alzheimer's disease and frontotemporal dementia patients (Thijssen et al., 2020). It may also separate, at the group level, Alzheimer's disease from patients with other clinically defined neurodegenerative disorders such as vascular dementia or Parkinson's disease (Karikari et al., 2020). Third, more recent evidence suggests that the combination of plasma p-tau181 and plasma neurofilament light chain levels predicts well, on an individual basis, which patients with MCI are likely to progress to Alzheimer's disease (Cullen et al., 2021).

In the January issue of Brain, Moscoso and colleagues reported results from a longitudinal study that plasma p-tau181 levels are significantly higher before CSF or PET markers of amyloid pathology (Moscoso et al., 2021). Annual increases in p-tau181 were higher in individuals with preclinical Alzheimer's disease and Alzheimer's disease than in those with A− phenotypes (for a Commentary see Tijms and Teunissen, 2021). Now in this issue of the journal, Keshavan and colleagues show also that plasma amyloid-β species levels measured with liquid chromatography-mass spectroscopy can predict individuals who are A+ on amyloid PET imaging (Keshavan et al., 2021).

These findings make it increasingly likely that a combination of plasma biomarkers—p-tau, amyloid-β species, neurofilament light chain and possibly others—will provide a means to detect Alzheimer's disease at early stages of the disease. Emerging evidence that p-tau217 actually surpasses p-tau181in diagnostic performance (Palmqvist et al., 2020) adds to the likelihood of diagnostic blood tests for Alzheimer's disease in memory clinics and even population screening for the condition becoming a reality within the next few years. Of course, a positive screen result would need to be taken into context, followed-up with 'high resolution' cognitive testing. These are likely to be computerized tests that are sensitive, have dynamic range and can track small changes in performance over time. With a combination of plasma biomarkers and such cognitive tests, it might be possible to recruit patients with Alzheimer's disease but without dementia into therapeutic trials far earlier and monitor the effects of experimental therapies far better than we can now. A serious vision is taking shape.