Good morning, everyone. Thank you, and welcome to this session on a debate regarding primary prevention ICDs. Is it required in all patients with reduced ejection fraction? My name is James Ip. I'm a professor at Weill Cornell Medicine. I'm here with my co-chair, Dr. Singsan. Yes, welcome this afternoon. We have for you an interesting three-way debate, and we cannot say who belongs on which side, but we will find out at the end of this. I'm Sanjeev Saxena from New Jersey. I happen to be perhaps one of the only person who was there when this low EF thing was started, and I'm looking forward to what my colleagues are going to talk about. We have a very distinguished panel. We have Sanal Khadib, our program chair here, professor of medicine at Duke, Gert Hendricks, former president of IHRA and professor of medicine in Germany, and we have Ilan Goldenberg, professor of medicine at the University of Rochester, who carries the mantle of Arthur Moss's major program. Go ahead. Dr. Al Khadib is the professor at Duke University and soon to be president-elect of Heart Rhythm Society. Welcome. Thank you very much. Good morning, everyone. What a great pleasure for me to participate in this debate, because this is actually a topic that's very near and dear to my heart, and my task today is to convince you that primary prevention ICDs should continue to be indicated for patients with HFREF to improve survival, and let's delve into it. I'm going to start with my disclosures. Here's my funding, but an important disclosure that all of you should be aware of is that I actually love debates, and I'd like to share with you, although this is private, a picture of me when I was a baby when my mom started to teach me about how to debate. So I've been doing this for a long time. Now one thing I do want to highlight is I am not including the following patients in any of my arguments. Patients with HFREF, with an LVEF of greater than 35%, and patients who have a life expectancy of less than a year. Well, let's delve into it. Well, before I start convincing you that my part of the argument is the right one, I'd like to start with this observation that I made. As I mentioned, I really have participated in many debates, and my experience is that I usually just debate one person, but as you heard, all of a sudden today, I'm having to debate two people. Okay. Well, that led me to question, what are some possible explanations for this astute observation? Is it because I chaired the 2017 AHA, ACC, HRS guideline on sudden cardiac death prevention? Is it because I have more than 400 publications, many on ICDs? Is it because I've lectured widely on ICDs? Or is it because I've participated in more than 20 debates? Most were on ICDs, or simply, and I think this is the correct answer, the evidence supporting the role of ICDs in HFREF is so strong. Okay. Well, let me invoke the evidence, and I don't think I need to tell anyone in this audience about the strong evidence that we have from two major randomized clinical trials supporting the role of primary prevention ICDs in patients with HFREF. So here, to your left, you'll see the results of the MAIDER-2 trial, showing a significant 31% relative risk reduction in all-cause mortality with the ICD in patients with ischemic cardiomyopathy. And to your right, the results of the SCADHF trial showing a significant 23% relative risk reduction in all-cause mortality with the ICD compared with amiodarone or placebo. So here, you see that the results are pretty convincing. And in fact, our group and several other groups were able to replicate those findings in real-world settings because people may say, well, those results were from clinical trials. Those don't apply to my patients. So that's what we did. And we published our work in JAMA showing that actually MAIDER-2-like patients and SCADHF-like patients in clinical practice who receive an ICD have better survival than patients who receive medical therapy in the clinical trials. And so it's important to keep that in mind. And we also showed that patients with an EF of 30% to 35% had a significant improvement in survival in association with primary prevention ICDs. So what do the professional guidelines tell us? In this guideline that I had the pleasure of sharing, we basically tell you that the ICD is a class 1 recommendation, guys, in patients with ischemic cardiomyopathy, EF of less than or equal to 35%. They could be less than or equal to 30% if they have a class 1 heart failure symptoms. All right. Well, what about patients with non-ischemic cardiomyopathy because a lot of this discussion is going to evolve around patients with non-ischemic cardiomyopathy. We basically still give you a class 1 recommendation for ICDs in patients with HFREF due to non-ischemic cardiomyopathy if the LVEF is less than or equal to 35%. And patients have class 2 or 3 heart failure symptoms despite guideline-directed medical therapy. Well, some of you may say, well, these are the U.S. guidelines. I'm going to really abide by the ESC guidelines. What do the ESC guidelines tell us? Basically the recommendations are very similar. Still for ischemic cardiomyopathy, a class of recommendation is a class 1, but they did downgrade in the most recent guideline the recommendation for ICDs in patients with non-ischemic cardiomyopathy. So why are we having this discussion today? I provide this very clear and strong evidence that I just shared with you. And what I'd like to do is consider the opponent's arguments today. And in fact, I really want to address the elephant in the room, and that's the Danish trial. Many of us are aware of this trial. It was published several years ago at this point and really focused on patients with non-ischemic cardiomyopathy, LFEF of less than or equal to 35%, randomizing them to either medical therapy only or medical therapy plus an ICD. And as you see here, the results were neutral. So if you just look at these results, you may actually agree with my opponents. However, it is important to keep in mind some important nuances of the Danish trial. Remember that they required an elevated pro BMP as an inclusion criterion. In my view, this certainly could have biased the results toward an increased risk of death due to heart failure rather than sudden cardiac death. And I tell you, patients in that trial received excellent medical therapy for HFREF, not achievable in clinical practice, and I'll show you that. And very importantly, 58% of patients in the trial received a CRT device. So to me, this was more a comparison of CRTP versus CRTD rather than medical therapy versus ICD. Now, it is important to remind you all of the results of the age secondary analysis from Danish, remember this was on the backdrop of excellent medical therapy for HFREF, showing that patients who were 70 years of age or younger had a significantly better survival with an ICD versus not from Danish guys. We didn't see that in the older patients, but in the younger patients, we definitely did. All right. Now, one question I do want to ask the audience here is how many patients in Danish were black patients? A very important question to keep in mind, because the risk of sudden cardiac death is higher in our black patients. And the answer is zero, not a single patient. So that actually led us to write this editorial, if people are interested in reading it, where we talked about advancing equity in sudden cardiac death prevention, beware of making assumptions about the effectiveness of primary prevention ICDs in black patients. I truly believe we cannot extrapolate to that group. All right. So my opponents are going to try to convince you that medical therapy for HFREF has improved drastically. And I agree with them. Of course it has. No question about it. They will also refer to this figure that now has been used so many times by people who present the same arguments, basically telling you that the risk of sudden cardiac death has decreased significantly. Now remember, these were all data from randomized clinical trials where we tend to enroll the healthier patients in those trials. And those patients likely don't represent the average patient that you and I take care of in clinic. And again, very poor representation of black patients in those trials. So we have to be mindful of that. Now as I said, medical therapy for heart failure certainly reduces the risk of sudden cardiac death. I will be the first person to admit that, looking at ARNI, looking at the SGLT2 inhibitors on the backdrop of beta blockers, mineralocorticoid antagonists, what have you. Now, these arguments, however, don't take into account noncompliance with these medications. They don't take into account cost. These medications are very expensive. They don't take into account that many HFREF patients do not tolerate optimal medical therapy for heart failure. We know it time and again. So I would like here to interact with the audience by show of hands. This is an analysis that was done using data from the CHAMP Heart Failure Registry, 3,518 HFREF patients from 150 practices. In those patients, I want to ask you what percentage were simultaneously receiving target doses of ACE inhibitor ARB or ARNI, a beta blocker, and AMARE? Let's show of hands, 75% of the patients were on all of them. Nobody is taking that. What about 50%? Any takers? 50%? One taker, a couple of takers, okay. Okay, great, thank you. 25%? Okay, a lot of takers here. 1%? Indeed, that is the correct answer. Only 1%, you guys, of those patients were on target doses of those medications. And look at target doses of any of those medications reached in these patients, also pretty low. We need to keep that in mind, guys. So we wrote this commentary. Medications are important for sudden cardiac death prevention, but so is the ICD. You need both to protect patients from heart failure, not one or the other. All right. Survival benefit from ICDs, interval GDMT for HFREF, looking at more contemporary data also showing that even when all these medications are being used in clinical practice, patients who receive an ICD have better survival than patients who don't receive an ICD. Now other points that my opponents will bring up relate to the higher risk of complications. They're going to tell you there is a high risk of complications associated with ICD implantation. Well, not if you know what you're doing and you know how to program those devices. We are much smarter about programming devices today. So we should indeed be able, if we implement the made-it RIT programming in our practice, to lower the risk of inappropriate shocks to less than 5%. We've seen that in practice. And of course, if you use any antibacterial envelopes here, these are data from the RAPID trial showing about a 40% significant relative risk reduction in the risk of infection. So we are actually really doing better in this field, and you cannot dismiss that. So in fact, periprocedural complications were 3% in this paper published in JAMA Cardiology in 2020, long-term complications 6.1 per 100 patient years. That required reoperation. Okay. So they're going to also tell you, of all patients who receive an ICD, only a small number have VT or VF terminated by the ICD. Well, I do want to remind you guys that during extended made-it-to and SCAD-HEFT, there were sustained improved survival of patients with an ICD versus no ICD, and in an analysis of a subset of made-it-to patients, the probability of a first appropriate ICD for ventricular arrhythmias during that extended follow-up period was actually 68%. And indeed, there are more recent data coming out contrary to that bullet point number one that they will absolutely bring up. So I also want to share with you some insights from experts. Our findings demonstrate sustained eight-year survival benefit with primary prevention ICD therapy in made-it-to populations. Circulation 2010, my colleague was the expert who really shared that insight with us. And another comment really from a very important expert in the field, comment on the SWEED-HF trial. These findings support current ESC guideline recommendation for primary prevention ICDs in patients with HFREF. Furthermore, these findings call for better implementation. This expert wants us to use more ICDs, guys, in clinical practice, no one other than my colleague here. All right. So let's actually talk about this argument and face this reality here that there are now randomized clinical trials that are being conducted to re-examine the role of primary prevention ICDs in patients with HFREF. PROFIT is one. Contempt ICD trials. They're going to present those data. And I had the privilege of being able to comment on the Contempt ICD methodology, a trial that was published in Heart Rhythm. And what did I say? I said a discourse on primary prevention ICDs for heart failure is understandable, but answers must be sought appropriately. Why did I say that? Because there are some major concerns about the Contempt ICD trial. There are using the made-it ICD benefit score. And this lacks validation by independent research groups. It only includes clinical variables. No imaging, guys. No nothing like that. No MRI. Nothing. So certainly not reflective of today's clinical practice. Keep that in mind. Implicit selection bias as sites try to enroll patients in the trial because they're not going to enroll patients that they're worried about. They're going to give those patients ICDs. How are we going to interpret the results of the trial at the end of the day? Not sure. Was developed, validated using made-to-trial data. They're going to argue those are old data, but they're actually using the same data in this trial. So the biggest concern I have about the Contempt ICD trial is randomizing ICD-eligible patients to know ICD. And they said that they're going to try to mitigate any potential risks by discussing what's known about the ICD in the consent form. Well, we as clinicians are responsible for protecting our patients. And we should prioritize our patients' safety above everything else. It is really, in my view, unfair to overburden patients with the full responsibility of evaluating and accepting the risks of not having an ICD. And ultimately, even after long conversations with patients, we've always been there, patients defer to you. What do you think? I'll do whatever you tell me to do. The need for more data on the outcomes of ICD in contemporary clinical practice, in my mind, is indisputable. We definitely need more data. But I think that we can and should generate data without exposing patients to unnecessary risk. While randomized clinical trials are indeed the gold standard for answering questions about therapies, well-designed and conducted prospective studies can provide invaluable data. We now have many more tools at our fingertips than ever before to conduct rigorous research studies that should provide opportunities for generation of reliable data that can inform professional guidelines and practice while protecting patients' safety and welfare. And I'd like to end with two things. The Hippocratic Oath, first, do no harm. We all practice based on that. And Gandhi, who said, the best way to find yourself is to lose yourself in the service of others. Our patients, you guys. Thank you very much for your attention. So now we are going to hear from Professor Hendricks as to his response to the questions and the preemptive arguments of Dr. Al-Khatib. Just to remind everyone, we'll be taking questions at the end. We can show a QR code, or you can use the app to enter your questions, and we'll go through them after all three speakers. Dear Mr. Chairman, dear Sana, very well done, very well presented line of arguments. I would like to shift the discussion now to a more evidence-based perspective on the question whether or not patients in 2025 under which conditions may profit from an ICD implantation and where we may have equipos in that direction. What do we know in 2025, and what is the need for additional data in order to improve the quality of care? It's not the question whether I or Sana likes or dislikes the ICD. It's a question of evidences, and it's a question of responsibility for the patients to provide an appropriate therapy. I would like to take you on a journey in order to share with you my perspective and my line of arguments that drive my equipos, whether it's adequate in 2025 to continue with what we are doing for 25 years based on old and outdated trial data. They have clearly reached their end of life. So where do we stand with risk ratification in post-myocardial infarction patients? It's a very, very complex scenario. The risk for sudden cardiac death in post-myocardial infarction patient is governed by multiple parameters that dynamically influence each other's on dynamic timelines. We are currently taking only left ventricular ejection fraction for risk ratification to answer the question whether or not there may be a benefit with the ICD in the absence of any evidences and opportunities to improve risk prediction in these patients. We have been working on multiple parameters, and for me as an electrophysiologist, it's very frustrating to look back into 30 years of research where we tried to develop personalized individualized risk ratification post-myocardial infarction that is better than left ventricular ejection fraction. And this is a statement that was given in an official paper by the American Heart Association, saying that given the relatively poor performance of risk ratification approaches for sudden cardiac death and the aforementioned challenges and limitations, it is reasonable to query whether further efforts should be devoted to this area. I completely disagree with this statement, completely and fundamentally, because I believe that we need to do more to better understand sudden cardiac death, to provide a better quality of care based on new evidences that guide therapies. What is the current role of the ICD? And Sana has shared with you her perspective that she focuses on the patient group with an ejection fraction of less than 35%. According to guidelines in North America and Europe, these patients should all receive an ICD if they have a reduced ejection fraction, plus clinical signs of heart failure, on top of medical therapy. The group of patients that contributes by absolute numbers most significantly to sudden cardiac death is completely left out of the scope. They are unprotected, and the result of this treatment regimen is that we have overtreatment in patients with reduced ejection fraction. The intervention rates of the ICDs in the last 25 years went down from 15% annually to 2-3%. There is substantial overtreatment. Most of the devices will never be used, and living with an ICD on board, if you don't need them, is a worse life as compared to living without an ICD. The other patients, the ones with an ejection fraction of more than 35%, are left out of scope and simply unprotected. What do current guidelines recommend? This is a quick view on the European guidelines. It is a 1A indication that was copied and pasted for the last 25 years in the absence of new evidences and new data. And we need strong and randomized data. We should not look to registries and so-called real-world evidences. This question is so important that we need randomized data in order to re-evaluate the use of the ICD in patients post-myocardial infarctions. Why do we still have a 1A indication? Because, and I'm fully, in this particular perspective, fully with Sana, because we copy and paste the results of trials that are completely outdated. It's MADI-2 and it's SCAR-T. These trials have been planned in the mid-90s, 30 years ago, where the quality of care was completely different, interventionally different, medically different. The biggest success in medicine over the last 30 years was the reduction in mortality in heart failure patients. It's unique, even if you compare it with other medical achievements. It's because of you, what you are doing to your patients. The reduction of total mortality, the improvement of the quality of care, of medication and interventions, has dramatically reset the scene. And the patients in 2025 can no longer be compared to the MADI-2 patients and the SCAR-T patients. And I will substantiate evidences for that. Quick look to more recent trial data, non-ischemics. Sana has mentioned that in Danish there was no benefit for the primary endpoint of all-cause mortality in the non-ischemics. Rising further questions on the role of the ICD if we continue to use this outstanding device in the same mode as we did in the last 25 to 30 years. It does not make sense anymore. What are the data supporting the guidelines side of the MADI-2 and the SCAR-T trial? Both completely outdated. I recommended in 2016 to the ESC to define end-of-life dates for clinical trials when we know that they don't have anything to do with the realities, with the treatment realities anymore. The relative risk of sudden cardiac death dramatically decreased. We had 12, 15% annual mortality in these patients 30 years ago. It's now down to 5%. Because guideline adhering medical therapy and interventions have been so, so beneficial to the patients. If Sana shares with you that there is only 1% on optimal medical therapy, it should push us to do better to provide the benefits of guideline adhering medical therapy to more patients. But it cannot be used as an argument to stay with the ICD when we have opportunities to improve quality and quantity of life by means of guideline-directed medical therapy. The results of the advances in medical therapy in heart failure patients are substantial. The best endorhythmic drug that we ever had in the field are SGLT2 inhibitors. Better than everything else. And the effects of these drugs and Zacobitril, Valsartan, have not been included in these figures. They go on top. So the relative importance and contribution of sudden cardiac death to total mortality will further decrease. And then it's clear that the total setting for the ICD must be different. It cannot be the same if the mortality rates are down to 5% and sudden cardiac death rates are down to 1% or 2%. The role of the ICD must be different. And the benefits of the ICD should be different in this setting. So what do we need to do? The last three minutes of my talk bring me to the biggest project that I have the privilege to drive forward. It's Profit Era. In 2013, we started to go into a process of re-evaluating the risk ratification and also the role of medications and the ICD post-myocardial infarction with a big consortium that is sponsored by the European Commission. And we are on a three-step project plan with the re-evaluation of existing evidences to better understand the risks for sudden cardiac death post-myocardial infarction, the development of a risk predictor that is better than ejection fraction, and then going into a randomized trial. We collected 220,000 data sets adjudicated for the event of sudden cardiac death from 19 databases worldwide. All the evidences on the planet that were available to answer that question, that covered all ejection fractions post-myocardial infarction from below 20% to almost normal. And we evaluated these data sets with the complexest available mathematical and data management models, including machine learning and AI approaches. And we added CMR data. Sana was also pointing out the fact that CMR may not be included into current risk ratification, but the value of CMR is not proven in that setting for risk prediction of sudden cardiac death. We will take a look at that as well. These are the data that, as they have been published in European Heart Journal recently, 140,000 patients post-myocardial infarction in three different groups, patients with an ICD and ejection fraction less than 35%, with no ICD, ejection fraction less than 30%, and 110,000 patients with an ejection fraction of more than 35%. Events and endpoints in these three patient groups. 12-month incidence of appropriate interventions, 9% in the ICD group, 18% at three years, only less than 2% in the non-ICD group, 3.5%, less than 3.5% in the non-ICD group at three years, and in the non-ICD group with an ejection fraction of more than 35%, it was less than 4.5% and less than 1% at one year. Now looking to the prediction of ejection fraction and the use of flexible parametric models, where we had more than 80 individual data available within this data set. 0.5 as a C-statistic is flipping the coin. So without CMR, it was exactly as flipping the coin, using these data for risk prediction for the event of sudden cardiac death in the ICD patients. It was 0.53% in the non-ICD patients and 0.6% in the patients with ejection fraction of more than 35%. And there was no added value of the CMR. We had CMR data from almost 3,000 patients. This complex graph should transfer that message. It shows you the C-statistics for all three patient groups with a model based on the ejection fraction in the upper row, the flexible parametric survival model in the middle column, and the random forest model in the lower column. It's 0.5 between 0.5 and 0.6. So there is no improved risk prediction for sudden cardiac death. Why was that the case in analyzing the biggest data set which was available on the planet? I think its lack of ground truth as the education of the data in the various patient populations for the event of sudden cardiac death was inconsistent. So we were unable to define a risk predictor that substantially improved risk prediction above the risk prediction of left ventricular ejection fraction, which was with 0.5 a poor risk predictor. Data were obtained in a static mode. Changes in profiles over time could not be obtained, and that explains why not only the statistic models, but also the advanced data management models involving machine learning were not particularly helpful. How do we transfer that into the future of therapy post-myocardial infarction? We have changed the trial plans for the profit-era trial, and we have stepped out the attempt to go into the patients with preserved ejection fraction or ejection fraction of more than 35 percent because the incidence of sudden cardiac death is too low for a clinical trial. And we have shifted all the energy into a single trial with 3,600 patients that are currently randomized in 12 European geographies and 180 centers to give us true evidences on the role of the ICD for the next years to come. If we do not succeed to complete the profit-era trial, we will not be able to improve the quality of care of post-myocardial infarction patients because we definitely need new evidences in order to know what the role of the ICD will be in the future. Thank you so much. Thank you very much. Up next we have Dr. Goldenberg from the University of Rochester. And again, please use the app to submit questions that we will have towards the end. Thank you, Chairman. Thank you, Sana. Thank you, Gerard. My topic is going to be focusing on the question whether there is a need to use risk certification strategies prior to a decision to implant a primary prevention ICD in both patients with ischemic or non-ischemic heart failure. And in this topic, I'm going to focus initially on the past. Like Sana said, the guidelines for primary prevention ICDs are based on clinical trials that were conducted more than 20 years ago that are not relevant to current data on the risk benefit of the ICD in contemporary patients with ischemic and non-ischemic heart failure who are treated with optimal GDMT but also with other modalities, such as improved revascularization, treatment of mitral valve disease, exercise, and diet. And then I'm going to move on to the present. Because there's not enough data from clinical trials that were conducted over the past 20 years, we need today to apply risk certification approaches for primary ICD implants. And this is actually a mandate by CMS that requires us to have shared decision-making with our patients. And then I'm going to discuss future directions, because the only way to update guidelines is to have appropriately designed randomized clinical trials. And observational data comparing ICD versus no ICD in the real world is going to be always limited by the bias of patients who are not referred for an ICD that can never be overcome by statistical adjustments. So I'm going to initiate my talk on discussion on the guideline basis versus the data on the risk benefit of primary prevention ICDs in contemporary patients with HFRIF. As you all know, ICDs are usually implanted today as an intravascular device with the sole purpose of terminating life-threatening ventricular tachyarrhythmias, thereby reducing the risk of sudden cardiac death. So if the risk of sudden cardiac death goes down, there's no way that ICDs can improve survival if they actually increase the risk of heart failure, as we have shown, and increase the risk of non-orific mortality. So this is the only way that ICDs can improve survival. And again, if the risk of sudden cardiac death is attenuated compared to what happened 20 years ago in our population, MEDI-2 would not have been successful showing a significant survival benefit, a 31% survival benefit, with only 1,000 patients. And we conducted, at the University of Rochester, the first MEDI trials that formed the basis of guidelines for ischemic heart failure for post-M.I. patients. In 1996, only almost 200 patients were randomized. They underwent EPS. In 2002, patients with ejection fraction below 30%. We also showed a survival benefit, again, with only 1,200 patients. And these results may not be applicable to contemporary patients who receive modern medical therapy for heart failure. And again, those three studies formed the basis for guidelines. MEDI-2, published in 2002, DEFINITE, that did not reach statistical significance, published in 2004, and SCADEFT, published in 2005. So over 20 years, there have been no studies showing a significant survival benefit in patients with ischemic or non-ischemic heart failure. So in summary, patients currently are being referred by current guidelines for primary seed implantation based on those studies. However, those current recommendations are based on clinical trials that were conducted more than 20 years ago that may not be applicable for our patients. And we know already that the mortality that based the statistical power of those studies that was high has decreased substantially over the past 20 years. And we see that even in those studies, only approximately 30% of the patients received appropriate ICD therapy during the trial, and 70% did not derive a survival benefit. In the long-term study that Sana mentioned, we did show, indeed, a survival benefit that would sustain over 8 years. But in the paper, and even in the discussion, we stress the fact that the survival benefit in those patients implanted in 2000 was only 4-5 months after 8 years. Subsequently, there has been, in real-world data, even in our MEDI trials, a significant reduction in the rate of appropriate ICD therapy and appropriate ICD shocks. And today, in contemporary registries, we see an appropriate shock rate of only 1%. And this is also related to what Sana mentioned, improved rates of MEDIT-RIT criteria, improved ICD programming. So again, maybe 99% of the patients do not derive benefit from a device. I mean, this is 1 year, but even more than 10 years. So this needs to be improved with certification in this population. And this is an extension of the study that showed a reduction in the rate of sudden cardiac death with a direct correlation with improved medical management to today. And in our own MEDIT studies, we have continued to run the MEDIT studies beyond the primary prevention study. These are all the primary prevention patients that were implanted, that were had in a primary prevention study over the past 20 years. And we see, again, a significant reduction in the risk of VTVF, life-threatening VTVF at the heart rate that is equal to or greater than 200 beats per minute, a reduction in all-cause mortality that is important for planning the power of clinical trials, a reduction in the risk of sudden cardiac death, and a reduction in the risk of non-arrhythmic mortality. So this data, again, suggests that the results of MEDIT-II may not be applicable to contemporary patients with HFRIF. And I know that we acknowledge that the DENIS trial has important limitations, including CRT, the inclusion of BNP, and even it was underpowered. But this is the only clinical trial conducted in interim over the past 20 years, published in 2016, and it failed. So over the past 20 years, there's no data from randomized clinical trials showing a survival benefit with an ICD in this population. And we know already, despite improved management, and this is from contemporary data, these are five-year rates of cardiovascular complications, including inappropriate shocks that are being reduced, but they still cause anxiety to our patients, device-related complications, a low rate of systemic infection, but still evident, lead-related complications, invasive interventions. And also we have shown that ICD therapies are associated with increased risk of heart failure admissions. This is data from our MEDIT trials. And now, today, there's improved medical management of HFRIF, the four pillars that we mentioned, ARNI, SGLT-2 inhibitors, MRAs, optimization of GDMT, together with beta blockers. And these medications have improved not only survival, so survival rates have gone down, cardiovascular mortality has gone down, heart failure admission has gone down, and all-cause mortality, and the same goes for SGLT-2 inhibitors. And we see how much we improved. We see that compared to medications that were administered 20 years ago, there has been a further more than 60% reduction in the risk of all-cause mortality if we implement four-pillar therapy. And importantly, ARNI and SGLT2 inhibitors were shown recently to also reduce the risk of sudden cardiac death and life-threatening ventricular tachyarrhythmias. So again, if the risk of sudden cardiac death goes down and the risk of life-threatening ventricular arrhythmias goes down, the potential benefit of the ICD will also be attenuated. And this is the reason we need personal selection for primary ICD, prevention ICDs. So this is the reason, actually, today the CMS mandates this type of discussion with the patients. First of all, we have to consider the risk-benefit of the ICD, that even in the major clinical trials that formed the guidelines, more than 70% of the patients did not receive any therapy. Today, there's even a very low rate of appropriate ICD therapy in our contemporary patients, we have to consider the potential adverse events of the device, device-related complications, effect of inappropriate shocks on anxiety and quality of life. We have to consider the fact whether our patients are being optimized on medical therapy and contemporary medical therapy has further reduced the risk of all-cause mortality and sudden cardiac death. And we also have to take into account that even in ischemic cardiomyopathy, in the med-it trials, in the med-it studies, almost no patients received coronary revascularization, there was no appropriate treatment of valvular disease, no other modalities for intervention. So these data may not be applicable for our patients. And again, there's a need for personal selection for primary prevention ICDs in contemporary patients with FREF. And I want to discuss, just touch briefly, on what current recertification modalities are currently available for our patients. And those are primarily three recertification modalities, non-invasive ECG recertification, imaging, and clinical recertification. So we'll start with non-invasive ECG recertification approaches. There are multiple studies that have shown that autonomic function, arrhythmia markers like non-sustained VTs, depolarization parameters, repolarization parameters, are associated with increased risk of sudden cardiac death and arrhythmia in patients with FREF. The problem with those measures is that they also increase the risk of low-cost mortality and heart failure mortality. So the competing risk of those modalities are not being translated with those measures. And actually, when clinical trials try to address them and include them in the inclusion criteria like cabbage patch and dynamite, they fail to identify a survival benefit. And we actually worked with the Aragon group to identify an electrical risk score using non-invasive ECG markers, identifying increased risk for sudden death in patients regardless of ejection function. But again, this risk score also identifies the risk of non-arrhythmic mortality. And we discussed also, and Sana mentioned cardiac MRI. It's a very powerful marker of the risk of sudden cardiac death, but it also predicts the risk of low-cost mortality, cardiovascular death, and heart failure death. So the competing risk here is not clear. And again, this risk stratification approach that is now being used sometimes in clinical practice, especially in patients with an ischemic cardiomyopathy is regardless of ejection fraction. And LG has been shown in meta-analysis to be a powerful predictor of arrhythmic events in patients with low ejection fraction, even high ejection fraction, and also in primary prevention ICD patients. So it does predict arrhythmic events, but it also predicts the competing risk of non-arrhythmic mortality. And again, we're seeing that LG patients who are positive, they have a much higher risk of arrhythmic events with a low ejection fraction, high ejection fraction in primary prevention ICD patients. So this is an important component that can help us now currently risk assess our patients. And we can consider patients who are positive on MRI and have ejection fraction below equal to or less than 35% to be at a very high risk for arrhythmias, but they also may be at a high risk for non-arrhythmic mortality. And this is the reason we developed an innovative risk stratification score, because we believe that the benefit of prophylactic ICD not only relates to the risk of arrhythmic events, but it also relates to the competing risk of non-arrhythmic mortality. And we believe that improved selection for primary prevention ICDs can only be achieved by weighing the patient-specific risk of life-threatening arrhythmias against the competing risk of non-arrhythmic mortality. And this is the basis of the MED-IT ICD benefit score. It's a clinical score. It's not perfect. It does not take into account imaging data. On the other hand, it was developed on all the MED-IT trials that were conducted over the past 20 years, and it was validated in the contemporary RAID trial, and in five independent contemporary real-world populations. And the score is very simple to use, because if you wanna conduct a clinical trial and enroll more than 3,000 patients, we cannot really force everybody to have MRIs, but we're gonna have the MRI data. So it's comprised of a TVF score that includes injection fraction, a history of atrial fibrillation, heart rate, blood pressure, age, sex, a history of prior non-sustained VT, and other markers that increase the risk of non-arrhythmic mortality, that those are comorbidities, including an older age, the presence of diabetes, and CRTD that decreases the risk of arrhythmic events. And patients, and when we integrate the score, we create a matrix, and patients with a score below 50 have a higher risk of non-arrhythmic mortality compared to the risk of life-threatening ventricular arrhythmic events. As you see here, patients at the highest score will have a very high risk of arrhythmic events and a low risk of non-arrhythmic mortality, and the other way around. And the score is available online, can be very easily calculated, and we see that patients with a score of 50, at three years, they have a risk of death, non-arrhythmic mortality, that is close to 30%, and the risk of life-threatening arrhythmic events, that is less than 10%. So in this population, the potential benefit, predictive benefit of the ICD should be very low. And we also have to remember, when we discuss shared decision-making with our patients, that the patients, everything is time-dependent, so the risk of the profile, risk of the patient changes, so we have to take into account the imaging data, the ECG data, the clinical data, and continually assess our patients who have personalized selection and treatment for an ICD. So what is the future? Because currently, guidelines are based on studies that were conducted 20 years ago, we developed a lot of risk scores, we have imaging data, but these are not perfect, and using observational data to compare patients who are referred for an ICD versus no ICD is always gonna be biased and can never be corrected by appropriate statistical adjustments. And this is the reason... We're gonna have to wrap up in a minute or two. One minute. This is the reason there are ongoing three clinical trials that are aiming to address this issue, the POFIDERA discussed by Gerard, the British, which is non-inferiority, the British trial that is looking at MRI, this is actually a superiority trial, randomizing patients positive in MRI to ICD versus non-ICD. In the Contempt ICD trial, we're gonna enroll more than 3,000 patients, and we hypothesize that in this population, patients with HF-REF who are at a lower risk based on the benefit score, who are non-ICD versus ICD is non-inferior with respect to the primary endpoint of all-cause mortality and superior with respect to the secondary endpoint of survival free of major cardiovascular events. I just wanna touch on the ethical issues that were mentioned here by randomizing patients to non-ICD and for here, we actually implemented a GDMT score that will allow patients to be included only if they have a GDMT score of more than six, meaning they're either treated on all four pillars and those were optimized. And we are tracking those patients in the trial right now and patients who have a low GDMT score we're actually improving to have their treatment and we're seeing a lot of patients surprisingly are eligible for the study and those who are potentially eligible, they are being further optimized and again, reassessed for the trial. We're employing the benefit score to make sure that only patients with a low arrhythmic risk and a high risk of non-arrhythmic mortality are included and following this, we're randomizing the patients, we're allowing crossover if the profile of the risk changes. Again, this is a non-inferiority trial with a very tight non-inferiority margin and it's supported actually and that's the reason I believe that this study is needed, supported by the Heart Rhythm Society. We have members in the steering committee so we got a lot of support from the Heart Rhythm Society, from AHA, from the ACC, the European Societies of Europe and in Canada. And the rationale paper Asana mentioned was recently published in Heart Rhythm with a simultaneous publication in the Journal of Cardiac Failure. So I'm gonna summarize. We started with the past, the data and the survival benefit of the ICD in patients with ischemic and an ischemic heart are limited to clinical trials that were conducted more than two decades ago that may not be relevant to our patients. So currently we need to implement shared decision-making and personal risk assessment in primary prevention ICDs. But for the future, at least for now, we have ongoing and complementary clinical trials that are expected to inform us appropriately on the benefit of primary prevention ICDs in patients with ischemic and an ischemic heart failure who are treated with contemporary GDMT. And this is the polling question that I was requested to have. I don't know if we have the app, but maybe by a show of hands, how many of you in the room agree that clinical trials aiming to reassess the need for primary prevention ICDs are in ischemic and non-ischemic effort for currently warranted to update current HRS guidelines? Okay. I think we're going to hold that for a minute and we're going to open it for discussion. Thank you. And first of all, I want to thank the three of you for really bringing everybody up to date with what are the real data available and what are the issues that confront us. So what we'll do is we'll, we may make a comment or two from the chair, but we have a number of questions that we want to get to you with. And then we will have a poll at the end of this session. So as probably one of the few people who did the MADE-IT-1 and the MADE-IT-2 trials, I want to make an important distinction. There was no requirement for defined symptomatic heart failure in those trials. A low ejection fraction is not HFREF. And I see all this discussion that goes on between, and why is this not a semantic difference? Because the moment you have some overt diagnosable heart failure, the competing risk between modes of death changes. So I'm saying to all three of you, I'm really interested to see why none of you showed the paradigm heart failure slide on modes of death that shows that even in the lowered risk of mortality, there's a two-to-one cardiovascular death, sway towards sudden cardiac death and pump failure. Sudden cardiac death is still twice that of pump failure death in paradigm HF, which is a subgroup of the low EF group. It is not the entire low EF group. So that's the first background I'm gonna put to you. And look at this picture, and then we have really a number of good questions for you. So I'll start with a question, and you can talk about my comment. Sanjay, it's a good point that you raise, and it strengthens my perspective on the need, on the limitations of existing evidences in the fields to really adequately implement ICD therapy. We do not know in 2025, also considering the argument that you have put up on the role of clinical heart failure, who at the end of the day will profit from the ICD implantation. And that's my, and I think all of us will agree, at the end of the day, that's my call into the community. We need the effort for new data that really reflect contemporary medical scenarios. And this should also include what you have mentioned with respect to the role of clinical heart failure. And again, this stresses the ethical issues because the risk of sudden death may still be high in patients with a low ejection fraction. The risk of all-cause mortality is high. And this is the reason when we designed the study, and we approached the NIH, we approached the Patient-Centered Outcome Research Institute, they asked us to address this. Initially, the Heart Failure Society wanted to include all patients. And we developed this score to identify this competing risk to make sure, I mean, as much as we can, that patients have a low risk for arrhythmias and increased risk for the competing risk of non-arrhythmic mortality. So if you look at the risk of sudden cardiac death, according to Neurocardiac Association class, I think the data have been pretty consistent in looking at the clinical trials, even in recent days, where you know that the risk of sudden cardiac death is highest among patients with class two and class three heart failure symptoms than in patients with class one. If you look at class four, they tend to die more of advanced heart failure, albeit the risk of sudden cardiac death is still significant in those patients. So I think it is important to keep that in mind. But as you pointed out, like those clinical trials did for the most part, including MATED-2, they limited it to patients with class one, two, and three heart failure symptoms. Class four were excluded, but they ended up having a couple of patients with class four heart failure symptoms. I know those trials by heart, by the way. And so, and of course, they definitely required class two or three. And I meant Skedheft. Thanks for the comments, and I think that's a very important point that's made. But I do make the point that patients who, I've done heart failure trials, these are people who are added into trials who are refractory to the existing treatment. And when you do RNA, you have already done diuretics, ACIRBs, and MRAs. So there is a significant difference in the population, and classic heart failure trials then was done in MATED-1 and 2, and that was my point. Now, we have some very important questions, and I'll have you read them out. Not too much time left here, but I'll ask this provocative question for our esteemed panel. In your opinion, in patients who've had an ICD already for primary prevention, they have not had any therapies, and perhaps their heart failure class has changed from the time they had implanted to the time of generator change. How are you approaching these patients, and what is your discussion with them in terms of going forward with the generator change or not? I mean, I completely agree with needing to look at what the patient looks like today, obviously, kind of talking about their risk. If you look at, actually, at published data, even newer data, show that if you do replace those devices, a lot of those patients move on to get therapy, appropriate therapy from those devices. Now, do I think that we need more work to better identify which patients will and which patients know? Yeah, of course, I do. But especially in the US, I think it would be very difficult to not replace a device in a patient if at the time of device replacement, their EF is still low, they have other high-risk features, I think. I think a lot of us, for the sake of the patient, but also for medical legal reasons, I think it would be very difficult not to. I agree. I think unless there are multiple comorbidities that prohibit another procedure, right now, there's a need to proceed with device replacement. I think it's a matter of the patient preference, and it's a matter of individual risk ratification as good as we can do it currently with all the limitations that all the three of us have shared. However, if you have a patient that does not qualify anymore for primary prevention ICD indication without any therapies, I think that there is space to discuss with the patient various options of continuation. In case of doubt, I'm fully with you, Sana. I would recommend to replace. Even super-responders of CRTD can normalize ejection function, yes. I think I'd make a comment that as data is being available in real-world databases, this kind of information will become larger and more available, so you will be able to see how many patients, and perhaps even identify these patients, who will get late therapies. Let's go on to the next question. Can I do one more? Well, we can take one more, okay. Go ahead. I guess one question here is, at what point are you seeing a decrease in sudden cardiac death among patients who have had their defibrillators for primary prevention? At what point do you then cross the threshold, I suppose they're asking, of perhaps not having that generated change at the time that it needs to be done? That's a good question. The question is, if the risk of sudden cardiac death is less than 1%, less than 2%, are they're asking about the threshold? Yes, I think so. Well, that's a conversation definitely to be had with a patient. I think each patient deserves this conversation, in my view. It's very difficult to quantify, though, for a lot of patients, in my opinion, because yes, you look at the data, and they show you some numbers, but for that individual patient, it's very hard to know, because a lot of the studies also don't take into account other factors that we have access to, such as the MRI results, biomarkers, perhaps some genetics data, things like that. So it makes it more difficult for us as clinicians to implement that in clinical practice. This is also a very, very important point in the light of the trial design. What is an acceptable difference in total mortality rates? Profit era is a non-inferiority design. And what is the margin that you consider acceptable for the proof of non-inferiority? If the curves are superimposed for mortality, then everybody's fine. If there is a difference in favor of either group, then we go into a discussion on what's the value of this difference. And that's why I understand very well your decision to focus on the, as good as we can, described low-risk group first. Yeah, and I agree. This is the reason the sample size, as opposed to made it in both of our trials, increased to more than 3,000 patients, because we took a very tight non-inferiority margin. But going back to the question about the risk of sudden death, with a risk of 1%, none of the clinical trials that formed the baseline of our guidelines would have been successful with 1,000 patients. Yeah, we calculated that. We calculated the size of the trial that you would have needed, based on the expected sudden cardiac death incidence for ejection fraction of more than 35%, and it would be something like 18,000. It's undoable. Thank you. It's been a great session. I'd like to have got to some of the other questions, and I apologize we were not able to get to those who asked those questions. The panel are available for you to talk to them individually, and how many of you are not going to put in primary prevention ICDs, based on this discussion? Can I have a show of hands? Not yet, not yet. I would thank you guys very much. Thank you.

ICD

heart failure

ejection fraction

Sanal Khadib

Gert Hendricks

Ilan Goldenberg

MAIDER-2

SCADHF

Contemp-ICD

risk stratification