false
Catalog
EP Fellows Curriculum: "Pill-in-Pocket" Anticoagul ...
EP Fellows Curriculum: "Pill-in-Pocket" Anticoagul ...
EP Fellows Curriculum: "Pill-in-Pocket" Anticoagulation for AF - Fact, Fiction, or Foolish?
Back to course
[Please upgrade your browser to play this video content]
Video Transcription
Good morning everyone. Thanks again for joining this morning. Today we have the privilege of hearing from one of my esteemed colleagues at Northwestern, Dr. Rod Passman. Rod is a professor of medicine and preventive cardiology. He's the director of our newly launched Arrhythmia Research Center. He's been an invaluable mentor to me throughout fellowship and then the beginning of my transition to attending here at Northwestern. So his research interests are in AFib management and ablation, digital health, and stroke prevention in AFib. Never seen him speak, you're in for a treat because it's always very entertaining and this is kind of a provocative topic. So the title of his talk is Pill-in-the-Pocket Anticoagulation for AFib, Fact, Fiction, or Foolish. Thank you Rod for being here and just so everyone knows he just gave grand rounds elsewhere so he's doing a doubleheader this morning. Thanks Rod. You're very welcome and I'm very excited to have this opportunity. That's one good thing about COVID is you give back-to-back grand rounds in different cities from the comfort of your basement. I purposely chose a provocative topic today and I hope that even though it's early in the morning this is sort of not going to be your typical review. I hope that we garner a lot of pushback and interest and let's take it from there. So I am going to be talking about this concept of pill-in-the-pocket anticoagulation that we've been working on and I'm going to begin with a case presentation because I'll tell you that regardless of what type of practice you have and how much research you do, I think all of us are at our core clinicians. We went into medicine because we want to take care of people and we want to help people and certainly in my career I remember my failures much more than I do my successes. So I'm going to tell you a failure. This is a patient who was sent to me in 2003 and not only was he sent to me but he was sent to me by one of my colleagues in my own cardiology division. This was a friend of the family and this colleague was going on to another job and his parting words to me were please take special care of this individual. So at the time the patient was 66 years of age, had a history of hypertension and diabetes, had highly symptomatic paroxysmal AFib. He was on an anti-rheumatic drug and anticoagulation. His symptoms resolved and every EKG and every Holter that I did for the subsequent three years showed normal sinus rhythm. So you know at that time I had two choices. I could continue him on anticoagulation indefinitely, thereby preventing stroke should he have recurrent AFib, but in doing so I would be exposing him to a dangerous drug that he may not be benefiting from. Or I could stop anticoagulation and therefore reduce the exposure to a dangerous drug but expose him to the risk of a stroke should he have recurrent atrial fibrillation. And my decision was easy then as it would be today because our guidelines are very clear. We don't make decisions regarding anticoagulation based on stroke risk. We make it based on CHADS-2 and CHADS-VASc score without any regard for AF type or duration and without any regard for whether someone is quote-unquote successfully rhythm controlled, whether it be with drugs or ablation. So I continued him on his Warfarin until he came in with his intracranial bleed with an INR of 2.3. I still see this patient. He is wheelchair bound, dysarthric, never worked again a day in his life as the lawyer that he was, and this event changed his life and the life of his extended family maybe for even generations. So why did I not stop anticoagulation on my patient? Well we have no cures for atrial fibrillation, at least in my institution, whether it be drugs or procedures. We understand a lot of AFib is asymptomatic, that when patients come to us and stating that we've cured them of their AFib because of the drug we put them on or a procedure that we do, we have to remember that a lot of our interventions convert symptomatic AFib into asymptomatic AFib. So patient symptoms are an unreliable marker about the burden of atrial fibrillation and the success of any intervention. We also are uncertain on the relationship between AF duration and stroke, and certainly in 2003 I had no ability to rapidly anticoagulate someone with an oral medication and had no way of monitoring this patient for long periods of time if they're sitting across the room or across the world. However, there are five reasons to consider stopping anticoagulation. Rhythm control does offer a chance for cure and there are some patients where we truly do cure them of their AFib. But even if we don't cure them, perhaps we can reduce the AF burden below a threshold with which the stroke risk is actually quite low. We also have to recognize that anticoagulation has risks, anticoagulation has costs, and number five should really be number one, that patients truly do not want to be on anticoagulation, particularly those in whom you're having them on a drug or ablated them and you've shown them that their AFib is minimized or altogether gone. But I would argue these first two are still quite true. However, we do have greater insights into the relationship between AF duration and stroke, and we do have the ability to rapidly anticoagulate someone and monitor someone over long periods of time. So where does this concept come from that it doesn't matter how much AFib you have? Well actually, it comes from the SPAF trials that were published more than 20 years ago. In SPAF, they found that patients with intermittent, or what we would call paroxysmal, or sustained or persistent AFib have a similar stroke risk regardless of whether they're low, moderate, or high-risk individuals. But remember that in the SPAF era, patients came into their doctor's office with ECG showing AFib, and therefore all of these patients probably had relatively high AF burdens. So our own guidelines suggest that there is a AF duration threshold below with which the risk of clot formation appears to be low. The 48-hour rule that talks about cardioversion within 48 hours of onset of AFib or AFlutter says that you don't need to do a TE or you don't need to anticoagulate someone for three to four weeks of therapeutic anticoagulation, suggesting that you need 48 hours of AFib for a blood clot to form and a stroke to occur. However, the era of implantable devices with leads in the atrium, or more recently the development of implantable cardiac monitors, suggests that 48-hour rule may not be correct. So for example, in this study of patients with dual-chamber pacemakers, right, those individuals who had 24 hours or more of atrial fibrillation had an increased risk of stroke, and those patients with less than 24 hours were as unlikely to have a stroke as patients with no AFib at all. So maybe that number is not 48 hours, maybe that number is 24 hours. Well, in the TRENDS trial, they took patients with dual-chamber pacemakers or defibrillators who had one or more stroke risk factors, and they defined the daily AT-AF burden as the longest total duration on any given day during a 30-day rolling window before the first thromboembolic event or the end of the follow-up period. For patients with any AT or AF, the median burden was five and a half hours per day. Look at the risk of thromboembolic events in patients with zero versus low AT-AF burden. It's exactly the same, and only those individuals with a high burden, i.e. greater than five and a half hours, had an elevated thromboembolic event rate compared to baseline. So maybe that number is five and a half hours. Well, in the ASSERT trial, they took patients with recently implanted dual-chamber devices who had one or more stroke risk factors and categorized them into one of two categories. Group one had at least six minutes of an atrial arrhythmia within the first three months of device implantation, and group two was everyone else. If you follow those patients who had at least six minutes of an atrial arrhythmia over two and a half years of follow-up, they had about a 2.49 relative risk of stroke, leading some to believe that all you need is six minutes or more of AFib to increase your stroke risk. However, I would urge you to look more closely at that data, because even though they drew the line in the sand at six minutes, when they looked at the amount of atrial fibrillation those individuals who had a stroke had, they found that only those individuals who had 24 hours or more of AFib had a stroke risk that was proportional to their CHADS VASc score, and those individuals with less than 24 hours were as unlikely to have a stroke as those with no atrial fibrillation at all. So I would actually think that there's not a clear AF duration that imparts a stroke risk equally to all individuals, right? And we wanted to evaluate the interaction between the AF duration, CHADS VASc score, and stroke risk. So what we did is we combined two big databases. One is the Optum database that had about 11 million patients with cardiovascular disease, and we combined that with the Medtronic Care Link database that had 1.4 million patients with two or three lead devices. And we had lots of inclusion and exclusion criteria. What we did is we plotted the CHADS VASc score at about 22,000 patients on the horizontal axis, and the maximum daily AF burden on the vertical axis. This is no atrial fibrillation, this is atrial fibrillation six minutes to 23.5 hours, and this is AFib greater than 23.5 hours. This paper was written by Rachel Kaplan, who is now one of our EP fellows and published a few months ago in Circulation. What we used as our threshold for anticoagulation was 1% per year stroke risk, which some have suggested should be the tipping point for initiating anticoagulation, particularly in the NOAC year. So with this in mind, an interesting pattern emerged. If you have a CHADS VASc score of 0 or 1, no amount of atrial fibrillation puts you over that 1% per year threshold, for which anticoagulation may be beneficial. For patients with a CHADS VASc score of 2, right, you needed 23.5 hours or more of AFib to increase risk, and patients with a CHADS VASc score of 3 or 4 who had no AFib were also at a reduced risk, but once they had six minutes or more of atrial fibrillation, moved into this high-risk category. Patients with a CHADS VASc score of 5 or more, no amount of atrial fibrillation reduced their risk of stroke. Even those without atrial fibrillation had a higher risk. What we don't know is, are these cardioembolic strokes, are these strokes responsive to anticoagulation? We don't know. But what I will tell you that, from a fellow's perspective, when you get a call from device clinic that Mrs. Jones had six hours of AFib on their pacemaker, the response shouldn't be start anticoagulation. The response should be, tell me about Mrs. Jones. Because if Mrs. Jones is in a lower risk category, right, six hours may be not enough to warrant anticoagulation. So we spoke about these first two things being true. We spoke about the greater clarity about the interaction between AF duration and stroke. We clearly have the ability to rapidly anticoagulate people. Obviously Coumadin would take several days to reach a therapeutic INR. The NOACs or DOACs can systemically anticoagulate someone within one to four hours of a single dose. We also have the way to rapidly or continuously monitor people. So if this were a live audience, I would ask who this is. But since it's not, let me tell you. This is Norman Holter. And what he's carrying is the first Holter monitor. Now Norman was not a physician. He was a tinkerer. And he was interested in designing a device that could monitor the rhythm of the heart outside the hospital setting. That was the good news. The bad news is the device weighed like 85 pounds, is shown here carried on his back, will only record 24 hours of the rhythm of the heart. And I'm not sure whether a bicycle came with the device, but you could see that it would be useful in this situation. Obviously we've done better since that time, right? We have a plethora of devices that can monitor people for days or even weeks on end and do ambulatory telemetry and whatnot. But these clearly are snapshots in time, right? And if we wanted to monitor someone for AFib and wanted to get information on AF burden, these are not viable options. Of course we can use implantable devices to know if someone's gone back into AFib and to assess AF burden or AF duration. So in fact, when we talk about reasons to stop anticoagulation, three of these things may be true. So can we leverage these advances in both pharmacology and technology to offer another option? Instead of continuously anticoagulating these patients, can we stop anticoagulation and reinitiate anticoagulation for only a period of time in response to a prolonged episode of AFib? So can we deliver personalized, targeted, time-delimited anticoagulation? So we spoke about two choices for my patient Mr. M, but maybe there's this third choice, right? Take patients with a low burden of atrial fibrillation and anticoagulate only as needed, thereby minimizing the exposure to chronic anticoagulation while still reducing risk of stroke. Now if we're going to do this with implantable devices, there are technological challenges here, right? These devices still are physician facing. Patients do not get the data from their device in real time. So we would need to have a system where the data is downloaded, it's adjudicated by device specialists, and then the patient is alerted very quickly that they've had a threshold event. So the first study that we did to assess the feasibility of this approach is called React.com. It was a pilot study using an earlier iteration of the Reveal Link, called a Reveal XT. This was a device about the size of a thumb drive, actually had to be sort of put in like a pacemaker generator, and patients had to manually download the data from their device every single day, which took a few minutes. But that's what we had at the time. So what we did is we took patients with these devices. If they had an hour or more of atrial fibrillation over a 24-hour period of continuous AFib, they would re-initiate their previously prescribed Noac for a month. So if a patient had an episode of AFib every 29 days, they would never get off their anticoagulation. However, if a patient like on the right, who had a single two-hour episode of AFib over a 14-month period, for him, if we did daily monitoring, we could anticoagulate for a month, hopefully protect against stroke while eliminating the exposure to anticoagulation the other 11 months out of the year. So in this pilot study, we enrolled 59 patients. The mean CHADS VASc score was 2.4. About two-thirds had undergone prior ablation. The median follow-up time was about 14 months. Now one point I want to make is that many of my colleagues said, you're never going to be able to do this, because patients are never going to want to wake up every morning and take the three or four minutes it would take to download the data from their implantable monitor. In fact, this wasn't the case, because as it turns out, patients don't want to have a stroke and don't want to be on anticoagulation. So being basically blessed by the medical community every day to say that this is a low-risk day, that you don't need to take your anticoagulant, that was enough to drive a daily compliance rate with transmissions of 99.7%. The other point I want to make is that many of my colleagues said, you'll never be able to do this, because all of these patients are going to have several hours of AFib in the middle of the night that you otherwise wouldn't know about. And actually, that too wasn't the case, because of the 59 patients, only 18 had at least an hour episode of AFib over that 14-month period. Those that did averaged about two episodes per patient. Most of these episodes spontaneously converted within 48 hours. Most of these episodes were asymptomatic. And if you look at the percent of time these patients should have been on anticoagulation versus the amount of time they actually were on anticoagulation, it represents a 94% reduction in the time on oral anticoagulation. There were no strokes. There was one TIA in a patient with a CHADS2 score of 1, who was on aspirin and had no atrial fibrillation in the year preceding the TIA. I will tell you, there were two major bleeds, both occurring off NOACs. One is a patient who I just saw last week. She's in her mid-80s. She's very tiny. And she was gardening around this time of year, was stung by a bee, fell and hit her head on a rock, and had a subdural bleed off the Xarelto she was supposed to be on. The other patient was working on the roof of their house, fell off, rolled into a creek behind their house, and hit their flank on a rock and bled into their kidney off the Eloquus that they would have otherwise been on. So both of these may have been very, very serious bleeds, even more serious than they turned out to be. The second study that we did, and this was with Jonathan Wax and Peter Zimmelbaum at Beth Israel Boston, was called Tactic AF. And here, it was a similar study to React.com, but this time done with pacemakers and defibrillators. And because the presence of the atrial bleed allows greater sensitivity for shorter episodes of AFib, the threshold for re-initiating anticoagulation was as short as six minutes, or a total burden of greater than six hours over 24 hours. Here, the average age was older. Most of these patients were on antirethnic drugs. The follow-up time was also about a year. And here, because of the higher number of comorbidities and the shorter threshold for re-initiating anticoagulation, about two-thirds of patients took at least a month of NOAC. But this still represented a 75% reduction in the time on oral anticoagulation. There were no strokes or TIAs, and one major bleed, again, occurring off NOAC would have been probably life-threatening had it been on NOAC. So, what have we learned from these pilot studies? Together, they enrolled 96 patients, had 112 patient years of follow-up. The reduction in the time on oral anticoagulation was 75% for one and 94% for the other because of the different patient populations and the different thresholds for initiating anticoagulation. There were no strokes in either studies, and I think that this suggests that pill-in-pocket anticoagulation is feasible. What we don't know is, is it safe from a stroke perspective? And how do we scale this to the millions of people in the United States and the tens of millions of people around the world who have atrial fibrillation and may be a candidate for this approach? Clearly, implantable devices are not scalable. Most patients don't need a pacemaker or defibrillator. ILRs are invasive, expensive, have a limited battery life, and still today are physician-facing, right, where we get the data. And therefore, it requires a vast infrastructure for timely adjudication and patient notification. I think most of us believe that the future of AF monitoring will be smartphone-based. This is a picture of the Pope's inauguration in 2005. If you look at the bottom right, there's one individual who has a device capable of taking a picture and making a phone call. If you look at that same exact scene eight years later, there's not one person in the audience who does not have a device capable of taking a picture and making a phone call. There are currently more cell phone accounts on the face of the earth than there are human beings on the face of the earth, and more than 80% of the U.S. population owns a smartphone. So how do we turn a smartphone into an AF monitor? Well, the easiest thing to do is to download an app that measures something called photoplethysmography, or PPG. You simply hold up your finger to the light of your phone. It measures the pulse waveform, and it has an algorithm that can decide whether you have sinus rhythm or atrial fibrillation. The sensitivity and specificity is reported to be quite high in the highly controlled environment of the research setting, but in the real world, the sensitivity is lower because of motion artifact and whatnot. And I think most of us would want to see an ECG if we were going to tell if someone had a fib or not. Many of us are familiar with this device. This is the Cardio device. You now buy this sensor, and they now make a one-lead and six-lead device. You download the app, it actually records the ECG, and will read it as either sinus rhythm, uninterpretable, or atrial fibrillation. The sensitivity and specificity are high, but in the real-world setting, it gives no diagnosis in about 15% because of noise and other issues. But whether it's PPG or Cardia, this again is simply a snapshot, and it's not a passive AF monitor, right? It's active, and it gives no information on duration or percent of time in atrial fibrillation. What we need are wearable, passive AF monitors with ECG confirmation, and those two are here. The two devices that I'm most familiar with are shown here. On the left is what's called the Cardia Band device, and this was a device that piggybacked onto the earlier iteration of the Apple Watch that didn't have the ECG monitor. And this was a really neat device because it was always in what's called workout mode. The PPG monitor was always on, and it could monitor your level of activity and heart rate. And if it saw a discordance between your level of activity and heart rate, it could give you a warning to touch the sensor on the band with a finger from the contralateral hand. It would record a 30- second ECG and tell you whether you've had atrial fibrillation or not. So again, a passive, continuous AF monitor. About two years ago, Apple came out with its own version, which has several important features. So first of all, the basis for all of this is, again, PPG that can be done from the wrist. Those of you who own these devices, you could see these green lights go on throughout the day, and particularly when you're working out. This can measure pulse regularity to see whether you're in sinus rhythm or atrial fibrillation. Now what you may not realize is that the current generation of Apple Watch actually has three mechanisms that pertain to AFib. One is the high heart rate notification. If you're not exercising and your heart rate is higher than expected, it will give you an alert. The other important feature is called the irregular rhythm notification. Throughout the day, during periods of time when you are inactive, it will check your pulse. If it sees an irregular pulse, it will increase the sampling rate. If five out of six samples suggest atrial fibrillation, it will give you a warning, and in the current generation of Apple Watch, that warning can tell you to check your ECG. You hold your finger on the crown of the watch, it records a 30-second ECG, and it will tell you right then and there whether you're having normal sinus rhythm or atrial fibrillation. So this is a passive AF monitor with ECG confirmation. So how accurate are these devices, and can we use them not only to detect whether someone has a fib as a screening tool, but can we use this to assess how much atrial fibrillation someone has? So this is work that we did with Jeremy Wesselhoff. Jeremy was a fellow and just became an attending a few weeks ago at Rush University in Chicago. And we were working with this earlier iteration. The Apple Watch hadn't come out yet. So we recorded the data from these watches of patients with a known history of atrial fibrillation and created a neural network to understand different patterns that they fit. So we could tell, right, this is an example. When someone went into atrial fibrillation, you could see the variability in the heart rate. The person gets a warning. They check their ECG, which clearly shows atrial fibrillation. You could also see when this person is in sinus rhythm, hear this individual check their ECG. You could see a beautiful P wave and QRS from the wrist. And not only can you tell that they had a fib, but you could tell how much atrial fibrillation they had, because this is the activity level in the green. You could tell when they're not wearing their watch, because they have no activity and no heart rate. So from a wearable device, you could tell AF burden. Jeremy and I actually gave 24 of these watches to 24 patients who had an implantable cardiac monitor in place. And Jeremy painstakingly went through more than 31,000 hours of simultaneous data. And what we found was that the sensitivity for episodes of AFib lasting more than an hour was 97.4%, with a near perfect correlation between the duration of AFib on the implantable device versus the wearable device. So here's a 14-hour episode of AFib on an implantable cardiac monitor. Here's that same 14-hour episode of AFib on a wearable device, again, with ECG showing AFib compared to normal sinus rhythm. Well, what's the sensitivity and specificity of the Apple Watch? We don't know. Remember that the Apple Heart Study, which enrolled more than 419,000 patients, used an earlier iteration of the Apple Watch that did not have the ECG feature. So what they did was that if you got the irregular rhythm notification on the watch, you were sent a patch that you wore for a week or two several weeks later. So if your watch told you on September 4th that you had atrial fibrillation, but the patch that you wore in early October said you didn't have atrial fibrillation, well, of course, that doesn't mean you didn't have atrial fibrillation. So the overall positive predictive value, even with that problem, was 0.84. To me, one of the most consoling messages from the Apple Heart Study is that we were all concerned that we would be overwhelmed with false positives. But actually, when you look at the percent notified in younger patients, less likely to have AFib, it was less than 0.5%. And the notification rate became 1%, 2%, or 3% only once you got into the older patient population. So now the question becomes, can we do pill-in-pocket anticoagulation with a wearable device? Can we personalize care instead of shotgunning the approach and treating all of these patients the same? So the study that we have designed is called REACT-AF. I will tell you that we've been working on this for a long time, but this has gotten through NIH study section, has gotten through NIH council, and now we're currently negotiating the budget. So this will hopefully see the light of day in the not-too-distant future. This is a pivotal one-to-one randomized trial that will compare chronic novel oral anticoagulation to watch-guided, targeted, time-delimited NOAC. The primary endpoint is non-inferiority for a combination of stroke, arterial embolism, and death due to cardiovascular causes. And the secondary endpoint is superiority for major bleeds. This study will enroll 5,500 patients at up to 100 US sites, with follow-up a minimum of three years in duration. As I said, it is funded by the NIH, with Johns Hopkins being the data coordinating center, and Northwestern and Stanford being the clinical coordinating centers. Now I want to make clear who we're focusing in on. Number one, these patients have a low AF burden. They've worn patches that show no AFib for three weeks prior to randomization. We're also focusing on a lower risk patient population, a CHADS-VASc score of 1 to 4 for males and 2 to 4 for females without prior stroke or TIA. Why? Well, these are patients more likely to require anticoagulant or to seek a rhythm control strategy. They're more likely to have a very active lifestyle and question the need for long-term anticoagulation. And these patients represent about half of the AF population. And these two are patients who have already shown they can tolerate an oral anticoagulant. Those patients who get randomized to the intervention arm would stop their previously prescribed oral anticoagulant, while those in the control arm will simply continue their previously prescribed NOAC. As far as the exclusion criteria, we're excluding patients who have an implantable device or have an indication, which means that we will be excluding patients with a low ejection fraction. And clearly, we're excluding patients who have a contraindication to NOAC. So conceptually, I want you to think about it like this. I mean, the irony is that we treat all of these patients the same, right? We have no regard to AF burden. And once you have a threshold CHADS-VASc score, no regard. So we treat the patient with a CHADS-VASc score of 8 in chronic atrial fibrillation the same way we treat a patient with a CHADS-VASc score of 2 who has a single six-hour episode of AFib once a year. They both get treated with a lifetime of anticoagulation. And these are the patients that we want to focus on. Now, if this were in person, I would expect that some of you would be raising your hand and saying that this is a foolish idea. And you would tell me about the data that suggests that there's a temporal dissociation between AF and stroke. So let me go over that data. This is from the ASSERT trial. And remember, ASSERT looked at patients with dual-chamber devices. And what they found in ASSERT was that there were 51 strokes. The strokes are shown here in these black lines. The darker lines that you see are the episodes of atrial fibrillation. And there are some patients who had AFib months before their stroke. There are some patients who had AFib days and weeks before their stroke. And there are some patients who had stroke and no AFib or first had AFib after their stroke. And many have taken data like this to support the concept that the AFib doesn't cause the stroke, that the AFib is simply a marker. But I would urge you to look more closely at this data because of the 2,580 patients, there were only 51 strokes. The mean age of those patients was 78. The mean CHADS-VASc score was 5. And only half of these patients had any atrial fibrillation, either pre or post-stroke. And of those, only a third had AFib before the stroke. What are alternative explanations for this data? First of all, it could be that many of these episodes of AFib lasted only a few minutes. And as I said, short episodes of AFib may not cause stroke, particularly in lower-risk individuals. We also have to recognize that not all strokes are due to atrial fibrillation. So if you have diabetes, hypertension, and heart failure, why should we be surprised that you may have strokes that are unrelated to AFib? And indeed, I think as electrophysiologists, we do ourselves a disservice when we lump all strokes together, right? What we're interested in is cardioembolic strokes. That would be like a neurologist saying, well, you have an arrhythmia and therefore you're at high risk of sudden death. Well, there are obviously many types of arrhythmia, just as there are many types of strokes. So when these authors actually went back and looked at that data, two interesting patterns emerged. Number one, if you look at the risk of stroke following a prolonged episode of AFib, i.e. greater than 24 hours, now the hazard ratio becomes very, very strong. 5.6 of a hazard ratio per stroke after a 24-hour episode. Now granted, when you adjust for comorbidities, it attenuates to nonsignificance and the numbers are relatively small. The other thing they did is they tried to retrospectively adjudicate stroke mechanism. Now this is critical because in all of these studies, these patients had pacemakers and defibrillators during a time where we weren't doing MRIs on those individuals. When they went retrospectively and adjudicated stroke mechanism, of those 51 strokes, only five were thought to be cardioembolic in nature and four out of those five occurred within 30 days of an episode of atrial fibrillation, suggesting indeed that there is a temporal association between AFib and stroke, particularly during longer periods of AFib. Now I actually think that the way to look at this better is to do what's called a case crossover, where you assess the impact of a short-term, time-varying exposure. What I mean by that is if you want to look at the impact of an episode of AFib within an individual, then look at the risk of stroke within 30 days of a prolonged AF episode, and then compare it with another time point in that same individual when they didn't have atrial fibrillation. And when you do that, an interesting pattern emerges. This is from Amintu Tarakia's work at the VA. He looked at patients with ICDs predominantly used as the AF threshold five and a half hours of AFib. He had 187 ischemic strokes. And what you see here is that the risk of stroke goes up about fivefold in the first few days and weeks following a prolonged episode of AFib. Well, you might say, listen, this is the VA. The number of strokes is relatively small. And these are all ICD patients. With Dan Singer, we recently repeated this. It was just presented at ESC. And publication is hopefully pending. But here, we combine the Optum database with the CareLink database. We had nearly 900 ischemic strokes. There was a smattering of pacemaker and ILR patients. And we used, again, that five and a half hour threshold for AFib. And we too found a fivefold increased risk of stroke in the first few days following a prolonged episode of AFib. But in our cohort, it quickly went back down to baseline after a very brief period. Again, suggesting a clear temporal association and a strong one at that between prolonged episodes of AFib and stroke. Now how do we know that these are cardioembolic events? Well, of course, we don't. But let me show you an interesting piece of data. Look at the odds ratio of that temporal association in patients off anticoagulation. It was 7.8. In patients on anticoagulation, it attenuates to nonsignificance. Suggesting, indeed, that these are cardioembolic strokes that are temporally associated with AFib and that may be responsive to anticoagulation. Of course, the other reason that this may be foolish is that AFib may simply be a marker for things like left atrial endothelial dysfunction, a local hypercoagulable state, inflammation platelet aggregation, whatever you want to call this left atrial myopathy that people talk about. What you need to ask yourself is which camp you fall into. If this is the stroke, and these are the episodes of atrial fibrillation in the black, do you believe that the risk of stroke waxes and wanes following each episode? Or do you believe that the risk is constant and that the episodes of atrial fibrillation are merely a marker for underlying risk? For those of you who would say this approach is foolish, let me remind you what I think is foolish. This is from the pinnacle registry of more than 650,000 patients who were eligible for anticoagulation. This is Coumadin in the orange. These are NOACs in the green and blues. Even in the NOAC era, only about 60% of patients who should be on anticoagulation as a result of Chads Vascor are actually prescribed anticoagulation. If you look at those individuals who are prescribed anticoagulation, and let's take Eloquus in the blue, about 50% of those patients will stop anticoagulation within about two years of prescribing it because of real or perceived risks. We have very clear guidelines telling us who should get it. We have very clear data suggesting a benefit, and yet we're only using it in half the patients, and half of those stop within about two years. The way we deal with ablation, post-ablation, I would say is foolish. This is a study of nearly 7,000 patients from Peter Noseworthy showing that in patients at high risk of stroke, i.e. Chads Vascor of two or more, only 37.5% remain on anticoagulation at 12 months, despite the fact that we know success rates of AAF ablation aren't that much higher than that. When you do stop anticoagulation, those individuals with a Chads Vascor of two or more have a increased risk of stroke by two and a half fold. So we are not giving anticoagulation to our patients as much as we should be. Patients are stopping it, either with our support or without. And post-ablation, we're stopping anticoagulation against the guidelines and exposing patients to risk of stroke. The way we practice this aspect of cardiology and electrophysiology doesn't make sense. We are treating patients in a vacuum of knowledge, when in fact we can treat them with clear data. These are two patients of mine, both 68 years of age, both females, both with a history of AFib and hypertension. The patient on the left has a no rhythm control strategy and has hours and hours of atrial fibrillation every single day. The patient on the right has an implantable monitor in place and has not one minute of atrial fibrillation post-ablation. Yet both of these patients are treated with a lifetime of oral anticoagulation based on the current standard of care. Well, if I am right, the implications I think are significant. We could reduce the time on oral anticoagulation, and in doing so, reduce bleeding, reduce costs, and potentially improve the quality of life. We could also change the indications for why we adopt a rhythm control strategy, moving it away from simply improving symptoms and in select heart failure patients, improving EF, and towards a goal of limiting or eliminating the need for chronic oral anticoagulation. I'm going to finish with a quote from the New York Times. It says, the notion of allowing patients to test themselves and treat themselves is outlandish to most doctors. I will remind you that this was not a quote from the year 2020 in response to what I am suggesting. This was written in the 1930s in response to this medication. I purposely left a lot of time for questions and answers, and we'll open it up to the group right now for discussion. Thank you very, very much for your attention. All right. Thanks, Rod. That's always a great talk. A couple of questions here. This type of strategy, maybe it's not a good idea in some patient populations. Who would you recommend that this is not considered in, and are there patient populations that you think it's particularly beneficial for? Yeah. Listen, I think that in older, people with lots of risk factors, particularly post who already had a stroke, I mean, I don't think most of us would feel comfortable with drawing anticoagulation. People need to be savvy with technology to adopt this. This is a lower risk, low burden population, but more and more, if your practice is like mine, patients come in with watches and phones and all forms of consumer grade technology to follow themselves, and we need to leverage this technology. This is for the younger AF patient with an active lifestyle, but that 60-year-old will be a 70 or an 80-year-old within the practice lifetimes of the people on this call. Any concerns that favorable data would get used too broadly? Well, I mean, first of all, this is experimental, right? I mean, we need a pivotal trial that shows that this is a safe approach, and I think like most things, there's a lot of leakage, and even though we may limit our study to a lower CHAZVAS population, I could foresee people saying, well, maybe I should put that 85-year-old with a fall risk on amiodarone, prove that they stay in sinus rhythm, and maybe avoid the need for anticoagulation or even Watchman. The question becomes, do you put Watchman's in patients who don't have AFib anymore, who have infrequent AFib regardless of their CHAZVAS score? I'll throw that out there for discussion as well, but yeah, we need to do a study allowing some creep into a higher-risk population before I think we recommend that approach. And then what type of infrastructure would be required on our end to accept these tracings, and would that be a cost-effective or something that we could actually do? Right. So if you're going to use like we did in React.com and federal devices, the answer is no, because you need people there weekends and holidays and devoted to this, and that's just not scalable. The ultimate plan is that this will be a closed loop, just as a diabetic checks their glucose and knows how much insulin to give because you've told them. My vision is that your watch will tell you you've gone to AFib. It will remind you to restart your NOAC, go into your Outlook calendar and remind you once a day or BID to take your NOAC for 30 consecutive days, and then tell you to stop. I don't want to know what my diabetic's glucose is every day, and I don't want to know whether they went back on their anticoagulation. So I view this as patient-facing and closed-loop without us needing to be involved at all. Now currently, many of us accept cardiac tracings and Apple Watch tracings at all hours of the day and night. They may or may not make it into the medical record. They may or may not build for it. The technology has outstripped our ability to incorporate this in medical care, and we need to catch up there because these technologies are only going to increase in use and be important in the management of these patients. There was a question here about if you could maybe more clearly delineate how you would use this data to instruct patients, how many hours of AFib, how would you tell them to do this? So it actually, you know, my vision is that the technology will tell them, right? It will say, hey, you know, whatever that threshold is, you know, at REACT, we're going to use an hour. It'll say, hey, you've been in atrial fibrillation for an hour, restart your NOAC. And the watch will tell you the instructions. One issue, of course, is you don't wear the watch at night. So we're going to miss those episodes. Now, that's important because, you know, in a lower risk population, I do believe that you need hours and hours of AFib, probably 12, 18 hours minimum. So I think, you know, the worst case scenario is that someone goes to sleep at 11, goes into AFib at 11.01, you know, it terminates at 7 a.m. and they put their watch on at 7.01, right? We may have missed an eight-hour episode. I think that's going to be the rare patient who's asymptomatic during that time. I do think, too, that the wearable technologies, you know, Fitbit has battery life of five days. So there's going to be other technologies out there, the ready rings and necklaces and whatnot that you could sleep with if you want, and long-term bands that will allow you to be monitored at night. So we're really just seeing the beginnings of these technologies. But my vision is that this will all be automated and you'll, you know, flip on a switch when you buy a watch to say that, A, this is an AF screening tool, like they're doing in the Heartline study, or B, this is an AF management tool, and that the algorithms will change the sensitivity and specificity according to how you're using your wearable device. Okay. Any concerns about underserved populations, given the fact we are relying on somewhat expensive devices here? Yeah. So I'm greatly concerned there. You know, clearly, you know, we've done studies and even, you know, adjusting for sort of levels of insurance, we know that underrepresented minorities don't get aggressive treatment, even when it comes to anticoagulation. The study that we're doing, we need to go with one vendor for simplicity's sake. But I do think that, you know, since many people own smartphones across the economic spectrum, that technologies that are compatible with Apple and Droid and whatnot, and can do the same thing will be forthcoming. I would argue that it's cheaper to, you know, own a phone and a watch than to pay for a NOAC every day and pay for the complications of bleeding from a system perspective. So my hope is that, you know, if a study like ours is positive, that insurance will pay for it. Because as I said, it's cheaper than the alternative. And how do the various companies feel about this? The NOAC companies and, you know, companies like Apple, do they want to step into this space? So it's funny, I approached all the NOAC companies multiple times, and they have literally thrown me out of their office. You know, I mean, I'm afraid to start the engine of my car every morning, because clearly, this is a threat. And while privately, they would say that they, or publicly, they would say they really want to target the therapy and minimize risk, you know, these drugs have limited patent lives, and they want to sell as much drug as possible. So to date, the NOAC companies have had no interest. The tech companies have a lot of interest. As you know, tech companies like Apple are sponsoring or helps run from the Heartline study, which is using Apple Watches as a screening tool. But you know, I think that we all believe that wearable technologies will ultimately be disease management tools, whether it's for AFib or high blood pressure, or maybe one day diabetes and sleep apnea. So you know, if they're going to get paid by insurance for these purposes, they're going to need pivotal studies to show that not only can you detect disease, but you can manage disease and improve heart outcomes. So anticoagulation companies, no. Tech companies, yes. There was a suggestion here that maybe one of the ablation companies would be interested in helping. Send them my way. I'll be happy to talk to them.
Video Summary
Dr. Rod Passman of Northwestern University presented a talk on the concept of "pill-in-the-pocket" anticoagulation for atrial fibrillation (AFib). The idea is to use wearable devices, such as smartwatches, to monitor AFib patients and deliver personalized, targeted anticoagulation as needed, rather than continuously. Dr. Passman shared the results of pilot studies that showed promising outcomes, including a significant reduction in the time on oral anticoagulation and no strokes. He also discussed the challenges and opportunities associated with using wearable devices for AFib monitoring and anticoagulation management. Dr. Passman emphasized the need for further research and larger trials to validate the approach and determine its safety and effectiveness. He believes that this personalized approach could reduce bleeding and costs, improve quality of life for patients, and change the indications for rhythm control strategies. However, he acknowledged concerns about the widespread use of the approach and the potential for misuse or overuse. Dr. Passman also addressed the need for healthcare systems to adapt their infrastructure to support the use of wearable devices and ensure proper data management and patient communication. While some pharmaceutical companies have not shown interest in the concept, tech companies like Apple are already investing in wearable technologies for disease management. Dr. Passman suggested that pivotal studies and evidence-based outcomes could encourage insurance coverage and integration of wearable devices into clinical practice.
Keywords
Dr. Rod Passman
anticoagulation
wearable devices
AFib monitoring
personalized approach
pilot studies
reduction in time
patient communication
Heart Rhythm Society
1325 G Street NW, Suite 500
Washington, DC 20005
P: 202-464-3400 F: 202-464-3401
E: questions@heartrhythm365.org
© Heart Rhythm Society
Privacy Policy
|
Cookie Declaration
|
Linking Policy
|
Patient Education Disclaimer
|
State Nonprofit Disclosures
|
FAQ
×
Please select your language
1
English