touchCARDIOÂ coverage of data presented at ESC 2024
touchCARDIO spoke with Dr David Conen, Clinician scientist, Population Heath Institute, Hamilton, Ontario, Canada, about his presentation on ‘Biomarkers for improved risk prediction and biological insights in patients with atrial fibrillation’. He summarizes his current research looking at multiple different outcomes, which biomarkers are associated and how they might improve risk prediction. He describes his findings that it isn’t a single pathway and currently there is a clinical need for more larger, comprehensive studies with multiple biomarkers and other phenotypic associations to determine the different outcomes more precisely.
Disclosures: David Conen has no financial or non-financial conflicts of interest to declare in relation to this video.
This content has been developed independently by Touch Medical Media for touchCARDIO and is not affiliated with ESC. Unapproved products or unapproved uses of approved products may be discussed by the faculty; these situations may reflect the approval status in one or more jurisdictions. No endorsement of unapproved products or unapproved uses is either made or implied by mention of these products or uses by Touch Medical Media or any sponsor. Views expressed are the speaker’s own and do not necessarily reflect the views of Touch Medical Media.
Click here for more content on atrial fibrillation and for further ESC 2024 highlights visit here.
Transcript
Introduction
Hi everybody, my name is David Conen. I’m a clinician scientist here in Hamilton, Ontario, Canada. I’m a senior scientist at the Population Health Research Institute here in the city, and I’m also Associate Professor at McMaster University [in Hamilton, Ontario].
Q1. What are the most promising biomarkers for atrial fibrillation?
I don’t think there is the single promising biomarker for atrial fibrillation because it depends a little bit on what you were looking at. There is risk prediction, there is looking at mechanistic pathways and there’s obviously also multiple outcomes that patients with atrial fibrillation face. In our work it’s interesting to see that despite looking at multiple different outcomes (such as heart failure, cardiovascular death or stroke) several biomarkers pop up for multiple outcomes. For example, GDF-15, which is a member of the GGF-β family and associated with oxidative stress, is associated with heart failure, cardiovascular death and also bleeding. So multiple different outcomes that probably have a different pathophysiological background, but seem to be covered by GDF-15.
GDF-15 was also, most of the time, stronger than somewhat more established biomarkers, such as, [cardiac] troponin T or NT-proBNP, which are strong predictors of heart failure and cardiovascular events, but they’re also associated with bleeding. So I think that there’s a growing number of biomarkers that explain mechanistic and risk prediction insights for atrial fibrillation outcomes, but there’s not a single one. I would actually say we need large comprehensive studies with multiple biomarkers, as many as possible, so that we can look at which clusters or which biomarkers are actually associated with different outcomes precisely.
Q2. How might the use of biomarkers assist in improving risk prediction in patients with atrial fibrillation?
I think risk prediction for various outcomes in patients with atrial fibrillation is a very important topic, especially if we have established treatments for patients who are at highest risk. For instance, the most famous issue in atrial fibrillation patients is stroke. For stroke prevention we do have anticoagulation, and there are well established risk prediction tools for this. However, even if those risk prediction tools are well established, they’re not very precise. For instance, the CHAâ‚‚DSâ‚‚-VASc Score, which in Europe is most often used, is a very poor predictor of outcomes of strokes in patients with atrial fibrillation. It’s mainly useful for excluding patients at low risk of anticoagulation. It’s not good at predicting who is actually getting a stroke in those who have a high score.
So I think in this area, multiple clusters of biomarkers, the more the better, are needed to try to improve risk prediction so that we can better target treatments that do have some side effects, but on average are very effective.
Q3. What biological insights may come from the identification of biomarkers in these patients?
So looking at mechanistic pathway is the second promising area where biomarkers are potentially very useful, as opposed to risk prediction. Obviously, when you look at pathological insights, we’re looking at causal associations, while for risk prediction we’re only trying to predict who is at risk, and it doesn’t really matter whether the risk marker that we look at is causally related to the outcome. For mechanistic pathways, obviously causal associations are required, and this is a little bit more challenging to do in observational studies where residual confounding and biases are always a problem.
However, by measuring more biomarkers at the same time, with as little blood as possible (and there are multiple platforms that will be able to do so) and then adding gene expression pathways, we can start looking at really what is causing those outcomes. When we know this, we could then look at what treatments we have available or what treatments we can create to improve outcomes or reduce the risk of outcomes in patients with atrial fibrillation, who do have a high risk of complications.
Q4. What are the challenges and future directions in integrating biomarker data into clinical practice for improved personalized treatment strategies in atrial fibrillation?
So far we have mainly talked about research on how we can improve outcomes. The second step is how we can apply this in clinical practice. For risk prediction, as I already mentioned, the CHAâ‚‚DSâ‚‚-VASc Score or similar scores are regularly used in clinical practice to decide on which patients should receive or not receive oral anticoagulation. The the good thing about the CHAâ‚‚DSâ‚‚-VASc Score is that it’s very simple. It can be used without calculation tools and you can do it in your head in front of the patient. When we start looking at more complex models, I think we will need integrated calculation tools in [electronic medical records] EMR systems or in clinical practice so that it can be easily calculated or presented to the clinicians when they are in front of a patient.
I think given the increasing use of technology, AI and other things, it should not be a big deal to implement calculation tools that are more complex than other tools. In our work, we try to apply some machine learning algorithm pathways to try to improve existing conventional Cox model tools. We found that some of them may be slightly better, and I think it’s a good start. [To achieve this] we will need very large datasets, very large studies with multiple biomarkers and other phenotypic associations to really make sure that we get these complicated associations right because, again, I don’t think it’s a single pathway. There’s multiple networks going on, and we need to decipher all of them, in the future.
Interviewer/Editor: Heather Hall
Cite: Conen, D. Biomarkers for improved risk prediction and biological insights in patients with atrial fibrillation. touchCARDIO, September 19, 2024.
