Good morning. It's a pleasure to welcome you to San Diego to the Heart Rhythm 2025 46th Annual Meeting of the Heart Rhythm Society. If you haven't already done so, please download the HRS 2025 mobile app from your preferred app store. This is how you can participate in the question and answer sessions. Please scan the QR code on the screen to access this question Q&A. When using a mobile app, log in with your HRS credentials. Please note that visual reproduction of HRS 2025, either by video or still photography, is prohibited. I would like to welcome everybody here. Kyra Kosola from Richmond, along with the co-chair of the section, Dr. Sumit Chowk, as well as the rest of the faculty. Today we will talk about sudden cardiac death. Some of the things that we already know and we're going to summarize, who are the patients who best benefit from workup, who are at risk, and what we can do to improve on this important condition. So without further ado, I would like to introduce the first speaker, Dr. Monique Starks from Duke University, who will discuss sudden cardiac death care, as well as the future plans for our society and cardiology. Thank you. Here are my disclosures. Next slide, please. I'm going to talk to you about the epidemiology of out-of-hospital cardiac arrest and the chain of survival that's critical for preventing death after sudden cardiac arrest occurs. And then we will discuss the problem of public access defibrillation, and I will provide examples of efforts that are ongoing to improve bystander CPR and AED use. Next slide. More than 350 Americans experience out-of-hospital cardiac arrest every year. Survival roughly ranges from 8 to 10 percent. There's known significant regional variation in survival, with the highest being in Seattle, lowest being in places like Alabama and Detroit. Most cardiac arrests occur in the home, and about 40 percent of those are witnessed. And there, importantly, has been no significant change in U.S. survival over the past 40 years, despite intensive efforts. We know that the chain of survival, a coordinated step of interventions that are needed to be integrated and done rapidly to improve survival, is critical to surviving cardiac arrest. We also know that in airports and casinos, we see the highest survival ranging from 40 to 60 percent, and if defibrillation can be done within three minutes, survival of more than 75 percent. We also know about the time-sensitive nature of out-of-hospital cardiac arrest from study 30 years ago showing that delays in defibrillating out-of-hospital cardiac arrest patients are associated with disproportionately low survival. And then a study out of Duke University of defibrillated patients show that when defibrillation can be provided very early, we can see high defibrillation rates, like in casinos and airports, in the community. The persons who are going to be responsible for that are bystanders and first responders, those blue and red bars. So what's the problem with AEDs in the United States? Well, the big one is we don't know where AEDs are largely, and if we do know where they are, they're not accessible 24 hours. 911 dispatchers often don't have information or knowledge of AEDs or the ability to refer bystanders to those nearest AEDs. We know that smartphone apps that refer bystanders to nearest AEDs work, but not in the United States. We have very sparse evidence, and we only have programs for public setting. And then the community largely does not know how to use AEDs. So in North Carolina, we are working on implementing the first three steps of the chain of survival in communities across North Carolina, and the first cluster randomized trial. We know that the critical barrier is not knowledge of what to do, but rather the widespread lack of systematic interventions known to be successful. So we are randomizing 62 counties, and we'll enroll more than 20,000 patients, and again, we will focus on community CPR, telephone CPR, and the ability for dispatchers to really rapidly recognize cardiac arrest and get first responders and EMS to the scene, as well as equip first responders to both treat cardiac arrest with high-quality CPR as well as AEDs. The study will be compared with Control County. It's a purely registry-based study where we will also look at quality of life, and the primary outcome is good neurologic survival. We also know from the Netherlands and places like Denmark that bystander referral programs work using smartphone apps. In Denmark, they are conducting the first randomized control trial called the Heart Runner Trial of smartphone activation of citizen responders to treat cardiac arrest. Importantly, not just in public areas, but also in private areas. They have impressively enrolled more than 150,000 citizens who act as responders and remarkably implemented more than 1,000 or more than 21,000 AEDs in the community with more than half of them having 24-hour access. They'll enroll 1,600 out-of-hospital cardiac arrest patients with a primary outcome of 30-day survival. AED delivery is really cool and really exciting. Sweden is leading the way and has been conducting their program for several years and have reported their first save in the New England Journal of Medicine. Other countries are currently working on this concept, including here in the United States. We have an American Heart Association grant to develop the protocols and to implement it and to test it real-time here in the United States. We're doing things a little bit differently in the United States. We know that we have big elephant in the room FAA barriers to flying beyond visual line of sight, but the FAA does allow public safety agencies to do this after intensive certification. We are adding drone AED delivery to a program that's ongoing in a few communities. We're developing the procedural and operational infrastructure for drone AED delivery, and we're preparing for our real-time pilot trial of drone AED delivery, which is going to be implemented in two urban regions and four rural regions in Forsyth County, North Carolina, as well as James City County, Virginia. We'll enroll 100 patients, and the primary outcome will be time from 9-1-1 arrival to AED delivery. Importantly, we have lots of partners who are making this possible, and we are regularly in communication with the FDA and the FAA. And then we know that artificial intelligence will play a critical role in the spectrum of out-of-hospital cardiac arrest. Maybe not at this time in 2025, but there's a lot of work being done across the spectrum. For example, they're already dispatched algorithms that are based on machine learning to recognize cardiac arrest and help dispatchers to dispatch first responders and EMS earlier. The sensitivity and specificity right now is too low for prime time, but we know that there's been some successes. Wearables are really important as a future mechanism for being able to recognize cardiac arrest. Those are in development, and the idea is that they would be connected to systems that can alert 9-1-1 dispatchers to get EMS to the scene. And then we know that AI will have a critical role in helping to predict sudden cardiac arrest as well. Now, importantly, as we talk about novel interventions that might move the needle, I want to talk about the common problem or the low-hanging fruit. We know that minorities in low-income communities are the least likely to be treated from cardiac arrest and the least likely to survive. But we also already know that these technologies will have the slowest diffusion to these communities. So we know that it's going to be really important that we're not only testing this in a larger scale, but we're trying to understand how we bring technological innovations to these communities. And these are just studies that I have been a part of or my colleague have done showing gross disparities between whites and blacks, blacks and Hispanics, as it relates to bystander CPR. And you'd think that the disparity would not be greater in the public setting, but it turns out that the greatest difference in bystander CPR is happening in public settings. And that's across neighborhoods and income spectrums of the community. We've actually shown that bystander AED use is the absolute lowest in minority communities at less than 1%. So in conclusion, in the U.S. in 2025, we need both old and new solutions for preventing cardiac death. Clinical trial evidence is underway for guidelines-based interventions that hopefully helps it to be easier for communities to systematically implement parts of the chain of survival. We know that public education campaigns and training to strengthen the public health response are still going to be needed and is critical. And then we need novel solutions to get AEDs into the hands of bystanders. Drone AED delivery, referral programs, the future of artificial intelligence are just a few examples. And then I implore us, as we're thinking about these novel solutions, that we include the communities that are most impacted by out-of-hospital cardiac arrest in our novel studies. Thank you. Dr. Starks, thank you for that really outstanding overview of resuscitation. Now we're going to move to prediction, and it's my pleasure to introduce Dr. Kumar Narayanan from Hyderabad, India, Medicover Hospitals, who's going to take us there. Kumar. Thank you, Dr. Chugh, and it's my honor and pleasure to be here. And we are going to talk now a little bit about the issues with identifying individuals at an increased risk of SCD. If you look broadly at the etiologies underlying SCD, we can really put them into three big baskets. That is the bulk forming with coronary artery disease. Then you have the non-ischemic structural heart disease. And then a small proportion, probably about 5%, with what we can call the pure electrical diseases or channelopathies. If you look at, in a way, our overview of present risk stratification, if you look at the populations which we are selecting for ICD implantation, you can see that we have certain specific, or let us say somewhat specific, risk stratification in terms of certain channelopathies, like CPVT, BRUGADA, or long QT, or certain specific risk scores, which we might use and say some of the non-ischemic structural diseases, like hypertrophic cardiomyopathy or ARVC. That could be said to constitute probably around 20% of the case scenario. However, if you look at the bulk of the remaining 80% with coronary artery disease or the nonspecific dilated cardiomyopathy, we are really relying on one particular risk marker, which is the presence of severe LV dysfunction. And we know over the last decade and a half that there are issues with this specific approach, and that is twofold. One is that we know that out of all those people implanted for primary prevention ICD with severe LV dysfunction, probably less than 10% really have an arrhythmic event down the line. So we are lacking the means to really identify among them the high-risk subset which is going to have SCD. And secondly, this approach doesn't really address the bulk of those people, two-thirds, who have normal LV or do not have severe LV dysfunction. So we are really needing some additional markers and approaches to improve our risk stratification at this point of time. One of the promising markers is potentially scar assessed by cardiac MRI. And we now know that it's not just the burden or the quantity of the scar, but also the quality or the characterization that matters. And this is one such example where the study showed that scar above a critical volume, but also the presence of intramural scar at an infarct border zone, conferred more than a tenfold increased risk of VT or VF. The other important thing is there is clearly a need to move away from a monoparametric LV-EF-based approach to a multi-parametric risk assessment, and this is something in which work has been going on for some time. This is one such possibility where work from Dr. Chook's group has seen that if you take multiple electric markers from the ECG and you build an electrical risk score, you can improve prediction about clinical variables. And going one step further, if you follow them longitudinally and see a worsening of the electric risk score, that potentially improves predictivity further, which takes us to the next important concept that we need to move away from a static one-time risk assessment to a repetitive or dynamic assessment. This is important because arythmic substrate can worsen over time, and conversely, your therapy can also probably improve some of those risk markers, and it's very important to keep this in mind with the progression of time, and then that can definitely help us improve risk stratification a little better. Another important concept to consider is the issue of competing risk, which we have been poorly considering so far in our risk approaches. If you look at the progression of heart failure from NYHA class 2 to class 4, yes, the risk of SCD increases, but disproportionately, the risk of non-arhythmic mortality also increases, which is why if you look at those red bars, which is the benefit you might get from the ICD by reducing arythmic mortality, it can get more than offset by an increase in non-arhythmic mortality, and therefore, you're not going to benefit certain populations where the risk of competing mortality due to non-arhythmic causes is high. It's also important to consider the influence of sex. Women have been poorly represented in most primary prevention ICD trials, and we know that there are structural differences in the structural heart disease in terms of the SCA substrate in women versus men, and this was an interesting meta-analysis in 2018 which showed that when you looked at all the primary prevention ICD trials, really, for women, there was equipoise. There was no great evidence of benefits, and this is another area where we really need to focus and look at specific risk factors in women. Another subset which has been really difficult is this one-third, potentially, who have SCA as the first manifestation of heart disease. With our current approaches, we are not even beginning to address this subset, and one of the potential contributors is possibly the issue of primary VF occurring in acute coronary syndromes, and how do we begin to address this? Now, we have always recognized, but we have kind of put it into two baskets, that there is the fixed substrate in the form of a scar, and there is the potential unstable plaque which may be leading to SCA in the situation of ACS, but this may not really be necessarily a dualistic issue. This is interesting data from the FinGESTR study recently which looks at 600 SCD victims who died of CAD, and it found that about 50% had unstable plaque, 50% had stable plaque, but what was there across the spectrum was the presence of adverse remodeling in the form of myocardial hypertrophy, as well as patchy interstitial fibrosis, not the kind of fixed scar which you might necessarily pick up by MRI or which is going to lead to a low LVEF, so I think we need to improve our understanding, potentially, of the underlying substrate in this situation. And also, in terms of addressing these patients who might have SCA as the first manifestation, as I said, conventional approaches are going to be disappointing. We do not have an advanced warning. So here's where some of the approaches towards near-term prevention could be very useful, and this is made potentially possible by the fact that up to 50% of patients might have warning symptoms in the hours or days preceding SCA, and if you combine it with the current generation of availability of vital data from wearable sensors, as well as the connectivity afforded by mobile devices, you can imagine that you can potentially combine these two in order to identify those patients who are imminently going to have an SCA and convert an unwitnessed SCA to potentially a witnessed SCA. We can also use these kind of data to potentially understand better the final pathways to arrhythmogenesis and improve our mechanistic understanding. AI can potentially help in this. This is just an example from a recent study where we looked at the ability of the first 24 hours of a Holter study to predict VT over the next 13 days, and we found that using an AI-based analysis, this was achievable with a very good AUC, and this kind of examples give us hope that we can intelligently leverage AI to automate some of these processes. In terms of looking at near-term risk, I think we also need to shift our thinking beyond just the electrical signals leading to arrhythmogenesis and also trying to understand what is the final vulnerability in the substrate which may be changing towards the end in order to lead to a lethal event. And this, again, very interesting data from the UCSF group, which did a transcriptomic profile of the hearts taken from SCD victims, and it showed that there was a differential upregulation of certain genes, especially those related to fibrosis, both active as well as mature. And again, this leads us to think that we need to do a better job in understanding not just the fixed substrate, but potentially also the dynamic substrate. We also have interesting data coming up in terms of some of the newer substrate risk factors, such as, for example, obesity cardiomyopathy, the role of fat, or very interestingly, microstructural or concealed cardiomyopathy, where you might have electrogram level abnormalities without an obvious structural defect in conventional imaging. The role of genetics is very clearly defined for monogenic disorders, but has been fairly disappointing at a population-wide level. But in the recent ASC guidelines, we have seen for the first time that the issue of assessment of genetics, especially for certain non-ischemic cardiomyopathies, has been given importance. And I think this is important to emphasize that this could have an emerging role, especially in this subset, where we need to identify certain high-risk genes, such as lamin or filament, and then use that for SCD risk stratification. And this is an evolving field. And lastly, we can think of the substrate and we can think of the trigger, but I think we should not forget the background role of the influence of autonomics. And this is, again, I think not being well looked at in terms of risk stratification. And this data, which is about 10 years old, but this showed us that in stellate ganglia removed from patients with CPVT or from long QT, there was inflammation, T cell-mediated cytotoxicity, and this leads to an increased adrenergic risk and potentially to arrhythmic storm. And this is something which I think needs to be researched further. So to conclude, current approaches to SCD risk assessment do have major gaps, especially in the region of normal EF. Scar burden and character is a promising additional marker. We need to acknowledge the heterogeneity in non-ischemic cardiomyopathy and look at the issue, especially of genetic risk assessment in them. There is an evolving understanding of the role of different new substrates as well as the role of autonomics. We need to shift from a monoparametric to a multiparametric risk assessment, incorporating dynamic risk as well as competing risk. And lastly, near-term prediction and prevention is a promising strategy which really needs further study. Thank you for your attention. Thank you very much, Dr. Narayan. It was an excellent presentation and a great overview. Let's switch gears here and look at some of the options that we have for device therapy and how we can make that better. Let me introduce Dr. Chuck Swerdlow from Cedars-Sinai Hospital. Yeah, I mean, if there's anything we can do about the echo in the room, that'd be great. Well, when I was at first, Dr. Chu, Dr. Zhao, a pleasure to give this talk in front of you and this audience. When I was assigned this topic of pearls and pitfalls of ICD therapy, my first sense was, I felt like Homer Simpson in the nuclear power plant. Too many choices, not enough time. And then as a nerd, my first thought was, well, the obvious pearl here is the technical evaluation of the ICD from the original AID to what we have today. And just taking a look at some of the components we can see by the tangle of wires, this was way before the era of microprocessors. Those of you in the front row can see that the labeling on the high voltage capacitors is the same as these photo flash capacitors developed in the 1970s for the photo flash industry so we could take birthday pictures. But before I got too far down the pitfall of this nerd hole, I thought to myself, well, I should probably focus this talk on clinical practice. And I wanted to motivate it by referring to some of the key findings of a paper published last year that I had the privilege of working with Bob Hauser on. And we looked at failure of ICDs to treat life-threatening ventricular arrhythmias by normally functioning ICD systems. So we reviewed the MOD database. We considered only reports that had been submitted by manufacturers and validated by their engineers to indicate that the ICD systems were functioning normally, meaning that generators had both normal specifications, they were turned on, the battery wasn't depleted, and the connections and leads were all working. And I wanna point out sort of three key findings from this slide. First, there are a lot of events reported. Second, the number is increasing over time of reports. And thirdly, the number of reports is increasing faster than the rate at which implants are increasing. So to provide some context, excuse me, to provide some context for this discussion, you know, I was thinking, you know, we spend a lot of time trying to identify patients who might benefit from ICDs who don't benefit. But to me, the people who already have them who don't benefit are potentially the low-hanging fruit. So I wanna first address our rules for programming implantable defibrillators. Our first guiding directive, right, is to detect all life-threatening ventricular arrhythmias. However, we have an important secondary principle to minimize shocks without violating the prime directive. And to operationalize that, about a decade ago, HRS came out with a comprehensive consensus statement in which recommendations are presented as generic programming. By that, I mean experts evaluated multiple clinical trials that use manufacturer-specific programming. They came up with these recommendations and are published in our consensus. And obvious, and some of these key recommendations are program faster detection rates for primary prevention patients, 185 to 200 beats per minute, program detection rates for secondary prevention patients, at least 10 beats per minute slower than the slowest documented VT, and to use SVT-VT discrimination algorithms on up and through very fast VTs. So I asked the question a number of years ago, is it really true? Is one set of programming safe for all? After all, what could possibly go wrong, right? So I wanted to share with you some of the pitfalls I identified in this type of analysis and share with you what might be a few of the pearls for avoiding them. The first pitfall is seen in this slide where we notice that the slowest VT rate of 320 milliseconds, 188 beats per minute, is within the range of program parameters. However, this zone uses consecutive interval counting, meaning that a single interval slower than the program detection interval will reset the count to zero. So you can see in the red boxes counter resets from intervals that are longer than the program detection interval. And this primary prevention patient from our institution died of untreated polymorphic VT. Now, if we look a little further, we can see, I apologize, the joker up at the top corner of slides just indicates these are my opinions, they're not validated in any studies. I would say this is true and the guideline is right, but it applies to zones using probabilistic counting. If we look a little deeper into why this is happening, you can see the primary cause of these long intervals is functional undersensing. That is, even though the signals are big enough that the ICD would sense them in isolation, they're not sensed because the dynamic sensing threshold doesn't adjust fast enough, it misses them, and so it registers a long interval even though there's not a true biologic long interval. The second pitfall I want to address is effective antiarrhythmic drug rates on VT. Here's an example of a baseline electrogram recorded in a patient who was treated at our institution. Rapid VT detected, treated, and resuscitated. Here is the patient's ambulatory monitor while he was dying of untreated ventricular tachyarrhythmia after prescription of an antiarrhythmic drug that slowed his VT, and as we can see from the only electrogram recorded during this event labeled by the device as a non-sustained over-sensing episode, both slowed the electrograms and resulted in marked variation in electrogram amplitude, again, functional undersensing. So I would say that the recommendation should be give a very wide berth if you elect to use type I, C, or type I, III antiarrhythmic drugs in ICD patients. The problem for us, though, is that many of these drugs are prescribed by some other physician, usually to treat natural arrhythmia without knowledge of the electrophysiologist, and without careful histories and drug histories, we'll just never identify that. A third pitfall, use of SVT-VT discriminators. Here's an example of polymorphic VT that was missed, not from our institution, by a defibrillator where the program stability criterion was set to 24 milliseconds, nominal is 40 milliseconds. Now, what that means is that any time an interval's varied by more than that amount, the ventricular, the VT count is reset back to zero. In fact, the measured stability here is 156 milliseconds, in part, again, because there's some intervals which are affected by functional undersensing as indicated by the asterisks. So my advice is program this algorithm only for rates less than 170 beats per minute, use only values greater than the nominal, be cautious with antiarrhythmic drugs, with EVICD, and as for other discriminators, never program single-chamber abrupt-onset discriminators or any discriminators in patients with complete heart block because they obviously don't need it. So you've identified a theme in this talk, hopefully, that in transvenous devices, undersensing of polymorphic VTVF is usually due to an interaction between the dynamic sensing threshold and some programmable or non-programmable parameter. In this particular case, also not from our institution, the parameter that is the culprit is the method the device uses while VF has been detected and the device is charging to re-declare sinus rhythm, again, from functional undersensing indicated by the yellow bars. So this patient, in fact, had repeated resets during charging and was eventually resuscitated by the paramedics after about 15 minutes. On another episode of ventricular fibrillation, this patient is shown here, very low amplitude signals, reliably sense because the change in amplitude is quite slow. If we go back and look at our mod study and ask the question, what were the causes of failure to detect VTVF? We'll see the most common one was SVT, was a misclassification as SVT, but the most common cause of death related to undersensing interactions. Of course, I will tell you that of the four cases I just showed you for this talk, and none of them were actually reported to mods, so this is, of course, even though there are a lot of cases here and 100 deaths, obviously a substantial undersampling. I want to close with a few thoughts just to remind us that not all PIRLs and pitfalls are technical. Pitfall number four and my last one for this talk, we spend less time talking to our patients because we're all so busy, but we don't want to fall into what Sam Sears refers to as the bionic blunder, namely by ignoring the patient experience, we actually undermine the benefits of the therapy we deliver, and I think nowhere is that more important than patients who have just experienced a shock, patients who've had especially multiple shocks. The world can look pretty dark. The pitfall's pretty deep, and the PIRLs can be swallowed up in them, so this is, without apologies, my distillation of what Sam has taught me about what to say in five minutes and still get through rounds. First, emphasize that ventricular arrhythmias is one of the challenges of heart disease. Shocks save lives and let you do more with your family and what else you want to do going on into the future, but it's natural to feel stressed or anxious after them. Then review the steps you've taken to minimize future shocks. Provide a realistic estimate of the likelihood of their success. Confirm what the patient's gonna do if there's another shock. This not only saves you middle-of-the-night phone calls, but it empowers the patient by promoting their own agency. Encourage them to have normal activity. You can see in this paper graph from a paper published by Dr. Sears and colleagues of almost 500 episodes of shocked ventricular arrhythmias and ICDs that it takes 90 days before normal activity returns to baseline in patients, and this is determined by looking at the accelerometers in the devices. And then finally, let people know, this can be tough, support's available. I mean, we need to remember that risk is perceived different when it's personal for our patients and that they'll make better choices if we acknowledge their feelings and prove worthy of their trust. If we do that, hopefully the world will look a little brighter to them, the pitfall's more manageable, and the pearl's a little larger. Thank you. Dr. Swirlo, thanks for a really thought-provoking and thoughtful talk. Our final speaker of the session is Dr. John Sapp from Queen's University in Kingston, Ontario, and he's going to talk to us about, can PVC and BT ablation reduce the risk of sudden cardiac death? Thank you very much, Dr. Chu, Dr. Kasala. It's a pleasure to be here. Thank you very much, Dr. Chu, Dr. Kasala. It's a pleasure to be here. I'm actually from Dalhousie University in Halifax on the east coast of Canada. It's always sunny, just like in this picture. I've been asked to talk about whether catheter ablation can reduce the risk of sudden cardiac death. My disclosures. To get some insight into this, I went back and looked again at the causes of death in the SCUDHEF trial, and Doug Packer published this really very nice paper looking at the causes of death. So in the placebo group, the tachyarrhythmias accounted for almost 40% of deaths, whereas in the ICD group, that was down to 20%, and heart failure began to dominate. But when you look at the absolute numbers of deaths, this is what it looks like. So a big reduction in tachyarrhythmic death and not a big change, a small maybe increase in heart failure deaths, probably people who were destined to die with progressive heart failure and arrhythmias if you had to over-interpret it. So ICDs remove most tachyarrhythmia deaths from patients with known ventricular arrhythmias. So is there room with catheter ablation for further improvement in survival? And that's the question we're looking at. So back it up a little further. VT treated by defibrillators is indeed associated with mortality, and that's in pretty much every study that looks at it. This is our own data. We saw that more shocks, higher risk of death, and it looks like this if you plot it on survival curves, especially if patients have clusters of arrhythmias. We looked at that again in the RAF trial in the nearly 11,000 ventricular arrhythmias that were experienced by patients in those trials. There's a clear dose-response relationship. The more ventricular arrhythmias, especially treated by shocks, and the more tightly clustered they are, the higher the risk of mortality. And we haven't published this yet. It's in a manuscript preparation stage, but the influence and the tightness of that relationship is very time-dependent, and it's acute. So higher concentration of shocks have a big, high-tight association with mortality that doesn't decrease back to baseline for about six months. And yet correlation does not necessarily imply causation. Ice cream sales correlate very closely with shock attacks, apparently likely something to do with warmer water and swimming. And so what is the relationship between increasing and worsening heart failure and VT recurrence? Does it simply cause more VT, and does heart failure also cause a higher risk of heart failure death, and so these track together? Or, as I would suggest, it's probably more complicated than that, and VT recurrences probably also feed back on that increasing risk of heart disease and a higher risk of death and heart failure also. So they all interact with one another if I had to guess what that relationship was. Mortality has been looked at in a number of randomized clinical trials of catheter ablation for ventricular tachycardia. We'll just go through them quickly. SMASH-VT, PAWS, VTAC, and SMS, all of the seminal first RCTs of catheter ablation for VT, none of them showed an important impact of catheter ablation on mortality. Now, this was catheter ablation compared to sort of no specific antiarrhythmic treatment. And then the VANISH trial showed, again, no significant improvement in mortality in comparison to escalating antiarrhythmic drugs for patients who had VT refractory to antiarrhythmic drugs. And more recently, the BERLIN-VT and the PARTITA trial showed discordant results with early versus late intervention for ventricular tachycardia. And then, most recently, the SURVIVE-VT trial, a trial of catheter ablation against antiarrhythmic drugs for first-line therapy, showed no change in mortality. And we saw a similar finding in the VANISH-2 trial, in which case, although the trend was favorable for catheter ablation in comparison to antiarrhythmic drug therapy, it was clearly not statistically significant. So does this mean that ablation does not save lives? This was one of my favorite books as a young guy. I had a soft cover book on the shelf above my bed. The great Carl Sagan, who is, I guess, the Neil deGrasse Tyson of the 1970s and 80s, said that absence of evidence of benefit is not evidence of absence of benefit. I think he was actually quoting someone else. In any case, so just because we haven't proven that treating VT with catheter ablation saves lives doesn't necessarily mean that we're not having a positive impact. In the VANISH trial, in the VANISH-2 trial, we randomized 416 patients. But among the half of the patients who were randomized to drug, a large proportion ended up going on to have an ablation procedure. And that's because we're not willing to sit by and watch our patients die of recurrent ventricular arrhythmias just to prove the point that we're helping them, in the same way that you can't really ethically do a randomized trial of parachutes for jumping out of airplanes. So I think that the impact of defibrillators is that they slow down the risk of death, and they convert a lot of those sudden cardiac deaths to a longer phase of heart failure. So let's transition, then, to thinking a little bit about that relationship between heart failure and ventricular arrhythmias. This is another look at the RAF trial, in which we looked at patients, the impact of cardiac resynchronization on mortality. So this is looking at it the other way. Sorry, on arrhythmias. This is looking at it the other way, the impact of treating heart failure on ventricular arrhythmias. And indeed, cardiac resynchronization delayed the new onset of ventricular arrhythmias, perhaps improving that negative remodeling that happens after cardiac injury, and also it reduces the burden of ventricular arrhythmias over time. I haven't reviewed it in this talk, but of course, pharmacotherapy that treats heart failure also reduces new onset ventricular arrhythmias. And Pugal presented yesterday, in a retrospective series, that left bundle branch area pacing seemed to reduce ventricular arrhythmias. So it seems likely that treating heart failure with all of the tools at our disposal influences ventricular arrhythmias, and then some relationship in the other direction as well. I was also asked to comment on PVCs. Well, PVCs are normal. In studies, if you monitor long enough, everybody has some PVCs. And yet, having lots of PVCs is associated with adverse outcomes. And again, it's likely that complex relationship between heart disease causing PVCs, and then, of course, do PVCs cause heart failure. PVC-induced cardiomyopathy is probably not that common, but it may be a small but constant role. And the higher the dose of PVCs, the worse the adverse effect. Going back again to old trials, the STAT-CHF trial showed that although amiodarone had no impact on mortality, it did, surprise, have an impact in ejection fraction. And of course, this has been borne out in multiple studies looking at treatment of PVCs with catheter ablation, improving ejection fraction. Successful ablation of PVCs is associated with better survival. But of course, this is not in a randomized prospective trial, but rather contaminated heavily by association. And in a systematic review done by Saurabh Kumar's group, clearly, the evidence is fairly weak to support a favorable mortality effect of catheter ablation. So I think my conclusion from this is if there's a mortality improvement from catheter ablation for PVCs, it's likely fairly small. I'll finish up with these comments. ICDs reduce sudden death, but not heart failure death. Treating heart failure with drugs and CRT and all of the tools at our disposal reduces ventricular arrhythmias and death. And it ameliorates that complex interaction between the two of them. It is likely that suppressing defibrillator shocks with ablation might reduce heart failure and thus mortality. And suppressing recurrent VT likely reduces arrhythmic mortality in some cases when defibrillators fail, as Dr. Swerdlow just showed us. It's also likely that suppressing PVCs may have some benefit in terms of heart failure, but it's over a long time horizon and very difficult to prove in a randomized trial. So finally, I think treating ventricular arrhythmias is likely to have benefit in terms of heart failure mortality, sudden cardiac death mortality. It's going to be very difficult to prove in any kind of ethical randomized trial. Thanks so much. Thank you very much, Dr. Sapp. It was a outstanding summary of the facts and our knowledge and treatment options. And I would like to open the session for questions. I encourage everybody to submit it through the online portal. And we already have one question, so maybe I just start with a quick comment and have a question then to Dr. Swerdlow about it. I think we are in a very technical field where we deal with management and re-stratification of arrhythmias and sudden cardiac death. However, if we look at the big picture, I believe the best approach to kind of affect a large number of patients is through education, you know, education of our neighbors, our relatives, our colleagues, making sure that people are on proper medications to minimize any coronary disease, which is really still the main cause of sudden death. So I think we all must remember that. Training, like training high school students for CPR and allowing a larger community approach for management of sudden cardiac death may be another way to improve our outcomes. In terms of the question, which we got from the audience, is about post-mortem ICD interrogation is missing in many institutions. Any advice on how to set this up? For me? No, that's an excellent question because I wanna, before talking about how to arrange it, I wanna just explore for 30 seconds the implications of this. Many patients with ICDs have, as we've heard and understand, have multiple possible causes for a unwitnessed sudden death or unwitnessed death that might or might not be sudden. We don't know whether or not the ICD system operated correctly, either because of some of the issues I showed or a whole other subject I didn't discuss because of some failure mode of the system so that there was a lead failure or a device failure or some other type of failure mode unless we do post-mortem interrogation. So our ability to identify failure to treat or failure to resuscitate of potentially treatable rhythms is limited, which means we under-diagnose this problem substantially, even beyond what we know about and don't report to MAUD. I wish our institution had a systematic program for doing that. My approach to doing that is to get approval from the family to ask one of the, and with that, our device reps or our device companies will send someone to the mortuary to do the post-mortem interrogation. This is actually important in many parts, even more important in other parts of the world than the US because cremation is much more common outside the United States than it is in the United States. But any device that, if a family's loved one is going to be cremated, the device has to be removed. The reason for that is that there's enough lithium in defibrillator batteries to explode, and actually explode through the body and blow away the tiles on the inside of the crematorium. So there are reasons to turn devices off for them to get explanted, and they're gonna get explanted in a large number of cases anyway. So the only question is, before they're explanted, we have to get someone to interrogate the device. We have a couple more questions. Sorry, I apologize. Maybe you can conclude. Thank you. I have a question for Dr. Starks, but first I'd like to apologize to Dr. John Sapp for getting his institution wrong. It's Dalhousie University in Halifax, Nova Scotia. Dr. Starks, we discuss cardiac arrest resuscitation in terms of response time, but really what's going on in a lot of situations is something called collapse time. Well, there's that sort of literally dead time between someone's upstairs sleeping and someone's downstairs. What can we do about that? Because every second can be precious. We hope you can fix the problem with AI. I think that's an incredible question that we actually don't have the answer to. We know that only 40% of events in the United States are witnessed, and that there's this incredibly large population that has this time that's so critical that leads to reduced outcomes. I do think the future of instances like wearables or the AI work that you are doing is so critical to helping with that time that's missed, that's truly critical in the chain of survival. So while I'm joking a little bit about you fixing the problem, I do think that the future of technology and wearables is going to be needed to help with that. Yeah, and it may not just be a wearable because it has to be something that's always listening or watching. Simin, I have a very quick thought. I apologize for not knowing the reference, but there is a study that demonstrates that AI analysis of breathing patterns on cell phones can discriminate agonal breathing from snoring quite accurately. And for people asleep, that might be a direction that's low cost. Absolutely, that's Dr. Jake Sunshine and colleagues from University of Washington. Let's move to the next question in the chat. Sure. What threshold for annual SED risk do you think is reasonable for primary prevention ICD? What are some of the risk stratifications you use? I think that's directed to Dr. Narayanan. Okay, that's a good question, and I'm not sure I have the exact answer to this. I think, as we pointed out, and yesterday I was listening to one of the abstracts of the APPRISE ATP analysis where they looked at the current rates of appropriate therapy with all the GDMT, and it was in the order of 7%. And they estimated that the median longevity prolongation in these 7% was about 23 months. So I think that kind of provides us a context. So before I think we can think of the threshold of risk for ICD implantation, I think the more important question we are faced is how do we accurately estimate the threshold? I think that's where we are kind of stuck with right now. So I think it's hard to put a number on a life saved but I think if you're looking at in terms of the intervention expense as well as the potential complications of an ICD, I guess an annual risk of somewhere over the order of 10 to 15% is what would come to mind, but this is just a number. I don't think there is any specific right answer to this. And I think additional risk stratification tools, some of them we discussed, but I think right now the reality is we don't have a one perfect marker. So I think it's ongoing work. Okay, all right. One more question. This is for Dr. Sapp. Patient with VT, recurrent VT with dual chamber ICD and cardiomyopathy and left bundle branch block, what do you do first? Do you treat the cardiomyopathy or the VT? That's a great question. Probably do both. VT is usually I think going to be the crisis that prompts treatment. And in fact, I think the decision to upgrade to some form of cardiac resynchronization is complex. There's actually surprisingly little data on the risk benefit ratio of doing the upgrade. I think all of us would suggest that a resynchronized ventricle performs better, has longer survival. But as far as resynchronization for treating ventricular arrhythmias, that's an open question. I actually think that once the substrate is present, once that ventricle has remodeled, VT is present and resynchronization and heart failure treatments are less likely to make it go away. Thank you. And one last question. I think we have time for that. What do you think about reducing the indication of ICD for primary prevention because of our four fantastic drugs? Yeah, maybe I can take that. I think we are getting there as the evidence accumulates. But I think the conundrum really is that these indications and guidelines came up based on large randomized controlled trials. And so I think to move the field again to shift our thinking, we probably need similar trials. So I think it's good that trials like the PROFIT trial, for example, in the European Union, is seriously looking at this question. So I think once we get some data and then we also have some ongoing trials, for example, on primary prevention defibrillators in the very elderly subset. So I think these are hard questions we need to look at and assess prospectively in trials. And then get the data before we are able to really change the guidelines with respect to this. So following on that, what about the 70% of people who have preserved ejection fractures? Are they going to be out of luck for the foreseeable future? Or should we be thinking of doing clinical trials with ICDs in those patients as well? Yeah, absolutely. And I think there is one arm of the PROFIT, if I'm not wrong, which is looking at preserved EF. So I think now the scientific community has woken up that we need to look at that huge subset. And I think the challenge really is going to be to get a set of markers based on which to select patients for such a trial. May I ask another question? So I think this has been a really wonderful look across the spectrum of sudden cardiac death. But perhaps I can try to connect an early part of this session to a later part. So my question is addressed to Dr. Starks and Dr. Swerdlow. So patients who have devices, whether it's pacemakers, and also we learned from Dr. Swerdlow that ICDs don't always succeed. Why can't they be connected to 911? I guess that's more for me. So as a first great question, enabling technology is finally available today to do that until the available of low energy Bluetooth is available in the United States. So if there's only available of low energy Bluetooth in devices, which is in the last few years, that really wasn't practical. But since many of the present generation of devices can access cell phone apps and can be evaluated through cell phone apps, there's no reason the next step can be taken for the generation of devices that have Bluetooth connectivity to cell phones. Because it comes to mind that when you have wearables, you can go wrong. But a pacemaker or an ICD can do a pretty good job of discriminating sinus rhythm from VT and VF. And I also think the triangulation of AI and technology and our expertise is really going to be important as we move forward. I think that's why I love this conference so much. I leave with a ton of ideas to test. And I think that these are incredible ideas that we actually need to be thinking about that we have not been able to do before. So on behalf of Dr. Kozala and myself, thank you to all the speakers and to the audience for coming in early. The session is now closed. Thank you.

sudden cardiac death

risk assessment

chain of survival

public access defibrillation

bystander CPR

AED technology

multiparametric approach

ICD programming

catheter ablation

ventricular arrhythmias