ARTESiA: Apixaban vs Aspirin Bleeding Risk with Dr D Siegal

Major bleeding remains the principal complication of oral anticoagulation. In patients with device-detected subclinical atrial fibrillation, the decision to anticoagulate requires careful balancing of stroke prevention against bleeding risk.

In this episode of CLOT Conversations, Dr. Deborah Siegal discusses a prespecified subanalysis of the ARTESiA randomized clinical trial, recently published in JAMA Cardiology. ARTESiA demonstrated a 37% reduction in stroke and systemic embolism with apixaban compared to aspirin — but at the cost of increased major bleeding.

This subanalysis goes deeper, examining the site, severity, and clinical course of bleeding events. Most bleeding was gastrointestinal and non-critical. Rates of intracranial and fatal bleeding were low and similar between treatment arms. The majority of events were not clinical emergencies, and many were defined by hemoglobin decline rather than catastrophic presentation.

We explore what these findings mean for individualized risk assessment, the importance of modifiable bleeding risk factors such as NSAID use, and how physicians and patients can approach shared decision-making in subclinical AF.

Abstract (subscription required for full paper): 

Siegal DM, Sticherling C, Healey JS, McIntyre WF, Christensen LS, Parkash R, Vanassche T, Conen D, Gold M, Granger CB, Nielsen JC. Major Bleeding With Apixaban vs Aspirin: A Subanalysis of the ARTESiA Randomized Clinical Trial. JAMA cardiology. 2025 Dec;10(12):1305-14.

https://jamanetwork.com/journals/jamacardiology/fullarticle/2841075

Transcript

Welcome back to CLOT Conversations, the podcast where we explore the latest evidence shaping thrombosis and antithrombotic care. I'm David Re, executive Director of Thrombosis Canada, and I'm joined once again by my co-host, Dr. Jameel Abdulrehman from Toronto General Hospital and Dr. Maha Othman from Queens University in Kingston.

Thank you for being with us today. We're discussing a clinically important and highly practical topic. Major bleeding with apixaban versus aspirin in patients with subclinical atrial fibrillation based on a pre-specified sub-analysis of the Artesia randomized clinical trial recently published in JAMA Cardiology.

Jameel, I'll turn it over to you to introduce our guest. We're very pleased to welcome Dr. Deborah Siegel. Dr. Siegel is a clinician scientist hematologist specializing in thrombosis medicine at the Ottawa Hospital. She's an associate professor of medicine and the school of Epidemiology and Public [00:01:00] Health at the University of Ottawa, and the scientist at the Ottawa Hospital Research Institute.

Her research focuses on improving the safety of anticoagulant therapy, including management of anticoagulant related bleeding complications, restarting anticoagulation after bleeding, preoperative management of anticoagulants and antithrombotic therapies for ischemic stroke in patients with cancer. She also serves on the board of directors for Thrombosis Cana da.

So Deborah, it's so lovely to have you today. And, why don't we start with the big picture. So, how big an issue is bleeding on DOACs, such as apixaban and on antiplatelet agents like aspirin, particularly in patients with subclinical atrial fibrillation. And what was the motivation for conducting this sub-analysis of Artesia?

Thank you so much, for the opportunity to be here to talk to you about this analysis and the importance of understanding bleeding. Bleeding is, even in 2026, bleeding, is the main complication of treatment with oral anticoagulants. And so patients who are taking oral anticoagulants like DOACs or direct oral anticoagulants.

For atrial fibrillation or venous thromboembolism experience rates of major, which we think of as being more serious, bleeding of about two to 4% per year. And that's quite variable depending on the patient population and, the individual drugs and the indication for treatment. But we still see that bleeding on anticoagulants is the number one adverse drug event that brings patients to hospital to the emergency department and actually causes death. So it is an important iatrogenic complication of treatments that patients need, to prevent clots. So we conducted this sub-analysis of the Artesia trial in order to better understand the major bleeding events that occurred in this study.

So you conducted a pre-specified sub-analysis of data from the Artesia RCT Could you give us a summary of the Artesia RCT and what was the goal of this sub-analysis? 

So patients with subclinical atrial fibrillation, as you mentioned, so those are patients who have atrial fibrillation that's detected on an implanted device like a pacemaker or an implantable cardiovert or defibrillator, for example. Those patients may have short episodes of atrial fibrillation that don't quite meet the definition of atrial fibrillation. We call that subclinical. Those patients have an elevated risk of stroke and systemic embolism. But the role of oral anticoagulant treatment in this population is not clear, because the rates of stroke were actually lower than those with established AFib.

And also there's an increase in the risk of bleeding, as we had already talked about. So Artesia was a trial which randomized patients who had device detected, subclinical atrial fibrillation to either receiving treatment with apixaban at a standard dose, so five milligrams twice daily versus aspirin.

And the Artesia trial enrolled, 3,961 patients, so it was a large trial and showed that the risk of stroke and systemic embolism was decreased by about 37%. With apixaban compared to aspirin, but that this effect was accompanied by an increase in the risk of major bleeding. And so it's important for us to understand not just, the stroke and systemic embolism side of things, but also the major bleeding side of things.

Because in clinical practice, we want to make sure that we're treating patients so that the net clinical benefit, favors benefit over harm, was why the sub-analysis was done. We have an understanding of what happens to patients after ischemic stroke and systemic embolism. We know that many patients experience, fatal events or, long-term functional impairments or disability as a result of stroke, but on the bleeding side, it's not as clear to us what happens, what those bleeding events are, where they arise, and what happens to patients after those bleeds.

So we thought that it's really important to understand not just the risk, like not just the absolute risk of having a stroke or a bleeding event, but also an understanding of the type and severity of these events. And that's really what helps physicians and patients make decisions about treatment.

So regarding major bleeding, how did the characteristics of patients who experienced major bleeding differ from those who didn't? Yeah, so we know that, as in other studies, and this has been pretty well established, we know that patients who have major bleeding tend to be older and have a higher degree of comorbidities compared to, other patients. And so we did see that in apixaban as well. There were 133 patients out of the 3,961 who experienced major bleeding. The average age of those patients was older. Their body weight was lower and their, body mass index was lower. They also had some differences probably in the distribution of CHADs-VASC or, or the, the stroke risk categories that we talk about.

Again, this is an older patient population with a number of cardiovascular comorbidities. And then we also saw that patients who had bleeding had a lower creatinine clearance, which implies that they had worse renal function than the patients who didn't have bleeding. So again, some of the major risk factors that we've seen, time and time again in other studies have shown up in our study.

Right. Okay. So with bleeding events, how did they compare between apixaban and aspirin and at what sites did they occur?

What was interesting is we had an opportunity in the Artesia trial to, characterize the major bleeding events in a different way maybe than some other trials because these events were adjudicated by a committee that was blinded to treatment allocation.

And this was an independent committee. And not only did we adjudicate whether a major bleeding event happened, so like yes or no, but we adjudicated what criteria were fulfilled in order for that bleeding event to occur. So with ISTH major bleeding, people are probably familiar, you know, fatal bleeding events, bleeding occurring at critical sites like in the brain or spinal cord, for example. A drop in hemoglobin of 20 grams per liter or more, or a transfusion of two or more red blood cells. And so what we wanted to understand was, you know, major bleeding is kind of a broad category, with criteria that are kind of heterogeneous in terms of maybe their prognostic importance.

So what we wanted to understand was, what do the bleeds look like, as opposed to just kind of calling them major bleeds. A big advantage of our study was to say, well, what is actually happening to patients? Who is experiencing these bleeds? Right. So what was interesting, in our study was that we saw patients who were treated with Apixaban had more bleeds at a non-critical site, whereas the rates of bleeding at critical sites was similar between the two groups.

So there was a similar, proportion of people in the apixaban group, 1.3%, versus 1.1% who experienced critical site bleeds. Whereas a higher proportion of people on apixaban, versus aspirin, so that was 3% versus 1.3% had non-critical site bleeds. And, majority of these bleeding events occurred, arising from the gastrointestinal tract.

That was kind of the single most common site of, non-critical site bleeding in this study. 

Right. So I guess you may have alluded to that the study has used the ISTH definition of major bleeds, which includes several components, which criteria most commonly drove the classification of major bleeding in Artesia.

Yeah, so that's where things kind of got interesting, I think for us to, we saw that actually the rates of fatal bleeding and critical site bleeding were low and similar between the two treatment groups. So between patients who received apixaban and those who received aspirin, where we saw a difference was, in the hemoglobin decrease of 20 grams per liter or more, where we saw that there was a difference. A higher proportion of people who were receiving apixaban had bleeds associated with a drop in hemoglobin compared to those receiving aspirin. And that was about 3.1%, versus 1.2%.

Can you tell us a bit about the severity and clinical course of these bleeding events and how they were managed?

So this is another kind of novel feature I think of this trial and, and how these events were adjudicated was that the adjudicators not only determined whether or not a major bleeding event occurred and what criteria were, were fulfilled for that bleed, but also, some measure of the clinical severity of presentation and the clinical, course of major bleeding.

And these were based on, criteria that were developed for other types of studies, in the, in the venous thromboembolism space, just to give a sense of what type of, events they were, the emergencies. Did patients require some treatment. So you'll see if you are looking in the study and we, we show the distribution of the severity of the clinical bleeds at presentation. The majority of bleeds were not determined to be clinical emergencies, although patients required some treatment, so that could be like a transfusion, for example. And that was the sort of the vast majority. So overall, 75% of events, bleeding events that occurred in this, trial were not considered to be clinical emergencies by the adjudicators, but some of those, folks did require some treatment and that could have been a transfusion.

So the multivariate analysis identified several predictors for major bleed. And you may have said something around this, but which factors are independently associated with the bleeding risk?

Yeah, so we did look at, some risk predictors. One of the things that came out to be a really important predictor with a high magnitude, risk was, the use of NSAIDs or non-steroidal anti-inflammatory drugs. And the reason I want to highlight non-steroidal anti-inflammatory drugs use is because this is actually a modifiable risk factor for bleeding. And we know that a lot of patients have risk factors for bleeding, such as anemia, renal failure, age, prior bleeding history, those are things we can't modify. But the use of NSAIDs is something that we can modify. So in this study, I think that was one of the most important findings is to say no. That's been shown before, but it's helpful for it to have borne out in our study to say, it reinforces the importance of us, when we see patients to make sure that we're questioning the use of these drugs and de-prescribing as necessary.

How do your findings compare to prior studies comparing DOAC to aspirin, including the AVERROES trial, which looked at apixaban to aspirin for AFib in patients unsuitable for BKA and the NOAH-AFNET 6 trial, which compared edoxaban, to aspirin or placebo for atrial high rate episodes.

Yes, the Artesia, it's interesting to consider the different types of data that we have in this space. AVERROES as you mentioned, was a randomized trial, slightly different population because those are patients with, kind of established atrial fibrillation, who were deemed not to be suitable for vitamin K antagonist or warfarin. And they were randomized to either apixaban five milligrams twice daily, which is a standard dose, or aspirin.

It was a really important trial because, prior to AVERROES aspirin was being used, for patients with atrial fibrillation considered to be at high risk of bleeding and not suitable for a VKA. But the AVERROES results showed that apixaban was essentially twice as good at preventing stroke and systemic embolism with a similar risk of bleeding compared to aspirin.

So essentially, after the AVERROES trial, really aspirin was no longer recommended to be used, for atrial fibrillation, reduction of stroke and systemic embolism. I think what's different, with regards to our AVERROES in our trial is that our patient population was older than the Artesia trial, and so it's a bit difficult to compare across trials in this way.

Our trials show that there was an increase in the risk of bleeding with apixaban compared to aspirin, whereas AVERROES didn't. Because that could potentially be explained by some of the differences in the population that was under study. We found similar results with NOAH-AFNET the use of a DOAC compared to placebo or aspirin reduced the risk of stroke and systemic embolism.

And there was an increase in the risk of bleeding. I should mention that in the Artesia study that we did, the risk of stroke and systemic embolism was decreased by 37%, which is important. And actually there was a 45% reduction in the risk of disabling or fatal stroke.

So, these were important, clinically significant strokes that were prevented. The increase in the risk of bleeding, was, a risk of 1.7 in the apixaban group versus 0.94% in the aspirin group. So a small absolute increase in the risk of major bleeding. So one of the insights that we had from this study is that it does matter what the bleeding events are and one of the maybe provocative things that we, proposed in the discussion of our study was that. It seems like the strokes that were prevented in Artesia with apixaban may have been, more severe than the bleeds that were caused by apixaban in this trial. And just highlights the importance of us doing further research to understand now we have the prognosis of these events. So that's kind of our next task is to understand, well, what happened to patients who had major bleeding, but right. But because we saw that, the majority of these bleeds were, were not clinical emergencies.

The majority of bleeds were characterized by a drop in hemoglobin, and that there were similar rates of intracranial and fatal bleeding events between the apixaban and aspirin arm. But again, we saw an important reduction in disabling stroke made us wonder whether or not, that is in fact the case and that this is something that we need to be thinking about when we're doing these trials is not just the absolute risk of an event and not just the hazard ratio, that characterizes the association, but actually what are the consequences of these events and how do they compare?

This is a good segue to my next question. Well, considering the overall balance, a reduction in the stroke and systemic embolism on one side and the increased major bleed on the other, what do you see as the key clinical takeaway for physicians managing patients with subclinical AF?

It's a good question because we, again, when Artesia came out, there was this conflicting results where you had, on one hand you see a benefit, and this is the case in a lot of our trials, when people take anticoagulants, they reduce the risk of thromboembolism. When, you take anticoagulants, it increases the risk of bleeding. So to me, the question here clinically is whether or not we can see a net benefit. I think it will be helpful for us to have data regarding the prognosis of these bleeds to confirm our hypothesis that there is benefit. And perhaps that's not a benefit for all patients, but thinking about patients, and their risk profile for both of those, might be helpful. So there are characteristics that are being uncovered in terms of the risk of stroke among these patients and certainly novel ways of understanding stroke risk is ongoing, but the same side for bleeding, right? So these results may apply differently to patients based on their bleeding risk.

If you have a patient who's very elderly with renal function, renal impairment, and prior bleeding and anemia, and cancer, those characteristics, might sway you in one direction or another. I don't think that these results are ready to be implemented in prime time in clinical practice, but certainly understanding what these events are for patients is helpful 

That means that both physicians and patients can be better informed about the potential consequences of decisions. So that's how I would apply these to practice. 

Perhaps individualized risk assessment. Yes, exactly. And consultation between the physician and the patient as well.

Yeah. And acknowledging uncertainty. But I still think a better understanding of the consequences of different events helps people understand the decision in a different way. And so it's actually something that in thrombosis medicine, I think we, we've done very well in understanding the risk of thromboembolism for  stroke and systemic embolism and AFib, and for venous thromboembolism. But we, I think we're still uncovering the characteristics, the severity, the prognosis associated with different types of bleeding events. And I think more information about that will help us make better decisions for our patients.

So is there anything that we haven't discussed today that you'd like to bring up? No, thank you for, those great questions.

Dr. Siegel, thank you for joining us and for helping us better understand not only the frequency of bleeding with apixaban in subclinical atrial fibrillation, but also the nature and severity of those events and how that should inform clinical decision making.

And to our listeners, you'll find a link to the full Artesia sub-analysis in the show notes, and we always appreciate hearing from our listeners. If you have any comments, questions, or ideas for future episodes. Please email us at info@thrombosiscanada.ca and be sure to follow or subscribe so you never miss an episode.

And visit ThrombosisCanada.ca for clinical tools, patient resources, and information on our educational programs.

Before we close, if you value this podcast and the work of Thrombosis Canada, please consider becoming a Thrombosis champion with a monthly gift for as little as $10 a month. Your support helps us to continue delivering high quality evidence-based education for healthcare professionals and patients across Canada.

Thank you everybody for listening, and we'll see you next time on CLOT Conversations.