DOAC in AF + One Risk Factor: A Choice, Not a Prescription

This transcript has been edited for clarity. John M. Mandrola, MD: Hi, everyone. This is John Mandrola from theheart.org | Medscape Cardiology, and I am excited to have here my friend, Dr Bogdan Enache, who is a clinical electrophysiologist in Romania. He takes care of patients and is one of the sharpest minds on social media. He has a degree from the London School of Economics.We're going to talk about the SINGLE-AF trial, which is a really important trial, that was presented at the European Society of Cardiology and published in The New England Journal of Medicine. SINGLE-AF looked at patients who have one risk factor and atrial fibrillation (AF), and asked whether we should anticoagulate these patients.It was a trial of anticoagulation vs no anticoagulation, looking at a primary endpoint mostly of stroke, systemic embolism, major bleeding, and cardiovascular death. The results are provocative. Bogdan, what do you think about the study question, first of all? Bogdan Enache, MD, MSc, PhD: This trial was needed because we have these patients called intermediate-risk, and everyday clinicians don't know whether to anticoagulate these patients or not.What is the balance between stroke risk and bleeding risk? Do these patients with modern treatment actually have a significant risk for ischemic stroke?Mandrola: The evidence base for treating AF patients with anticoagulation really stems from older trials in patients with higher CHA ₂ DS ₂ -VASc scores. For patients who have only one risk factor, the yearly stroke risk may be lower, and so the net benefit [of stroke reduction vs bleeding increase] might be different.Enache: Correct, because the CHA ₂ DS ₂ -VA and CHA ₂ DS ₂ -VASc scores were established epidemiologically a long time ago when stroke rates were higher. For contemporary patients, who are diagnosed with AF sooner and receive today’s treatments, we don't really know their stroke rates, but we still use the CHA ₂ DS ₂ -VASc score.SINGLE-AF Trial PatientsMandrola: The trial enrolled 1800 patients from multiple centers in South Korea. They were 60 years old on average, mostly male (76%). These were slim patients by US standards, BMI 25 on average; 72% of patients had paroxysmal AF and about a third had persistent AF. Their HAS-BLED scores were very low, so these were also low-bleeding-risk patients. Enache: I also found it interesting that the median time to randomization from their first diagnosis of AF was 21 months — almost 2 years. Twenty percent of them already had an electrical cardioversion and 30% of them had a previous ablation for AF, because this trial, of course, is in conversation with the recent trials asking if we should stop anticoagulation after AF ablation.Mandrola: It's good for external validity because these are representative of the patients we see, right? They've had cardioversions, they've had ablation, and these are the patients that we see in clinic.Now the results. After approximately 2 years of follow-up, a primary endpoint occurred in four of the 900 patients in the DOAC [direct oral anticoagulant] arm; that's 0.5% vs 1.5%. So the absolute risk difference was -1.0%, the confidence intervals go from -2 to -0.1, and the P value is.03. Stroke events drove the difference: three vs 10 events. Major bleeding was rare in both groups (three vs four). Although stroke drove the difference, it was very few events.Enache: They did manage to show statistical significance. But in absolute terms, if you calculate the number needed to treat (NNT), we would need to treat about 100 patients for 2 years with anticoagulants to prevent one ischemic stroke at, if I'm not mistaken, the risk of two non-major bleeds, or one to one.Mandrola: We really want to know about major bleeding, and that was three vs four. There was a 0.5% annual stroke risk in the control arm vs 0.15% in the DOAC arm. And so the NNT per year is 200-250 patients. Enache: Correct. Mandrola: That's a really small effect size. And as far as statistical robustness, you hardly even need a calculator if you have three vs 10 events in a trial of 1800 patients. That could be noise.I calculated the Fragility Index Enache: So did I.Mandrola: I got two events. Did you get two events? Enache: No; I got one, actually.Mandrola: Your calculator gave you one event and my calculator gave me two events; that could flip the difference to nonsignificance. That makes it hard to make conclusions about what to do in clinic with these patients. Do you agree? Patient Choice or Guideline Recommendation?Enache: I agree. It's hard to use this data to make clear recommendations. I'm not even talking about guideline change. However, I can use this data every day when I see patients like this: men with CHA ₂ DS ₂ -VASc score of 1 or women with CHA ₂ DS ₂ -VASc score of 2, because that’s who we're talking about. I explain, "We don't really know the balance between ischemic stroke and bleeding in your case. We do know that if you take blood thinners, your already low risk for ischemic stroke will drop lower, and this is supported by SINGLE-AF. However, the choice depends on your lifestyle, your fears, your values. You might be afraid, let's say, of bruises or bleeding, or just not want to take an extra drug every day."Now I'll be able to talk more specifically, of course, with the caveat of a large NNT and so on. But now I can give them some recent numbers. We all have patients who say, "Oh, I had a stroke in the family and I'm traumatized by this. I've seen my grandmother bedridden for years," or something like that. Others say, "I'm very active. I ride my bike. I do ski jumping, boxing — and whatever contact sports — and I am not willing to take anything that would make me bleed or bruise easily." Mandrola: We know from the NOAH and ARTESIA trials of subclinical AF and anticoagulation that subclinical AF or shorter-duration AF — say, 2-4 hours on a smartwatch — found a yearly stroke risk of only 1%, even though those trials enrolled patients who had a CHA ₂ DS ₂ -VASc score of 4. So subclinical AF looks somewhat lower-risk. Let's consider two patients who could be in the SINGLE-AF trial: One is a 50-year-old with hypertension who has 4 hours of AF on their smartwatch vs a 65-year-old insulin-dependent diabetic with persistent AF. I would see those two patients as having a different stroke risk and different risk-benefit balance, but they both would qualify for SINGLE-AF because they have one risk factor. Enache: Correct. This is one of the limitations of the CHA ₂ DS ₂ -VASc score, right? It does not cover all the risks that we now know are associated with stroke in AF. We should judge it case by case. Mandrola: It's almost cliché to say that we should have our patients help us by sharing their values and preferences, because that makes a difference.Would you agree that this data, given the effect size, is not statistically robust enough to create a guideline or quality measure telling us that prescribing anticoagulation is something we should do in these patients?I think it's something that we can discuss, and different patients, different scenarios, might warrant different choices. Enache: What does it actually mean to "inform guidelines"? Should we change the recommendation? No. Should it be part of the discussion section in the guidelines — you know, the section that nobody reads anymore because they just read the colored boxes, which makes me sad.I think it should be part of the conversation, within the guidelines, just as it should be part of the conversation with our patients. And under no circumstances do I think it should be a quality measure, which is a topic for another day but which I'm basically against.Mandrola: Me as well. So awesome. Great discussion. Thanks for coming. Enache: Thank you, John. Thank you very much. John Mandrola practices cardiac electrophysiology in Louisville, Kentucky, and is a writer and podcaster for Medscape. He espouses a conservative approach to medical practice. He participates in clinical research and writes often about the state of medical evidence.