Philip B. Adamson, MD: Hello, my name is Phil Adamson. I am from the Heart Failure Institute at the Oklahoma Heart Hospital in Oklahoma City.

Paul Hauptman, MD: I am Paul Hauptman from St. Louis University School of Medicine in St. Louis, Missouri.

Scott D. Solomon, MD: And I am Scott Solomon from Harvard Medical School and Brigham and Women’s Hospital in Boston, Massachusetts. We are here today to discuss the very real risk for sudden cardiac death (SCD) in patients who have heart failure (HF). Nearly 5 million patients in the United States have HF, and nearly 500,000 are diagnosed with HF for the first time each year. These patients are particularly vulnerable to sudden cardiac arrest. Survival rates are very poor in these patients.

Our objectives today are to identify persistent treatment gaps within the HF population; to evaluate the potential mechanisms underlying the risk for SCD and HF; and to assess the role of implantable cardioverter-defibrillators (ICDs) and wearable cardioverter-defibrillators (WCDs) to address the risk of SCD in patients with ischemic and nonischemic HF. Why don’t we begin. Phil, Paul, thanks for joining. Let’s talk about the risk for sudden death in the patient with HF. Paul, who is at greatest risk, and why?

Dr Hauptman: That is really the million dollar question. We know who is at risk in a global sense. We know that ejection fraction (EF), for example, will predict SCD risk. We know a family history of SCD is a strong predictor. Aside from that, it gets more difficult, which is the clinical challenge. We do not have a good way to divide patients who are, let us say, at very high risk, from high risk, to moderate risk, to lower risk. If you are on the higher end of the low EF group, you are probably at lower risk. That is about the extent of our ability to navigate the prognostic waters.

Dr Solomon: The risk of SCD is certainly higher if your EF is very low. What about HF with preserved EF? Are those people also at increased risk?

Dr Hauptman: That is less clear. My feeling is that those patients at risk are largely mediated through hypokalemia and metabolic abnormalities. I do not think that risk is anywhere near what it is for the low EF group. I should also add that the degree to which patients are symptomatic with HF is very important. The more symptomatic the patient is, the less likely they will have a sudden death event.

Dr Adamson: That is a problem with how we come up with risk stratification. We study populations and assess a relative risk but then when we apply that to an individual, it is very difficult to hone in on that individual’s risk, to gain insight into the decision-making process for an ICD, focusing on those at highest risk, the post-myocardial infarction (MI) patient with left ventricular systolic dysfunction. Identifying those subsets of the HF population may help us guide patients with advice, and help us with our judgment as well.

Dr Solomon: I agree that when we try to figure out what factors influence the risk for sudden death in a post-MI population, we cannot at present do a very good job.

Looking at the VALIANT trial, where there were approximately 14,600 patients, nearly 1100 had an event, and more than 900 people died suddenly.^[1] We tried very hard to identify which factors predicted sudden death. Unfortunately, all the factors that predict sudden death in that population are also the factors that predict death of any sort. We did not have very sophisticated electrocardiographic measures but even if we did, I am not sure that we could have discovered the factors. Certainly there are no biomarkers that predict SCD.

Dr Adamson: For transparency, we need to indicate that when we say sudden death we mean preventable sudden death. That would mean the population of patients in whom lethal arrhythmias cause an abrupt problem that can be treated. Other things cause sudden death: cerebrovascular accidents, pulmonary embolism, myocardial rupture, and aneurysms. We have to clarify which group of patients who die suddenly can be helped. We need to identify the arrhythmic group, because we have ways to treat them.

Dr Hauptman: We should also clarify that we are talking about a primary prevention strategy. We know that if you have had a primary event, you are at high risk. Sustained ventricular tachycardia (VT), symptomatic VT with presyncope and syncope, and an actual SCD event puts you at the highest risk. We know that there are certain periods of time, like early post-MI, when you are at higher risk than later, following your MI when the patient is at risk.

We know that a wider QRS probably portends a poor overall prognosis, including SCD. Those patients are most concerning. New-onset cardiomyopathy is a very mixed heterogeneous population. They can include some patients at high risk, but they can also include some patients at low risk. Depending on presentation, up to 50% of those patients will improve their EF with aggressive medical therapy.

Dr Solomon: The issue of timing is incredibly important. We looked at the timing of sudden death relative to an MI in the VALIANT trial.^[1] We found that the risk of SCD was dramatically highest in the first 30 days, then declined fairly rapidly. Even if your EF was > 40% in the first 30 days, your risk of dying was > 1% per month. These were people who were discharged from the hospital, not expected to die, not expected to have SCD, went home, and then were found dead. The other point that you made, Phil, is that we want to consider those SCDs we can prevent. In the post-MI patient population they are not all due to arrhythmia.

We looked at this again in the VALIANT trial where we had autopsies on a subset of patients. We saw that in about 50% of patients who died suddenly, we could not identify either a new MI, a rupture, or any other clear reason for death. We assume that those are the arrhythmic deaths. Whether we could prevent all of those, we do not know.^[2] In the chronic HF population, we see the same thing. When they are hospitalized with HF, then get discharged, their risk is extremely high early and then comes down gradually with time. Is this a window of opportunity?

Dr Adamson: I think it is. It is difficult. Clinical trials applying therapies like ICDs, in that vulnerable period, the highest risk period, do not reduce mortality. That is a paradox. It is difficult to understand. Why, for example in DINAMIT was there no overall reduction in the risk for death with early application of an ICD?^[3] There are many possible answers, or at least speculations. That vulnerable period from MI and left ventricular dysfunction until 30 to 40 days is a very high-risk period, but is not best treated by an implantable device. This presents the opportunity to use alternative therapies that we have not had before.

Dr Solomon: There was another trial besides DINAMIT. It was called IRIS and it showed essentially the same thing.^[4] Early post-MI ICD did not benefit these patients.

Dr Hauptman: If you want to extend it, although it is an old trial and we do not practice this way, the CABG Patch Trial suggested that there is no benefit from early ICD placement post-revascularization.^[5] Obviously this was in the pre-transvenous defibrillator age.

Dr Adamson: That leads to the concept that there are retrievable arrhythmias, or treatable arrhythmias. Then there are other causes of SCD that may not be best thought of as an arrhythmic process. That leads us to make recommendations for all post-MI medical therapies such as β-blockers and angiotensin-converting enzyme (ACE) inhibitors, to affect the remodeling process that could lead to another MI or vascular remodeling that leads to SCD. When we get into the technocratic stage of devices that do wonderful things, we forget that the medicines we have used after MI are very effective at lowering mortality. We should focus on using those medicines and encouraging people to focus on them as well. True application of an antiarrhythmic intervention is where the conundrum comes. How do we choose which patient should get an ICD? How do we choose which patient should go home with a WCD? Now that we have WCDs, where does that fit in with the processing of our clinical judgment?

Dr Solomon: We have come a long way with medications. In the VALIANT trial, the majority of patients were on β-blockers, statins, aspirin, and, by definition, getting an inhibitor of the renin-angiotensin system; yet we had an 11% sudden death rate.^[1] Granted, we cannot necessarily affect the outcome in all of those patients, but are there strategies we should think about in these high-risk patients, and then move to the broader HF population?

Dr Hauptman: VALIANT as a clinical trial is not indicative of a real world clinical practice. In 2010 we published a paper in Circulation: Cardiovascular Quality and Outcomes that described β-blocker use in a generic HF population; < 40% of patients had a β-blocker prescription covering at least 80% of the 90 days leading up to defibrillator placement.^[6] β- blocker compliance/adherence was high in VALIANT because it was a clinical study; in the real world it is much lower. There is clearly an opportunity to maximize medical therapy post-MI and in new-onset HF. That is the lowest tech way to prevent SCD: put a patient on a β-blocker, maybe also an aldosterone antagonist following MI. EPEHESUS showed similar findings to VALIANT.^[7]

Dr Adamson: If we have an ischemic patient in that vulnerable period where ICDs do not seem to reduce mortality, but we have high risk, what do you recommend in that group?

Dr Solomon: I would first note that these studies tended to implant devices relatively late post-MI. The average time of implantation in DINAMIT was 17 days.^[3] What we found in the VALIANT trial was that the risk started very early. We are discharging patients now -- I do not know about Oklahoma or St. Louis, but in Boston, we are discharging patients within 3 or 4 days post-MI. That policy rarely changes, even if their EF is reduced post-MI. That puts them in an extremely vulnerable period. For each individual patient we need to consider whether they would benefit from some type of temporary treatment before we can make a determination of whether they might need an ICD long-term.

Dr Hauptman: Do you think these electrical events post-MI, when they do occur, are all VT or ventricular fibrillation (VF)? Do you have any idea what the incidence of bradycardic arrest, pulseless electrical activity (PEA) arrest, is? Neither the WCDs nor ICDs are going to save the patient. Could that explain the failure of these trials to show a benefit?

Dr Adamson: That is hard to say. It seems that of the arrhythmic or electrical events that lead to sudden death, it is thought that around 85% to 90% are tachyarrhythmias. Sustained bradyarrhythmias account for maybe 10%. Then the electromechanical disassociation PEA-type arrhythmias, in which there is no treatment really, is a small percentage. I think 2% to 5% are probably unrecoverable.

Dr Hauptman: Scott, you mentioned the autopsy study in VALIANT. There was an autopsy study in ATLAS.^[8] In nonischemics in ATLAS a fair percentage - - I do not recall the exact number, died of acute MI even though they were nonischemics. They went on to develop plaque rupture. We have seen that as well. Not all sudden death events in the nonischemic population are electrical.

Dr Adamson: We do not have time to talk about whether it is an ischemic cardiomyopathy, or cardiomyopathy in the presence of ischemic heart disease, and what risks that portrays. It becomes very difficult to sort out outcomes from clinical trials based on those definitions. What group of patients are we even studying? Then, when you put it into practice it becomes quite a conundrum.

Dr Solomon: Paul, we have been talking a fair amount about the post-MI patient. Obviously the chronic HF patient is at risk as well. In your practice treating chronic HF patients, do you see windows of opportunity for shorter term therapies than implantable devices?

Dr Hauptman: You are talking about the de novo HF presentation. These patients are challenging. By the Centers for Medicare and Medicaid Services (CMS), the national coverage determination, you have to wait 9 months before you can put an ICD.^[9] I have always argued that there are 2 types of patients who present de novo. There are those who have truly de novo cardiomyopathy and HF. Then there are those who have established cardiomyopathy but a de novo presentation of HF. How do you distinguish those two? That is a challenge. For the group that you think has established cardiomyopathy, to wait 9 months is asking a lot. It is also asking a lot from medical therapy. I do not know about you Phil, but I would say if you were to see a left bundle branch block pattern, an end-diastolic dimension of 7.5 cm to 8 cm -- that did not just occur in the last 30 days. That person has established cardiomyopathy, especially if you throw in a family history of cardiomyopathy. Those patients are especially concerning. In general, those patients are not going to be among the 50% that improves their EF. A few of them might but most will not.

Dr Adamson: The 9-month wait period is a real risk. When you look at the DEFINITE trial,^[10] looking back at the people who had de novo -- as close as they could call de novo HF -- in that trial, there was still a benefit of an ICD within the first 3 months of implantation. There are data, albeit retrospective, that an ICD reduces SCD in nonischemic patients requiring a shorter waiting period from the time of their diagnosis or closer to their diagnosis. It is a conundrum. Our risks are critical, possibly a SCD that could have been prevented. Yet, the alternative is 9 months of wearing a WCD. That sometimes is difficult to recommend to a patient.

Fortunately, patients with prolonged QRS durations have to wait a shorter period of time for resynchronization therapy. We have a general consensus that there should be a 3-month waiting period for a cardiac resynchronization therapy (CRT) device. These are not inconsequential questions because of the potential financial liability if we put ICDs in wrong, at the wrong time, and have to pay back the money plus penalties. These are issues we all have to face in this country. Timing is not just a consensus question, is a regulatory question, as well as a financial reimbursement question.

Dr Hauptman: Phil, that is really a great point. As a clinician you have a challenge. You have American College of Cardiology (ACC)/ American Heart Association (AHA) guidelines.^[11] You have the recently published ACC appropriateness paper across 369 different indications.^[10] You have the national coverage determination,^[9] you have clinical judgment, you have patient preference, and then you must synthesize these and not expose yourself to risk. Yet ultimately the issue is about the patient. Making sure that the patient is not put at risk is the real clinical challenge.

Dr Adamson: It is a challenge.

Dr Solomon: Many of these guidelines are based on EF, which we know is variable. It can change in the course of a patient’s illness. It is also a measurement, and I am an echocardiographer, but it is a measurement that has an error around it of probably 7 points either way. Can we be completely sure that we know what category a patient fits into when we try to assess the risk?

Dr Hauptman: I love this point because, unfortunately, we paint by numbers now. We act as if something magical happens. At 36%, you are not at risk. At 34%, put in a defibrillator. That presents two completely different approaches to the patient. Is 36 really 36? Is the 34 really 34? Is the 36 going to stay 36? Is the 34 going to stay 34? We act as if there are these cutoffs that are meaningful for the patient. The patient does not know what the EF is. The patient just does not want to have a SCD. That is a true surrogate. Scott, I am sure you have never had a patient who shows up in the office and says, “Hey doc, my EF is 29%. I think I need a defibrillator.”

Dr Solomon: The problem is that we focus on that number. We ignore other things that we know about the patient. I can tell you that from large clinical trials we can see that a patient with diabetes and an EF of 40% is at risk for sudden death just as much as a patient without diabetes and an EF of 25%. Renal function is another incredibly important modifier of risk that we are not paying attention to. What about the data for the use of WCDs? There have not been the type of trials we saw with the ICDs. We assume that if you are wearing one and you have an arrhythmia that they are effective. That depends on many things, including the patient’s adherence. Do patients like to wear these, or should I say are they comfortable wearing them enough so that they leave them on when they are at rest?

Dr Adamson: That is a big deal, both of those questions. Number one, do we know that it is effective? As Paul mentioned, it is not effective against bradyarrhythmias.

It is not effective if it is sitting on the bedside while the patient is sleeping. Your data, and others, have demonstrated that many SCDs, if not the majority of true out-of-the-hospital/at-home sudden deaths, occur while patients are sleeping. If they do not feel comfortable enough to wear it to bed while sleeping, which is one of the most vulnerable times, they could die suddenly. The post-MI period to me seems to be a no-brainer. Patients really buy into the WCD because they know in 30 days, or 40 days, or whatever their waiting period might be, they see the light at the end of the tunnel. It is the nonischemic patients who are at risk where you are compelled to wait 9 months.

Dr Solomon: You mention sleeping. It is apropos that you alluded to many of the patients who had SCD during sleep. Obviously people ask all the time if they should buy an automated external defibrillator (AED). Well, you need 2 things then: that the patient is having an arrhythmia, and that there is a spouse who wants to resuscitate them.

Dr Adamson: And who is awake?

Dr Solomon: Right.

Dr Hauptman: We know that that approach does not work from the HAT study.^[12]

In fact, it was a failure. The number of events was relatively small. The number of events for which the external defibrillator was actually applied was small. You need an educated caregiver, who is at home at the time, who can recognize when it happens. It should not be upstairs and the patient downstairs. That is often lacking. The early implant will not work. The home defibrillator will not work. In the WCD registry, the compliance was remarkably high. It was nearly 22 hours a day in something like 900 patients.^[13]

I was actually surprised when I saw those data. We are concerned if patients say, “Well, I will not wear it today.” Obviously we tell them when they take a shower they can take it off. I think 22 hours is really a positive finding. The duration of use, in general, is not 9 months. It is somewhere between 1 to 2 months when you consider the post-MI population.^[14] Clinically there is hardly anything more rewarding than to be able to say to a patient, “You know what, you can take the WCD off occasionally, and, we do not have to put in a transvenous device.

Dr Adamson: Or the one who had a shock.

Dr Hauptman: Good point.

Dr Adamson: Now you have a secondary prevention strategy for an implantable device. That is also very gratifying. Those anecdotes, I think, drive the utilization of the WCD.

Dr Solomon: Is it fair to say, in summing up, that for these very high risk patients there are options other than implanting an expensive device, and perhaps we need to grow awareness of these alternative options, emphasizing the risk to these patients.

Dr Adamson: Pressure comes from matching quality markers and mortality post MI, and the high risk in that 30-day period for sudden death, using the WCD is a grand opportunity for institutions to recognize the opportunity to reduce mortality post-MI. I think that is a great starting point, and how we apply the WCD to the longer term chronic HF patients in a primary prevention strategy will evolve over time. It is a great opportunity for us to improve quality and reduce mortality after MI.

Dr Solomon: Paul and Phil, this has been a great discussion. Thank you for participating in this activity. To proceed to the CME posttest click the earned CME credit link on this page.