Good morning, everyone. Appreciate those who are here for the early morning. I'm Michael Gold. My co-chair is Ben Steinberg. And we are going to be talking about, I think, a very interesting topic of subclinical atrial fibrillation on multiple fronts. And where we're going to first start is to talk a little bit about the sub-studies from these two major trials, the ARTESIA trial and the NOAH trial. And Jeff Healy will be the first speaker as a PI of the ARTESIA study. While he's getting his slides up, if we can remind everybody to download the app. There's some audience participation questions, and that's obviously also another venue to submit your questions for the speakers. All right. Thanks, Michael, for the kind introduction. As Michael said, thanks to everybody who's made it to this early morning session. So with that, by way of introduction, I'm going to give you some insights that I have taken from the ARTESIA trial here. You can scan the QR code for your questions and give you my declarations of interest here on this slide. So to start at a high level, ARTESIA was this large randomized trial looking at the use of apixaban for patients with subclinical atrial fibrillation between six minutes and 24 hours. And you can see the main results of the trial here. Looking at the outcome of stroke and systemic embolism, this was reduced in relative terms by 37% with the use of apixaban compared to aspirin. You note on the right the relatively small event rates here, which were similar to what we saw in the ASSERT trial with a aspirin treated rate of only 1.24% per year. So low risk, yes, but significant reduction in stroke with the use of direct oral anticoagulant. Now this slide gives two key pieces of information. First of all, to the surprise of many, the strokes that were prevented using apixaban included a fair proportion, about 40 plus percent, of very large strokes. So the modified rank in three to six ranges from moderate permanent disability all the way up to fatal strokes. So this was surprising that such a large proportion of the strokes in this trial were of such severity. And the severe strokes, they're slightly more reduced at 49% reduction with the anticoagulant. Again, there's always a price to pay in life, and here we see a 36% increase in the risk of major bleeding going up to a rate of 1.5% per year with the anticoagulant. So at the top level, people have struggled over the last year to make sense of this. Now my co-presenter will present the data from the NOAA trial. Here just showing the summary meta-analysis that was published in 2023, just before the end of the year, showing that the results of the trials, although seemingly different at the bottom line, really is purely a matter of statistical power, and you can see here with these forest plots for both the outcome of ischemic stroke and for the NOAA composite there below, that those confidence intervals for the two trials widely overlap. The statistical assessment of heterogeneity of the I-square is 0% in both cases, suggesting that the trials are in fact displaying the same result. The pooled estimates are both significant for stroke and for the NOAA composite, and with approximately two-thirds of all the events in this analysis coming from Artesia, and one-third coming from the NOAA trial. So here's the polling question. We've got a 66-year-old woman, no history of bleeding, normal renal function, so we'll make this easy. She has a history of ischemic stroke five years ago, also has hypertension and coronary disease. She has a dual-chamber pacemaker with two episodes of subclinical AF, one lasting 30 minutes, one lasting two hours. What do you do? She's already on aspirin. Do you keep her on aspirin, or do you switch to a DOAC? Do people have the poll ready to go? Let's see. We'll have a small sample size here, so every vote counts. We've got 10, 9, 8, 7. It's going to kick us out here. Okay. So either one person voted, or there's unanimity in the room. So it's switched to DOAC, so let's unpack this. So the guidelines here, 2020. I'll just give you the 2020, because I want to show a little bit of perspective. We've always been on this idea that there's going to be an intersection of some measure of burden here on the X-axis and some measure of cardiovascular risk, and this is where, you know, I think John Lukabota first presented this type of proposal back in the early 2000s. There might be others. It came, many good ideas come from multiple sources at the same time, but you can see here the red zone where in 2020 we thought we would treat netted out around stroke risks of around 1.28 to 1.79 percent per year, and as we all know, there's been some shift in thinking over the last four years between this guideline and the last, but I throw this up because it's interesting to challenge ourselves to think about, you know, why our goalposts seem to have shifted in the last four years. Polled analysis, so we look at it at a high level. Why do we start with CHADS-VASc? Because this is how we manage atrial fibrillation, right? Do you anticoagulate every person you see with atrial fibrillation? Of course, the answer is no. We largely use CHADS-VASc to define a group with a very low risk, less than 1 percent or, you know, nowadays with shared decision making somewhere in the 1 to 2 percent range, and you can see here, I'll focus your attention to the right curves, and I'll preface it by saying this is a subgroup analysis of a positive trial, okay? So the trial showed a result, so we're not fishing for little groups of patients where there was benefit where the overall trial was neutral. Those would be considered hypothesis generating. Here our main goal is to look at the absolute risk reduction, so where do patients get the most benefit? And you can see here for stroke and systemic embolism, the reduction over 3.5 years for the CHADS-VASc greater than 4 group is about 4 percent absolute, and you can see, you know, for the interaction p-value, you can see a .06, suggesting that as CHADS-VASc goes up, the absolute bang for your buck goes up as well. The left panels show the relative risk. They're actually not that important because, again, the overall trial result is that there's a reduction, so we're not looking for the group that benefits, you know, the interpretation of that statistically is all of those groups seem to have the same relative benefit. It's just the higher CHADS-VASc patients have more baseline risk, and so you get bigger effects. The bleeding, on the other hand, you can see, just eyeballing it, is pretty, you know, it doesn't change much across the risk profile. So if you want to put this in more digestible terms, the CHADS-VASc greater than 4 defines about 27 percent of the Artesia population, a number needed to treat to prevent a stroke over 3.5 years of 25, a number needed to harm to cause a major bleed of about double, and again, the strokes tend to be more severe than the bleed. So that looks like a good place to treat. On the other hand, I think this analysis is just as useful for showing who doesn't benefit, and when the CHADS-VASc is less than 4, that's about 40 percent of people, there's really nothing to be gained there. The NNT is shockingly high, and the number needed to harm is under 100. So, you know, the CHADS-VASc 4 is a little more tricky to understand, but I think, you know, at the end of the day, we've kind of found an edge to treatment, you know, with this lower CHADS-VASc group, there's clearly no benefit, with the higher, there seems to be something going on. Quickly, this was recently published a couple months ago, looking at patients with prior stroke and embolism. One of the reasons the event rate was a little bit lower in Artesia than in Assert is we didn't enroll these patients as much, right? We had about 18 percent patients in Assert with prior stroke and embolism, here it was under 10 percent. Why? Because the number of patients were screened for this trial, people didn't want to enroll patients with a stroke, but they did enroll enough here, that's about 330 patients with prior stroke, and you can see there, their absolute risk reduction is about 6 percent over the duration of the trial, 60 percent relative, with a number needed to treat to prevent one stroke of 16 patients here, with the no prior stroke patients barely, you know, meeting significance, but with much lower event rates. So, you know, in the prior stroke subgroup, that's about a 3 percent per year event rate, over 3 percent, in fact, so it's quite high, and on the bottom, you can see those severe and disabling strokes, the rank in 3 to 6, those were reduced by 74 percent in the prior stroke subgroup, so we were preventing strokes in a much bigger way, and preventing big strokes. Just to show the analysis, this was part of the supplement for the Lancet paper, you can see that, again, the same story, that the NOAA results and the Arteza results are statistically similar, and when pooled, come out with this 57 percent reduction in stroke. Lastly, this issue of vascular disease, this is a joint paper between the two studies, it shows that the patients with a history of vascular disease, either peripheral, coronary, or cerebral vascular, have a much higher event rate, and there seems to be this trend, it's not significant, but a trend to greater reduction in stroke and embolism with those patients that have vascular disease. The big story here is on the bleeding side, of course, all the patients with vascular disease are already on aspirin, so as you know, the NOAA trial could have aspirin or placebo in the control arm, and so if you're already on an anti-platelet agent, the incremental risk of bleeding when you switch to an anticoagulant is much less than switching from nothing. Take it with a grain of salt, however, there's a huge degree of overlap, so in the Arteza paper, those with a chance of VASC greater than four, 72 percent had a history of vascular disease, in this definition, 20 percent had a history of stroke of that subgroup, so we're really talking about mostly the same patients, right, 90 percent or so of these patients are exactly the same as the high-chance VASC patients, so these are where I'd net out, you know, chance of VASC greater than four, the event rate's over 2 percent per year, that looks like a good group to treat, the prior stroke, 3 percent per year, looks like a good group, the vascular disease, it's also looking good at 1.9 percent per year, and this is, you know, where I would propose the treatment should lie, where it doesn't lie is burden, at least in that range between six minutes and 24 hours, we couldn't measure any appreciable difference, even in the event rates, let alone the relative risk, so that's where I would net out, so overall the Artesia trial showed a significant result on both stroke reduction and increasing bleeding, the net benefit seems to be best in the patient groups that we've identified here, burden, at least in the way we've analyzed it so far, doesn't seem to work, and I think you can, using the subgroup analysis, make clear decisions on about 60 or 70 percent of patients where the risk benefits are much more clear based on subgroups, so thank you, and I'll pass it to my next speaker. I just want to make one comment while you're switching, interestingly in the, you know, in the screening trials, screening for AFib and treating with anticoagulation was a benefit in patients with vascular disease, and not a benefit in patients without. Thank you, Dr. Glasser and Dr. Healy, our next speaker, we'll move on is Dr. Blomstrom-Lundquist, giving us insights from the NOAH trial. Thank you, my disclosures. So I'm going to discuss with you the substandards of the NOAH trial and my disclosures again, and before doing that, I would like to remind you that atrial high rate episodes, or device-detected AF, is common. It's present in about 10 to 30 percent of patients with cardiovascular disease, as we've seen in the ASER trial and in the LOOP trial, using continuous monitoring. We should also remember ourselves that they have an increased risk for clinical atrial fibrillation and an increased risk of cardiovascular stroke that is, of course, lower than clinical atrial fibrillation. And going back again to the meta-analysis of the NOAH and TCIA trial, anticoagulation did show that there was a reduction in all-cause stroke or systemic embolism, and fatal stroke was lowered by 49 percent at the expense of increase in major bleeding, but there was no difference in fatal bleeding, and that is important to recognize. So these are the four sub-studies of NOAH trial. Which device-detected AF patients would benefit from DUAC? Would it be those with episodes of longer than 24 hours, those with prior stroke, those with increased number of comorbidities reaching a chance of a score above four, or in the combined analysis, would it be patients with vascular disease? And before going into these different studies, I have questions for you. Which patient category would you think has the strongest indication for anticoagulation? Is it one patient with episodes of 24 hours duration, patients with prior stroke, or patients with vascular disease, or patients with many cardiovascular comorbidities and high age chance to ask about four? So please vote. though the vote would be for those with prior stroke. So let's look at the sub-studies. The first one is about the subgroup of low or high rate episodes exceeding 24 hours. 11% of these patients were present in the NOAA trial. And if we look at the number of our episodes, the baseline on the left side, we can see that there were quite few, not so many, in both the low group and the high group. And if we look at the longer single episodes, they were quite short. The median duration of longest R.A. was 53 hours in patients with R.A. above 24 hours and 2.2 hours in those with a shorter duration R.A. Primary efficacy outcome, composite of stroke, systemic embolism or death showed no interaction between allocated treatment and R.A. duration, meaning that patients with long R.A. above 24 hours did not benefit from anticoagulation therapy. And the same was true for the ischemic stroke. And if you see that the ischemic stroke rates were quite low, even in patients with long R.A. duration. And if we look at safety outcome, R.A. duration did not interact with safety. Of course, we had an increased bleeding rate in those on anticoagulation therapy as compared to those off in both subgroups. More ECG diagnosed AF occurred over time in patients with long R.A., which is important because we have to follow these patients. 17% per patient year versus half, 8.2% per patient year. So the next study compared those with and those without prior stroke. There were 10% with prior stroke in the NOAA trial. And if we look at the primary outcome of prior stroke or TIA, oral anticoagulation did not interact with the presence of primary ficus events. So there was no benefit for those with prior stroke with anticoagulation. And the same result was also achieved for the ischemic stroke, no interaction here either. We can see that the rate of stroke was lower than expected in those with prior stroke. And the severity of strokes did not differ between those with or without DUAC. There was no interaction of the treatment with prior stroke or TIA with regard to safety, major bleeding, or all-cause of death major bleeding. Major bleeding occurred more often in patients receiving anticoagulation, of course. And there was a high rate of safety events with DUAC in patients with prior stroke, which was a bit unexpected perhaps. So conclusions made in this trial was there was no treatment interactions with prior stroke or TIA for any of the primary or secondary outcomes. The rate of stroke was lower than expected, and there was a high rate of safety events with DUAC in patients with prior stroke. So the third study that included patients with or without CHAT score above four, there were 29% of these patients in the NOAA trial. And as we can see, the event rates on the y-axis and the effects of anticoagulation in each CHAT-SWASC score category on the x-axis, and we have stroke and ischemic stroke on the left side and the bleeding events on the right side. And the fixed and safe outcomes rates increased with increasing CHAT-SWASC scores, but without treatment interaction. So these patients did not benefit of oral anticoagulation according to the study. These are the major predictors for stroke and major predictors of bleeding. So concluding remarks in this study, no support that oral anticoagulation is more effective in patients with high CHAT-SWASC score above four. Stroke rate was quite low in patients with high CHAT-SWASC score over four without DUAC. Reasons for low stroke rate may be that the patients may have better treatment of comorbidities if you detect them and treat early. Frequent careful ECG assessment for AF to switch over to treat with oral anticoagulation, and a low arrhythmia burden. So the last group was a combined analysis of NOAA and Artesia patients that you have already heard of just recently with regard to with and without vascular disease. And vascular disease was defined as the presence of stroke or myocardial infarction or peripheral vascular disease or revascularization of ischemic heart disease. So we could see that the composite of stroke, systemic embolism, MI, and cardiovascular death tended to show that DUAC did do a benefit as compared without in patients with vascular disease, but there was no difference in those without vascular disease and the p-value for interaction almost reached statistical significance. And the same observation was also made for stroke. There was a tendency for DUAC to be effective in those with vascular disease, but there was no difference in those without vascular disease and again, p-interaction was 0.11, not significant. And with regard to major bleeding, DUAC equally increased major bleeding in both groups and the same observation could be made for the combined major bleeding and death. So the patients with vascular disease had a high rate of thromboembolic events as compared to those without vascular disease and the major findings is that the anostroke risk exceeded the threshold for oral anticoagulation therapy in those with vascular disease, but not in those without. DUAC increased major bleedings in both groups and DUAC may have a greater absolute benefit in these patients with vascular disease, but those without should probably not have any antithrombotic therapy. So my conclusion in weighing risk-benefit is in the stroke reduction against bleeding in device-detected AF patients with vascular disease is that they may well benefit. They prevent, DUAC prevent 3.5 strokes as compared to three in the ordinary general RA population and at the expense of induction of 1.7 major, but not fatal bleedings as compared to seven in the general population. So that will be my last slide and my conclusions. Thank you. And while we're getting ready to change over to that, there was actually an interesting question relevant to the two quick talks and I just want to give hopefully a 30-second response to our two speakers already. And the question was, are all CHADS-VASc scores of four the same? Is the history of stroke and age over 75 the same as diabetes, hypertension, heart failure, and age over 65? So the short answer is no. And I think this might explain part of the reason why the CHADS-specific stroke rates were different than we predicted based on the observational studies. We under-enrolled prior stroke, which is a higher two points than say well-controlled hypertension and well-controlled diabetes. Good. All right. And we're gonna now give a series of case presentations. Rob Passman will go first from Northwestern. So thank you. I'm going to present a case that I saw just a few weeks ago of an 83-year-old female with a history of hypertension. She had a pacemaker for sinus node dysfunction. She has an unsteady gait. She had two mechanical falls, one over a rug in her home and the other on the ice during spring in Chicago. And she had a recent, I would say, moderate GI bleed with an endoscopy showed nothing. So I'm going to remind you to scan the QR code if you haven't already and I have a question coming up. So this was her device download. And if I could draw your attention to the far left, you see that since October of 2024, she had these kind of small episodes of AFib, I would say less than an hour. And I was sort of just watching those. But then her last device download just about a month ago showed this about six-hour episode of atrial fibrillation. Of course, the first thing you do is never trust the device. Take a look at the EGMs yourself, but there's no doubt in this case that it was atrial fibrillation. And the patient, when we called her, had no symptoms during any of these episodes. So I have a lot of options here, maybe more options than there are audience members, but let me just ask you what you would do in this next case. Would you give her a DOAC? Would you say that you can't get a DOAC and you should occlude her left atrial appendage? Would you wait for longer episodes? Would you do pill-in-pocket DOAC, only around the time of a multi-hour episodes of AFib? Would you try to control her rhythm, hoping that we could maybe reduce her stroke risk? Or would you do nothing now or ever? So let's see. I know there's lots of choices, a small N for each, but let's see if we could have a consensus here. Can you remind us our history? Oh yeah, 83-year-old with hypertension, six-hour episode of AFib, but she's had two mechanical falls and a recent GI bleed of unknown source. Oh, okay. Okay, so about a third said wait for longer episodes, 20% said a clue to left atrial appendage, another 20% said do nothing. So not surprisingly, we are somewhat without a consensus. But let me sort of highlight some data that I think is important in the decision making here. So if you look at the guidelines, our guidelines say that a class 1 indication for clinical atrial fibrillation is a risk of thromboembolic events of 2% per year or more, 2% per year. Let's keep that number in mind. And it was elucidated earlier, if you have a risk of 1% to 2%, that you get a 2A indication. So this was a study that we did, and the lead author is sitting in the audience, Rachel Kaplan, who was a student and resident and fellow with us. And I think that this paper gets cited a lot because it gives us a lot of insights into telling a patient what we believe their risk to be. I want to give you some background here. What Rachel did was combine two very large databases. One was the Medtronic Care Link database that had device data on a few hundred thousand patients with dual chamber devices. And she combined that with the Optum database that had clinical information on more than 11 million Americans with cardiovascular disease. And from that data, we could plot who had atrial fibrillation. This is between six minutes, sorry, no AFib, six minutes to 23.5 hours, and greater than 23.5 hours. And we knew their CHADS-VASc scores, and we knew who had a stroke. And we could tell what their yearly risk of stroke is, and we used as a threshold here sort of a 1% per year risk of stroke, which some have argued should be the tipping point in the DRAC era. And the point is that if you have a low CHADS-VASc score, 0 or 1, no amount of AFib puts you over that 1% per year threshold. Conversely, if you have a CHADS-VASc score of 5 or more, even the absence of atrial fibrillation doesn't put you less than 1%. Now, we don't know if these are AFib-related strokes or sensitive to anticoagulation. But in this range of 2, 3, or 4, duration matters. Those with a CHADS-VASc score of 2 needed to have episodes of AFib pretty much 24 hours or longer to increase the risk. And those with a CHADS-VASc score of 3 or 4, even episodes as short as 6 minutes could get you over that 1% per year threshold. But remember that here, this median duration in this 3 to 4, between 6 minutes and 23.5 hours, was measured in hours. We're not talking about minutes of AFib. We're talking about hours and hours of atrial fibrillation in sort of that high intermediate risk stroke patient. So you heard the meta-analysis data, but remember that 2% per year number. Because in the NOAA trial, the median duration of AFib was 2.8 hours. The stroke rate in the control group with a mean CHADS-VASc score of 3.9 was only 1.2%. In Artesia, the median AF duration was 1.5 hours, and 80% had episodes less than 6 hours. But here, the stroke rate in the control arm was 1.1%, despite a CHADS-VASc score of 4. So one of the take-home messages I hope to convey is that everything we know about stroke in AFib is based on clinical atrial fibrillation. This is a different animal altogether. And the CHADS-VASc score that was derived in those with clinical atrial fibrillation does not apply to these individuals. Their risk of stroke in study after study is probably half, if not more than half lower than what you might expect based on their CHADS-VASc score. The other point to emphasize is that on anticoagulation, in the meta-analysis, the relative risk reduction in stroke was about 32%. What is it in the general AF population with clinical AFib? It's more than double that. So you have a lower stroke risk to begin with. You have a lower response to oral anticoagulation, and yet you still have a 60% bleed risk increase. So you have lots of harm with, in my opinion, less benefit. So we have to tell this patient that you would have a 0.4 relative risk reduction in ischemic stroke per year, but an increase in major bleeds of 0.7%. So for those of you who said, let's put in a left atrial appendage occlusion device, this is a meta-analysis of three trials enrolling 933 patients, mean follow-up of 39 months, with no significant difference in stroke or systemic embolism rates or major bleeding between the groups, but a significant reduction in hemorrhagic stroke and all-cause mortality. That's important. But there are downsides here. Right? The current overall rate of acute complications from left atrial appendage closure is 2.2%. And in a meta-analysis of over 10,000 patients, the pooled incidence of DRT was 3.8%. And if you have DRT, you have a fourfold increased risk of stroke. So I think that this is an interesting patient. There are a lot of directions you could go. The easy thing would be to say, let's do shared decision-making. But I'm not sure I could potentially explain to this individual what the benefits and risks are in a coherent manner that would allow this individual to make an informed decision for herself. So I'm eager to hear what the panel has to say. And I don't know whether we're going to discuss these cases at the end, or you want to discuss them now? Now? Sure. OK. Your turn. So I was surprised. I would put her on a beta blocker. She has a pacemaker. And I would, I mean, when you say rhythm control, that's what you said, right? I wouldn't put her on amiodarone. But I would certainly increase her beta blocker when she was in my office. So the point you're making is that maybe if we eliminate her a-fib burden altogether, that we could reduce her stroke risk. Yeah, that's my point. And is it safer to put her maybe even on amiodarone than an anticoagulant or a left atrial appendage occlusion? What other people on the panel think of rhythm control to reduce stroke risk in device-detected a-fib? Sounds like a good study, Jeff. I think it might be tenuous. I can't even imagine doing the study, because the event rates would be relatively low. I mean, I certainly wouldn't give her an anticoagulant. Rhythm control to prevent stroke risk, probably not. Rhythm control to prevent heart failure risk might be different, because that's much more common and plausible. But I don't think we're there yet. But I'd address her risk factors. I wouldn't do an atrial appendage closure. I wouldn't put her on an anticoagulant. I would just follow along to see if she develops, because they're going to develop clinical a-fib at a rate of somewhere between 4% and 8% per year. So that's clinical a-fib. But so let's talk about the guidelines say that if you're over 24 hours, right, that that's a different entity altogether. But you showed that 24 hours didn't seem to mean very much in this population. So for those of you who said, wait for longer episodes, is there a point where you would pull the trigger, right? If she's subclinical and it goes on for four days, is that a different individual that has a six-hour episode? Or does duration really not matter if it's still subclinical a-fib? You'll have to wait for clinical a-fib. So you should follow the patient and observer for that, I think. But is a 24-hour asymptomatic episode of a-fib any different from a 24-hour symptomatic episode? No, I wouldn't say so. It depends how close you are to an ECG machine. We define progression in Artesia two ways, 24 hours or a clinical method of ECG-detected a-fib. The overlap was about two-thirds. So if you had 24 hours, you know, about two-thirds of those patients also had an ECG. So there's a lot of overlap. Ron, can you clarify your comment about Dr. Healer's data not showing a difference at greater than 24 hour? My interpretation was there was no difference between six minutes and 24, but at greater than 24, you couldn't really conclude anything. We encouraged withdrawal from drug therapy, like study drug therapy, and use of open label. We did have a few people that didn't follow that guideline, and Giuseppe Boriati has a paper coming out on that. It's complex, but like I said, it's just there's a lot of overlap. Many of the greater than 24-hour subclinical a-fibs were in it for weeks. And that was based on the ASSERT study as well, and that cutoff was made. Right. So to me, it seems like... Right. But am I interpreting Dora correctly that when you look at those 11% that had greater than 24 hours, that there was no significant interaction between, you know, use of anticoagulation in that group? So is 24 hours the threshold? I think there's a lack of data in that group all told, right? I mean, 11%, remember the event rates, the number of events in NOAH was relatively modest, and then that 11% was even more modest. So it's hard to say. Even Isabel Van Gelder's paper, you know, it's not hundreds of events, it's dozens of events. So, you know, it's a good question. But like I said, I think, you know, that greater than 24-hour group comprises different types of patients, many of whom had weeks and weeks and weeks of atrial fibrillation. And, you know, I don't think that patient who's in AFib for five weeks is any different than someone with a 12-week ECG. No, that's true, but the number of episodes were quite few. We're going to need to get on to the next one, and I'm going to try to pepper in just a few of these quick audience responses or other responses while we're trying to get a change over here, which is one that's intriguing, because the two studies we've been talking about, one was Apixaban, which is a dominant anticoagulant, and the other is Indoxaban, which is hardly used at all in the United States, at least. So the question was, would you anticipate similar results with Rivaroxaban or Dabigatran? Anyone have any comments? The only comment is it's a relatively older population. There may be some nuance to Indoxaban bleed rates maybe being a little bit more favorable in the very elderly and some of the other DOACs, you know, certainly with Dabigatran we see a clear effect with age. So, you know, it's beyond the reaches of our studies, but, you know, there may be some subtle nuance. I don't think it changes the top line, though. But if you're going to medicate with Dronedarone, I think Indoxaban is a more favorable drug. Good. Correct. Okay. Well, next is going to be Rupi Sandhu from University of Calgary, who's going to give us the next case presentation. Great. Thank you. Well, we know that even in the most well-designed clinical trials that the translation to real-world practice is quite variable. And so I know that you've heard some of the sub-study data, but I really want you to answer these questions with what you would do in your clinical practice. So you have a 67-year-old male with a past medical history significant for hypertension, diabetes, and a prophylactic dual-chamber ICD in the setting of coronary artery disease on guideline-directed medical therapy and an EF of 0.25. He's seen in the hospital during an admission for syncope. And in that workup, they interrogate the device, and this is what they find. So you can see from the device interrogation that there is an episode of AF. It's about 59 minutes. It's not related to the time and date of the syncope event, so completely incidental. And so I would ask you, would you initiate oral anticoagulation therapy, a yes or no, just based on this presentation? So again, scan the QR code. It's not clicking in. Okay. But let's just do a show of hands. So based on that history, who would initiate oral anticoagulation? So probably about a quarter. Okay. So during the hospitalization, just like most places where somebody presents with syncope, they get quite the workup. So he's getting additional testing, an echo that's unchanged, carotid ultrasound that shows 30% stenosis of the right proximal RCA. It's a CT scan of the brain that shows a remote infarct in the inferior temporal lobe. So just in terms of what your management would be at this point, would you initiate oral anticoagulation here? So yes or no? And let's just see if this poll works. So you would call that a subclinical stroke? Like he never had a history of a stroke? Yeah. Didn't have symptoms. Had a subclinical stroke, exactly. A non-Lacuna, given the location. And maybe you're intentionally omitting anti-platelet. Is he anti-platelet regimen right now? Yes. Not currently information that's provided. Right. Subclinical stroke, the same as a stroke. Okay, so 91%. Okay, great. So we already heard about the sub-studies that were presented earlier on using NOAAFNet and previous history of stroke or TIA. Neutral study did not show that previous stroke increased your risk for future events, whereas in Artesia, we did see that a previous history of stroke did. So I want to ask the audience, what if the brain imaging demonstrated a remote lacunar infarct in the right basal ganglion and thalamus? So what I'm getting at here is this is a really cerebral small vessel disease, which is a marker of when you have lacunar infarcts, it's a marker of cerebral small vessel disease. So does the type of stroke matter? We saw some sub-study data talking about vascular disease in general, but that was a compilation of different vascular etiologies. So if they did not have a non-lacunar infarct and had this on their imaging, how many of you would initiate oral anticoagulation? So again, a yes or no. So still yes. I think I would ask the experts here whether there was enough information to really identify the type of stroke in terms of your future risk. Because we know that cerebral small vessel disease, only 25% are responsible for ischemic stroke. So how would that change your management? I actually think this is a really under-recognized issue. We tend to lump all strokes together in the EP world. That's like in the neurology world, lumping all arrhythmias together without any consideration of what type of arrhythmia. That's what Chad Vass did. Right, and that's what Chad Vass did. But in the stroke AF trial, when you put lupocords in patients with stroke from large vessel disease, their instance of AFib is the same as the general population, as the same as patients with cryptogenic stroke. Those curves are almost superimposable. So those strokes don't respond to anticoagulation, and I think we should view them differently. I think the other issue in looking at NOAA, as I recall, there are patients who had strokes decades before. I guarantee you, actually, those aren't due to AFib, because those patients never declared themselves as having atrial fibrillation. So we tend to lump stroke and TIA, and of course, TIA is a hard diagnosis to make retrospectively. But we should be considering the mechanism of the stroke. But is the reason to think that this is the patient that might be a particular risk for ICH on the anticoagulation you consider? Yes. Right, absolutely. Just to throw a little spice into the mix, I mean, first of all, we do a modest job of subtyping strokes. If you take a stroke adjudication panel for a large clinical trial, you get moderate agreement. It's not as cut and dry as you think it might be. Second, there's no reason to believe that an 82-year-old with hypertension with a lacunar stroke, that the next thing they have is an embolic stroke or an ischemic carotid atheroembolic stroke. So patients, when you get up into this multi-morbidity, they have stroke from a variety of mechanisms. The final point is around the covert stroke. We know that for every clinical stroke, when we do serial MRIs on people or screening MRIs in populations, we see about five to six times more covert stroke. They do predict future events. So there's no free pass if you just have a large cortical infarct just because you didn't have symptoms. It's landed on grade 10 calculus and you never use it, so you don't know it's gone. But they are prognostic. We use them in some clinical trials. The OCEAN trial, which is going to come out in the end of the year, uses it for AF ablation populations. So they're there. So it is quite a complex issue. But I think stroke does matter because it's the sum total for this outcome. It's marking all the things that lead to, you know, the person's had a stroke. At any time, you have a baseline variable that is also your outcome. What's the best predictor of heart failure? Having heart failure in the past. So whatever mix of things got you to heart failure, you've done it before. So they tend to be strong predictors in multiple disease states. Yeah. And I would just add that it's vascular disease. Whatever the stroke is, it's someone who you can say has vascular disease. You know, it's kind of cute, right? We're looking at subclinical AFib in patients with subclinical stroke and subclinical vascular disease. It's like we don't have enough to do, right? Yeah. But the reality is there are some events out there, and it's just a matter of, you know, the data you know versus the data you don't know. No. It's a great case. Yeah. No, no. Why don't we move on? Great. Thank you. Thank you. I'll have a quick question in between as we're turning you over, but we've got to get Talia's presentation done because she's got to go to a wedding, which is the truth. First, I'm going to have to argue with Jeff in the Hyde Park, so I might never get there. Just a quick question. Do you consider a wearable detected AF, a wearable device, as device detected or clinical AF when estimating stroke risk? Any comments on what we do about wearables? Well, I think wearables are not tuned to pick up minutes of AFib. They're going to pick up hours of atrial fibrillation. So if you get a notice from a wearable that you have AFib, A, you probably have more than you realize, and it's probably longer than you realize with commercially available algorithms. So obviously you need ECG confirmation before you consider treatment, but I think that those patients have higher burdens than what we're detecting with implantables. I think it's a shifting target, though. The devices are getting more comprehensive in their screening, and we've all had patients who make 1,000 live core ECGs per day, and they're getting pretty good coverage. Well, right. I guess I should distinguish those that are obsessive and record 1,000 cardiac tracings versus those that have a watch that alerts them that they have an irregular PPG. Okay, so I'm going to go through. I have two cases. The first one I'm going to go very quickly, if I can do this. Okay, so this is an 85-year-old male with heart failure, coronary disease, diabetes, hypertension, Chad's VASC of six. We see these episodes with these durations. We definitely look at the stored intercardiac electrogram, and the question is, would you treat? Just show of hands, yes, no. Yes. How many people would treat? Great. This is the same person, same exact case with a Chad's VASC of three. How many would treat? A few still would treat, okay, but we can talk about why. Then the Chad's VASC of four, same exact case. Again, Jeff already showed this data, but in patients with a Chad's VASC less than four, the number needed to treat to prevent a stroke is huge, and the number needed to harm is very few. If the Chad's VASC is greater than four, there's much more benefit in treatment. I have found this in my clinical practice to be incredibly helpful. We talked about 24 hours. We need something to hang our hats on. We need to know a cut point, so 24 hours has kind of become it, whether it's right or wrong. I would say in the subclinical device-detected AFib arena, this is something you could really hang your hat on. I would say if the Chad's VASC score is less than four, maybe don't treat. If it's four, that's difficult. If it's greater than four, you probably should treat, and if it's four, it's a little still unclear. That was that point that was made already, but this is the one that I think is kind of interesting. This is an 85-year-old male with heart failure, hypertension, coronary disease, Chad's VASC of five. You saw these episodes in the past, and they're on anticoagulant because their Chad's VASC score is high. Now you're seeing them in February of 2024, and they haven't had any AFib in the last nine months. Would you change your therapy? Would you stop the anticoagulation? They've had no bleeding events in the area. Well, that's a good question. I don't want to make it more complicated by discussing bleeding, but let's just say that this is what you have. They had a six-hour episode, a two-hour episode, a four-hour episode, and they haven't had any more. So would anybody stop their anticoagulation? Nobody. What is it the patient said? I want to stop my anticoagulation. I've not had AFib. Even if you assessed it in May 2023, it's going to be incredibly low. But if you extend that a year, we're talking about a burden in the .001 range. So if he said, what is the evidence that I'm benefiting from this, I think you'd be hard-pressed to say that you have a clear benefit. Well, I think there's one other consideration, which is his ischemic cardiomyopathy and whether his heart failure or general cardiologist thinks he benefits from antiplatelet therapy. So I certainly wouldn't have him on both. But if somebody's committed him to antiplatelet therapy and we could go with only a DOAC instead of antiplatelet therapy or instead of dual therapy, I think that's another consideration, which is his bleeding risk. Because there is some atherosclerotic benefit, I think, to DOACs. It will be an interesting discussion. I mean, we do have a nice collaboration and a planned patient-level meta-analysis of these two large studies. We struggle with burden, right? There's many task forces looking at burden. Where do you draw the line? But zero burden is pretty unambiguous. When you have a pacemaker and you've had nothing for a year, you're clearly in a different group of burden than the other person. And there's no ambiguity about line drawing there. So that would be one of the most interesting things is these regression patients, what happens? It's certainly a big part of the loop trial, different population, of course, but screening population. But there are these patients where it seems to go away, whether that's GDMT or something else. Yeah, I mean, 25% of the patients in the loop trial, the AFib went away. But this is data from Artesia, data showing that patients who had no AF in the last six months did not have benefit for anticoagulation. And only those who had any episodes in the last six months had benefit. So this is something else that I hang my hat on a little bit. And, you know, they still have their device. We're not going to take the pacemaker out. And we can set our alerts. And we do remote monitoring. And so the question is, if you keep them on anticoagulation, you increase their risk of bleeding. So I might consider stopping it and watching it based on this. That's what it was. You don't have to repeat it. That's it. It says I have three more slides because they're blank. Okay, good. So I think we still have a little bit of time. And we still have a few questions, I think some interesting ones from the audience. Because it's not quite significant. One is define clinical atrial fibrillation. Multiple definitions already. Is there any consensus from the audience of how we're defining clinical atrial fibrillation? I'll just say one thing that's a huge pet peeve of mine, and that's symptoms. Symptoms don't mean anything in terms of stroke risk, in my opinion. I don't know what other people think. But, I mean, just because you're asymptomatic, if you're in AFib, you're still in AFib and you still have a stroke risk. So when you say clinical AFib, to me it means an EKG of AFib. But it has nothing to do with symptoms. I agree, unless the symptoms are like a TIA or something. But it's a marker of burden is what you're saying. If you happen to have gotten an EKG, then you have to be in a longer amount of atrial fibrillation. If you're able to record AF on ECG, you would have enough burden to increase the stroke risk. That's it. Yeah. Symptoms are still in so many papers, and people still talk about it. And that's why it's a pet peeve, because... Yeah, we should get rid of subclinical. Well, I just... People say you're treating symptomatic AFib. I mean, all AFib should be thought of as AFib. I mean, I don't think... I think it's actually kind of interesting, our results. I mean, they make for great debates. So Karina and I get invited to all these things, and Tia as well. And it's great. But I think we've kind of hit the edge of treatment, right? Like, for the last 20 years, we've been doing trials in the sweet spot of anticoagulation for atrial fibrillation, where everybody seems to benefit. The event rates are shockingly high. And to the credit of all these individuals, we've knocked down the stroke rates substantially. Here, we're getting on the fringe of where, for some people, it matters. I think it's... You know, I like CHADS-VASc. It just shifts up a couple of points. It's not a 2-3 discussion. It's a, you know, a greater than 4 discussion. But again, I think the trials are much more cohesive than they at first blush seem to be. I mean, again, you know, 90% of the people with the CHADS-VASc greater than 4 in Artesia had either a history of stroke or a vascular disease as defined by NOAA. So, you know, we can talk about the 86-year-old woman with well-controlled hypertension, well-controlled diabetes. She's a CHADS-VASc-5. Maybe you treat her, maybe you don't, right? But I think, you know, I think there's more agreement than there is disagreement in the group of patients where the benefit seems to lie. Dr. Healy, I'm curious if your thoughts would... If any of you anticipate your thoughts changing with factor XIa inhibitors and possibly different risks there, benefits. Do you anticipate this body of evidence changing? We need efficacy first. Yeah, exactly. I think we had oceanic AF that had asyndexian. It terminated early, and maybe it was something to do with the dose that was used and the patient population that was enrolled. But we don't have any current evidence to suggest that the novel factors would be better. The bleeding risk seems lower. The bleeding risk is lower, and they seem to be good. And then the studies came out, and they basically flamed out. Right. The Phase II studies showed that bleeding was significantly reduced, but the efficacy still hasn't been proven. Right. But one could hypothesize that the efficacy would be better than nothing. Right. So, I mean, I think there's a lot of shifting. Like, we have to sometimes sit back and appreciate our own biases, right? We're influenced by you. We walked around this meeting. There have been big stage presentations and things getting in our heads, and it does shape the way we ask questions like this. So, if you look at the top line of Artesia, you know, there was about a half percent reduction in stroke with the treatment in the all-cause population. And so you could say, well, am I willing to walk away from half a percent of stroke risk for the sake of bleeding prevention? Right. You know, where does this come up? Well, we saw it in the first question. Watchmen, right? We can, you know, maybe unprotected prevail. We give up a little bit of ischemic stroke, but there's less bleeding. We talk about it with 11As, right? We saw the presentation of the first couple of phase three trials. You give up about half a percent stroke, but to save bleeding. So, I think this is partly why this, you know, we've had this shift, because for many years, you know, 1% was our line where we drew the line for initiation of anticoagulation, and that was with warfarin, right? The drugs, the DOACs are already better for bleeding compared to warfarin. When you look at the Artesia data compared to, say, Avaroas, in age-related patients, the bleeding risk we saw in Artesia was lower than Avaroas, largely because we didn't use open-label aspirin on top, which was very common 10 years ago. So, just keep in mind that, you know, the thresholds are shifting. We're much more focused on bleeding, but it's partly because we have the promise of safer approaches to stroke prevention. The last point I'll make is we really do need to address the patient perspective here. You know, what do patients actually care about? Patients tend to accept somewhere between four and five major bleeds to prevent a stroke, based on the data in the clinical AFib realm. And the physicians are more one-to-one when you ask physicians what their rating of stroke versus bleed. We see them equally. In fact, we're a little more afraid of causing bleeds, because that's on us. So, definitely shared decision-making is important, as we're learning. I'm going to ask one more question that was asked, which I think is an interesting one, which is how do we balance the data showing lack of benefit of oral anticoagulation in embolic stroke of unknown sources, ESIS? Yeah, that's a great question. And neurologists are very strong about this. They won't anticoagulate these patients. How do we balance that with only device-detected atrial fibrillation? So, we're talking about device-detected atrial fibrillation, doing things, and someone who's had an embolic stroke by a neurologist, they're not going to treat that, no matter what. If you wind the clock back to, say, CRYSTAL-AF, it was probably the biggest trial that was there. About one-third of patients with cryptogenic stroke, ESIS, had subclinical AFib. About half of that AFib that was detected was detected more than 18 months after the stroke, and we don't know what they looked like before the stroke. So, it's a little bit like a screening trial. You're treating with an anticoagulant knowing that only 30% of people have the condition you're targeting, and only half of those, so 15%, actually, might have it before they have their stroke. So, you dilute down your effect. Arcadia, we thought we could do better. Unfortunately, we missed the mark because we largely recruited patients based on BNP, which didn't identify a much better group. But ASSERT also, interesting thing in ASSERT was that there were as many AFibs post-stroke as there was pre-stroke. So, there were a bunch of patients in the ASSERT study who had no history. We had a year or more of monitoring them with no AFib, and then post-stroke, they developed AFib. So, there's no free lunch here with this stuff, for sure. It's complicated. And, you know, but the good news also is that the stroke rate has gone down a lot. Our anti-hypertensives and our anti-lipid drugs and, you know, we are doing, let's not get depressed, we're doing a good job as you, I think you made a very good point, that we're on some fringe now trying to really balance something that we may not be able to balance. Well, I think we'll probably, we'll need out of time, but I think this turned out to be a great session, and I appreciate all the discussion that we've had from everybody, both in the audience that sent questions as well as the panel here. So, I think we all learned a lot, and we all learned a lot more we have to learn. So, thank you everyone for attending. And I'll let... Thank you.

subclinical atrial fibrillation

ARTESIA trial

NOAH trial

apixaban

stroke prevention

anticoagulation

CHA2DS2-VASc score

bleeding risk

patient-centered care