Screening for Alzheimer’s disease: plasma biomarkers and what else?

It is coming. There is no doubt it about it. Within a few years we will be entering a time when screening for dementia becomes routine. First, a series of reports on plasma biomarkers for Alzheimer’s disease (AD) have now convincingly demonstrated that levels of phospho-tau (P-tau) and Aβ-amyloid species, neurofilament light chain (NfL) and glial fibrillary acid protein (GFAP), on their own or in combination, can be useful to discriminate between people with AD or mild cognitive impairment (MCI) and healthy individuals.1 Plasma P-tau 181 also correlates with both Aβ-amyloid PET and tau PET signals2 and is significantly raised compared to controls in pathologically confirmed cases.2,3 These findings provide strong evidence of a pathophysiological link between plasma biomarkers and brain AD pathology. Second, even without the recent findings on the anti-amyloid monoclonal antibody lecanemab,4 there is now accumulating evidence that modification of lifestyle factors might have an impact on the trajectory of cognitive decline.5 Thus detecting people at risk of dementia early might be worthwhile in order to optimize vascular risk factors, diet and exercise, which can be effective interventions in maintaining or improving cognitive health.6 While these results represent a remarkable step forward in translation of basic science findings, they also pose some serious questions about where we go next in the clinic—and beyond. One potentially important application might be the use of blood-based biomarkers for recruitment into trials, particularly of people who perform within the normal range on cognitive tests but have pathological levels of plasma biomarkers. These individuals have been reported to have a higher risk of developing AD7 and it would be of great interest to investigate whether disease-modifying therapies that clear amyloid and/or tighter control of vascular risk factors would have any impact on the likelihood of subsequent dementia. This form of enrichment into clinical trials also allows for stratification of individuals, e.g. those with positive plasma biomarkers for AD, or MRI markers of small vessel cerebrovascular disease, both, or none, with selective intervention on the basis of which group they belong to. Some very recent evidence also points to the potential use of one plasma biomarker, P-tau 217, in tracking progression of disease in both preclinical and symptomatic AD.8 If confirmed, this could provide an important surrogate for the effectiveness of disease-modifying therapies. A second potential use of plasma biomarkers is for the more ambitious aim of screening for dementia, or people at risk of developing dementia, at a population level. To do this requires a national infrastructure. We saw this is available, in some countries at least, during the COVID pandemic. However, it has also become evident that risk evaluation by plasma biomarkers alone is unlikely to be completely effective. In this issue of Brain, Stevenson-Hoare and colleagues9 provide evidence that adding genetic testing to blood-based biomarkers can improve accuracy of AD diagnosis further. While the combination of APOE genotype, polygenic risk score and plasma Aβ40 and Aβ42, P-tau 181, NfL and GFAP provided a receiver operating characteristic area under the curve (AUC) of 0.81, the prediction accuracy of all biomarkers without genetic factors was not as good (AUC = 0.75).9 COVID has also taught us of the value of another important tool: remote or digital cognitive assessment.10 Although this was developed in a few select centres before the pandemic, there are now several platforms which allow for effective screening of cognitive function using online tools. Cognitive biomarkers therefore provide a third potential index to add to plasma biomarkers and genetics for dementia screening. While access to computers or other digital interfaces is limited in some communities, it is also important to remember that assessment of blood-based biomarkers and genotype, without a measure of cognitive performance, would rather be missing the point of screening for people with dementia or those at risk of developing it. Combining remote cognitive assessment with plasma and genetic testing provides hope for a practical and effective means of screening for dementia at scale.