There are lots of activities ongoing towards implementation of the blood-based biomarkers for Alzheimer’s disease. I alluded on that in my own presentation about activities for clinical implementation. We need to understand well what the behavior is of the plasma biomarkers relative to PET analysis, for example. And that was focused especially on the plasma p-tau markers. We found that plasma p-tau is equally good as tau PET for identification of early stage pathology, so amyloid pathology in those who are cognitively unimpaired...
There are lots of activities ongoing towards implementation of the blood-based biomarkers for Alzheimer’s disease. I alluded on that in my own presentation about activities for clinical implementation. We need to understand well what the behavior is of the plasma biomarkers relative to PET analysis, for example. And that was focused especially on the plasma p-tau markers. We found that plasma p-tau is equally good as tau PET for identification of early stage pathology, so amyloid pathology in those who are cognitively unimpaired. But that tau PET would be more useful for prognosis and monitoring disease progression in later stages. But it’s always a balance between what’s best and what’s feasible. And blood barrier markers are, of course, more convenient than tau PET analyses, and less invasive.
On the other hand, we are also looking at other biomarkers, for example GFAP. And that’s an astrocyte protein, and it may sound or may appear to be non-specific for Alzheimer’s disease, but it has a strong prognostic value. And we observed in the DIAN cohort for familial AD that there is a relation between GFAP and cognitive decline in mutation carriers, and especially in those who are amyloid positive. So we do think that there is a strong interaction between… Or some specificity, if you may, between amyloid pathology and the relations of GFAP and cognitive decline.
And yeah, based on these results, we further our studies also, the relation of GFAP and cognitive decline in normal population-based cohorts. Because those normal aging populations are very interesting if we are thinking of implementation of the blood-based biomarkers in primary care, for example. Or for screening for trials in a more general population. And we observed that it was a relation also longitudinally between p-tau and GFAP, and also amyloid beta, and the presence of longitudinal change also in those who are amyloid positive and amyloid negative. And that, again, especially those who are amyloid positive, the relation was stronger for GFAP and also p-tau. This again stresses the relation between those biomarkers and cognitive decline.
So those are some of the results that we obtained to understand better how the plasma biomarkers are changed in the early stages of Alzheimer’s disease. But on the other hand for implementation, we also need some other aspects to be addressed. And one of these aspects are the pre-analytical protocols. So for example, what kind of tube do you need to obtain the blood? How do you need to process the blood? Can you allow some variation? For example, if you do blood draw in a more general population, then the conditions are less well-defined, or controlled, than when you are in a memory clinic close to the lab.
So we assessed the effects of delays in the processing time. And we observed that delays in processing time could affect amyloid levels in plasma, but not so much GFAP or neurofilament light, or, yeah, in the first studies, we also observe not so much in effect on p-tau. So based on the results and we analyzed other conditions as well, but this was the most important that really had the strongest effect on the amyloid levels. But based on the results we defined SOP, so standard operating protocols for how the sample should be processed. So that is a technical aspect, was very relevant also to know if you think of implementation.
And another study that we presented also at AAIC, by one of our researchers in our group, Inge Verberk. She showed that we are now defining cutoffs based on a cohort of 1200 patients from the Amsterdam Dementia Cohort, for the optimal discrimination in the early stages and later stages and what combination of biomarkers would be best. Because we need that for implementation, because we want to start this fall with a real-world evaluation of the use of these biomarkers. So this is preparative work to understand how we should make our decisions in the memory clinics. And that’s a study that we will start because real world evaluation is very much needed. The studies so far have been performed in quite selected cohorts. Mostly volunteers who participate in the research, with also clear diagnosis based on CSF and PET, and also on the clinical picture. But in the real world, we will encounter also other patients with cognitive problems. So we want to understand how the biomarkers work also and how it in any case increases the confidence. Maybe the doctors are so happy with the results and have a lot of confidence in it that part of the AD patients, we don’t need to do some additional testing. And that may be the results of the study. And that’s what we hope for, and with additional testing I’m thinking of CSF and PET analysis.
So yeah, that should be one, or we will get more information based on those studies. So that’s a little bit what we are working on for the plasma biomarkers and how to implement them, but there are many other studies. And one important issue is also to include it in guidelines and to provide guidance, how to use these biomarkers. And for that, a paper was published during AAIC and Oskar Hansson has presented the appropriate-use recommendations. So that’s not yet appropriate-use criteria. So that’s next level that, is to follow – the criteria.
But the recommendations are clearly, now that you can use the biomarkers, start to use in specialized clinics where we have the possibility for confirmation by CSF and PET. But it’s early days, so we think that the confirmation is still needed. We can use them also in trials for pre-screening and again, the results should be confirmed by CSF and PET, but those who are negative and not included in the trial, then you can move on with those who test positive, but those need to be confirmed.
And the last indication is whether we can use it in primary care. And our recommendation is that it’s at this stage not yet possible. Given the paucity of data that we have, and the difference in prevalence in the settings, in the primary care that we cannot translate easily results from the memory clinic to the primary care. And also that you don’t have a direct possibility for confirmation with CSF and PET. So we think it’s too early to implement in primary care.
I do think that the plasma biomarkers, they are clearly very important and gaining a lot of attention. They are developed at very high speeds, very promising also. So we need to be careful also to not to implement them too quickly and to gain enough information so that we have enough confidence for using them. And at the same time should also not be reluctant. Because the results are very consistent, very promising. So yeah, it will not take too long anymore, I think, before we feel confident enough for implementation in clinical practice.