Gregg W. Stone, MD: Hello, I am Gregg Stone, professor of medicine and director of Cardiovascular Research and Education at the Center for Interventional Vascular Therapy at Columbia University Medical Center in New York. I would like to welcome you to this program titled "The Role of the Left Atrial Appendage in Atrial Fibrillation-Related Strokes."

Joining me today at the American College of Cardiology Meeting in San Francisco is Ted Feldman, director of the Cardiac Catheterization Laboratory at Evanston Hospital in Evanston, Illinois; Brian Whisenant, medical director of the Structural Heart Disease Program at Intermountain Healthcare in Salt Lake City, Utah; and Vivek Reddy, professor of medicine in the Department of Cardiology at Mount Sinai School of Medicine in New York. Welcome gentlemen.

The goals of the program today are to describe the role of the left atrial appendage (LAA) in the development of stroke in patients with atrial fibrillation (AF) and to discuss treatment options for reducing the stroke risk. We know this is a major healthcare problem that we are dealing with; as the population gets older, the number of patients with AF continues to exponentially climb. Strokes are going up. We have to effectively be able to manage this problem.

Before we begin with an in-depth discussion, I would like to note that this program will include discussion of devices not yet approved by the US Food and Drug Administration (FDA). We are going to get into an in-depth overview of AF and its effect on the population, and then consider the treatment alternatives.

Ted, tell us a about AF stroke. What is the etiology of it? How common is it? What is the effect on society?

Ted E. Feldman, MD: The magnitude of stroke related to AF is really pretty staggering; currently, approximately 2.5 million people in the United States have AF, and this is likely to more than double as we pass the year 2020. The incidence of stroke in this population, as a per-year instance, is 2% to 3% in the broad AF population. I think a more stark way to put it, the way that I talk about it with a patient, is a 10-year 30% stroke risk untreated, which is really a pretty frightening number. The magnitude of the problem, in terms of numbers and patient impact, is huge.

The costs are enormous because a large proportion, about a third of these strokes, can be disabling. Another third can be fatal. This is all critical because we know a little bit about some of the targets of therapy. We know that a large portion of these strokes are due to thrombi that arise in the left atrium and that 90% of those are thrombi that arise in the LAA.

Dr Stone: What is the LAA? Describe that structure.

Dr Feldman: The term appendage, or the appendix of the left atrium, is exactly descriptive. It is a projection from the roof of the superior surface of the left atrium that comes off anteriorly and basically points toward the apex of the heart, and, in fact, ...

Dr Stone: Like a finger?

Dr Feldman: A fingerlike projection. If you look at the ventricle it points toward the apex over the superior aspect of the ventricle. The similarities among patients end there because the anatomical variations in the left atrium are really myriad. The prototypical appendage is just this fingerlike projection. Various animals and vegetables are used to describe the anatomical variations. They can look like broccoli. They can look like chicken wings. A cactus is another.

Dr Stone: Cauliflower.

Dr Feldman: Cauliflower, yes, another good one.

Dr Stone: OK. Normally on a routine transthoracic echocardiogram (TEE) we cannot see the LAA. Is that correct?

Dr Feldman: Correct. The typical transthoracic, we do not. I think we are all accustomed already to thinking past that when we want to examine a patient with AF. A TEE, to look for not only left atrial but also valvular etiologies. I think here we are focusing on the nonvalvular population. An interrogation of the left atrium is critical, a TEE.

Dr Stone: Right. OK. Ninety percent to our best understanding; 90% of the strokes in patients with AF arise from thrombi that form in the LAA because of stasis, etc.

Vivek, with that in mind, what are some of the therapeutic alternatives in patients with AF to try to first prophylax against the occurrence of stroke in patients who have never had a stroke?

Vivek Y. Reddy, MD: Gregg, as you know, there are really 4 different options, each with various degrees of evidence and invasiveness. The modality that we have the greatest amount of information for is, of course, oral anticoagulation. I know we are going to get into a discussion about this. Certainly, there is warfarin, for which we have the greatest amount of information. Then, of course, during the past couple of years, we have a variety of novel anticoagulants. This is the modality for which you have the greatest amount of randomized clinical data.

Dr Stone: Pharmacologic?

Dr Reddy: Pharmacologic therapy. The second, actually the second and the third together, are ablation therapies. There is catheter ablation and there is surgical ablation. Certainly, surgical ablation came first. We do not have very many data. There is surgical ablation typically in concert with the surgical excision of the LAA. We again started approximately 3 decades ago, with relatively small numbers of patients, but we have relatively long-term data suggesting that when such a procedure is performed that patients fair pretty well. Again, most of those data were in relatively low-risk patients.

Dr Stone: Let's step back. You are talking about electrical therapies.

Dr Reddy: Yes.

Dr Stone: What about just for the patient with new-onset AF? What about cardioversion? What about rate control and getting them out of AF?

Dr Reddy: Yes. Sure. I think all of those approaches, whether it is again cardioversion, rate control, or antiarrhythmic therapy, these are all important approaches. They are approaches to treat symptoms. This is one of the lessons that we learned from AFFIRM, which looked at rate control and rhythm control.^[1]

One of the lessons we learned again is that while this may control a patient's symptoms, it does not prophylax adequately against stroke. We really have to think about AF and the treatment of AF, we have to think about it as preventing symptoms and then preventing stroke. Although they are related, they are not the same thing. Again, when you divorce these two, there are the pharmacologic therapies of oral anticoagulation and the interventional therapies of surgical and catheter ablation.

I should just say, there are a lot more data in terms of catheter ablation and how that may affect subsequent stroke. All those data, albeit in large numbers of patients, all those data are observational. We have no prospective, randomized data in terms of the utility of catheter ablation to prevent stroke.

Dr Stone: OK. I will ask you more about this later.

Dr Reddy: The last approach is appendage closure. As Dr Feldman just pointed out, the appendage is the origin of clots in most patients with nonvalvular AF. There are a host of new therapies directed at occluding, tying off, or somehow excluding the appendage from the circulation.

Dr Stone: Perfect. Brian, now let me go to you. There is AF, then there is AF. In some patients, they have a very low risk of stroke if they are in AF or paroxysmal AF. Other patients have a much higher risk of stroke. The CHADS₂ score and the CHA₂DS₂-VASc score have become part of our parlance. Can you tell us about how we should grade the risk of stroke and how that informs us as to who needs therapy?

Brian K. Whisenant, MD: Absolutely. As we initially make decisions about whether a patient should be considered for anticoagulation, we immediately gravitate to the CHADS₂ score, which is where the US guidelines are based. This is the risk score that has the best validated data. The guidelines now call for anticoagulation in patients with a CHADS₂ score of 2 or higher. That is pretty clear. There is a very good long-term data set that tells us what a patient's predicted stroke rate would be based on the CHADS₂ score. In a patient with a CHADS₂ score of 2, with aspirin therapy alone, their annual rate of stroke is certainly in excess of 3%, probably closer to 5%. It goes up very linearly from there as patients accumulate different risk factors.

Patients with a CHADS₂ score of 1 we often think of as not requiring anticoagulant therapy. The guidelines actually say that anticoagulation is optional for these patients. In fact, in practice, I think we see many patients who initially present with their stroke with a CHADS₂ score of 1. Of course, after their stroke their CHADS₂ score is 3. That demonstrates a gray zone. The primary reason why anticoagulation is optional in these patients is because of the well-documented bleeding risk of oral anticoagulants that is approximately 3% per year.

Dr Stone: Sure.

Dr Whisenant: If these patients have a stroke rate that is less than 3% per year, and their major bleeding risk is approximately 3% per year, it becomes a judgment call in patients with a CHADS₂ score of 1. That is where CHA₂DS₂-VASc comes in. The CHA₂DS₂-VASc helps us distinguish patients with a CHADS₂ score of 1, where we probably should anticoagulate those. It looks at those who are age 65 to 75 years, women, those with vascular disease already who are at a higher risk for stroke. Those patients probably need some stroke prevention in addition to aspirin therapy. That is what comes out of CHA₂DS₂-VASc.

Dr Stone: A CHADS₂ score of 2 or greater needs therapy. If a CHADS₂ score of 1, then use CHA₂DS₂-VASc to try to differentiate?

Dr Whisenant: That is my practice. Is that consistent with most of the practices?

Dr Reddy: I think so.

Dr Whisenant: I think that is a good rule.

Dr Stone: Ok, great. Then there is also bleeding risk as you mentioned. There is the HAS-BLED score. There are other scores. When you are trying to weigh the risks or benefits of -- let us start off with chronic anticoagulation therapy -- are you looking at how severe the CHADS₂ or CHA₂DS₂-VASc score is compared with the HAS-BLED score, or are you trying to just get really more of a sense looking at the patient and judging their risk? Bleeding and stroke risk, many of the factors are the same; not all, but many.

Dr Whisenant: I think most of us have incorporated the CHADS₂ score into our practice and into our decision making. Most of us probably have not incorporated HAS-BLED and the other bleeding scores into our actual day-to-day decision making. We tend to fall back on our gestalt, which most of the literature suggests undertreats patients. Patients who tend to be at the highest risk for bleeding also tend to be at the highest risk for stroke.

Dr Stone: Right.

Dr Whisenant: The HAS-BLED and the CHADS₂ score share a lot of the same letters because they share a lot of the same risk factors. Our elderly patients with the risk factors for stroke are also at risk for bleeding. For this reason, study after study, year after year, has shown that approximately half the patients with a high CHADS₂ score are actually not anticoagulated with oral medications. There is a huge unmet clinical need there.

Dr Stone: Yes, absolutely. It is probably true that the higher the bleeding risk, the higher the stroke risk. In fact, the stroke risk probably outstrips the bleeding risk. We are really undertreating a lot of patients because of the concern of bleeding. Ted, let's start now with the pharmacotherapy discussion because this is really the foundation of the way that millions of patients with AF are treated or undertreated. Warfarin is still the most commonly prescribed anticoagulant. That may be changing. It is changing over time. Tell us a little about warfarin. How efficacious is it? What is its adverse effects profile? What are some of the limitations to its use?

Dr Feldman: This is a really fundamental topic that actually does not get enough discussion because there is a reason that warfarin is a standard of care. That is, that it is a remarkably effective therapy. Broadly, warfarin therapy with a patient who is in the target international normalized ratio (INR) range reduces the stroke risk across the board by approximately two-thirds, which is a highly effective therapy. This is double-edged on the efficacy side because people still have strokes when taking warfarin. This is one of the difficult parts of the discussion with the patient. To say that you have to take this medicine that is difficult to take, requires frequent dosage adjustments in most patients, requires a huge education about interactions with foods and other drugs, and requires regular blood testing, which means trips to a facility of some kind. On top of that, it is not a perfect therapy. To try to get that whole message across to a patient, I think, is a process. It takes many visits. It is part of your relationship with the patient.

Dr Stone: Almost like insulin for a diabetic patient.

Dr Feldman: Short of the injection, it is that kind of a deal. That is actually a very good analogy. That is on the efficacy side. As Brian had mentioned, there is the risk side. That is a very consistent, again broadly average, 3% per year bleeding risk. That is much less than that in a 55-year-old with no risk factors, and certainly more than that in an octogenarian with a laundry list of comorbidities.

All that said, you go through the whole exercise of getting somebody started on warfarin therapy. It is very difficult for us as practitioners to figure out in our own practice what proportion of those patients are actually compliant with the therapy.

There is a vast literature that makes several points clear. Brian pointed out that half the patients who are candidates for warfarin therapy are not treated. That is our error.

Dr Stone: 50%?

Dr Feldman: Half of them. We do not initiate therapy. Then, among those who are receiving warfarin therapy, a small proportion is out of target INR on the high side, which is, of course, a risk. A large proportion, a third to half at any given time, is out of range on the low end of the therapeutic side. Then, further is this whole issue of medication adherence. We used to call that compliance. That was our way of being pejorative in saying the patient cannot afford a medicine (not often a problem with generic warfarin) or the patient is poorly educated or they do not want to take it. We have discovered that it is a much, much more complicated piece than that. Without jumping ahead into the results of some of our trials, one of the most striking findings in some of the recent device trials is quality-of-life assessments in patients taking warfarin. From the time of study initiation to the 1-year end-point, every measure of quality of life in people in a warfarin arm of the trial goes down. It is a downer to be on warfarin.

Dr Stone: The time in therapeutic range is difficult to achieve much beyond 60% to 65% in the real world, and even patients who are achieving a good time in therapeutic range also tend to fluctuate. It is hard to keep them perfectly controlled. They have to watch their diet, food interactions, etc.

Now, Brian, enter a whole new era of new long-term oral anticoagulants. We have 3 new ones that have been approved in the past several years. We have the direct oral thrombin inhibitor dabigatran, and then we have 2 factor-Xa inhibitors, rivaroxaban and apixaban. Tell us about some of the advantages and disadvantages of these new agents.

Dr Whisenant: Results of the clinical trials for the novel oral anticoagulants were published in the New England Journal of Medicine. We have all seen the Kaplan-Meier curves of the primary event rate of stroke. The Kaplan-Meier curves show a reduction in stroke with each one of the approved drugs compared with warfarin. Right there is an advantage. I think that is, perhaps, the most important advantage.

There is also, as Ted mentioned, a quality-of-life advantage. We see our patients who come in taking the novel anticoagulants. They are thrilled that they do not have to come in and have their INRs checked all the time. That they do not have to worry about what vegetables they eat. Young people want to live, eat a salad, and be healthy. They do not have to worry as much about that. Physicians are happy with these drugs because monitoring warfarin is not only difficult, expensive, and hazardous; it is unpleasant. There are benefits for everyone in these drugs. They are clearly an incremental step forward.

If we look at the discontinuation rate of these novel anticoagulants, interestingly, it is actually higher than with warfarin. Clearly, it was with dabigatran. I am not as familiar with the other drugs. That is a combination of adverse effects. The gastrointestinal rate of intolerance is not trivial with dabigatran. Each one of these drugs has its own profile and its own reasons why patients do not tolerate these drugs. If we look at a population -- there have been several publications about this -- from a population perspective, the novel anticoagulants have not increased the overall rate of anticoagulation among our populations. Rather, they have taken treatment away from warfarin, but they have not extended it to the population that was not treated.

We are left with a huge unmet clinical need of addressing these patients, who for bleeding risks primarily are not taking oral anticoagulants. That addresses the issue of bleeding with these drugs. That is, that they still have a bleeding risk. In the ARISTOTLE trial, it was a little bit lower than with warfarin. It is all in the ballpark. Overall, it has been generally a little bit lower than with warfarin, except for the higher dose of dabigatran. It is in that ballpark. If you drop the major bleeding rate from 3.5% to 3.1% per year, it is important. It is helpful but it does not extend that therapy to patients who are not candidates for oral anticoagulation.

Dr Stone: Right. They have decreased intracranial bleeding, and that is a good thing.

Dr Whisenant: Also increased gastrointestinal bleeding.

Dr Stone: Right. Dabigatran, in particular, has increased gastrointestinal bleeding. They may be as or more effective in decreasing stroke. Apixaban had a slight mortality benefit, but it is expensive. They have other issues, which as you say is keeping them away from most patients.

Dr Whisenant: We also have limitations in elderly patients, particularly with dabigatran. We have a data set showing that the bleeding rate in patients older than 75 years, who really have the highest stroke rate, is higher than with warfarin. That is where our biggest unmet clinical need is, in these elderly patients.

Dr Stone: The last point also is, of course, that there is no established antidote for these newer agents, and for patients who bleed, for the factor Xa inhibitors, the prothrombin concentrates, may be effective. That is being investigated. For all of this, I think most of us would feel that patients who can afford them should start taking these newer agents as opposed to warfarin; the advantages and benefits outweigh the risks. It is still not a panacea or a cure-all.

Before we get to LAA closure, Vivek, let me come back to you about AF ablation. I think this is something that most people do not think about routinely. Most hospitals do not have experts yet. There are questions as to the initial success rate and the durability of the procedure. How technique-dependent and operator-dependent is this as an option?

Dr Reddy: Very. Look, on the one hand it makes sense to think if AF is causing stroke, then get rid of the AF with ablation and you should be done with the stroke risk. In the simple scenario, that would be true. Unfortunately, from a practical perspective, our success rate with AF ablation has continued to evolve, but it is not 100%. It is important to recognize that we do not have randomized data in terms of stroke risk after AF ablation. Having said that, there are some data out there. They do include thousands of patients. There have been several studies published, again nonrandomized, that have looked at stroke risk after catheter ablation. In fact, the stroke rate seems to be quite low after catheter ablation.^[2]

Having said that, again, there are a couple of problems with these data. The first is, if you look at the patients who actually undergo ablation, they tend to be younger patients. They tend to have a relatively low CHADS₂ score. Although it is certainly useful in the treatment of any given patient as long as there is careful follow-up, you cannot really apply this to the population of patients who have AF. That is, I think, probably the most important limitation. The second is, again, most of the patients who have been treated with ablation are relatively low-risk patients. They are not the patients who have a CHADS₂ score of 2 and higher. There are relatively few patients in whom we have data in that group.

Probably the third and most important, again, if you look at patients who undergo ablation, it is not elderly patients who are at the highest risk, who are the biggest clinical unmet need. We are not doing catheter ablation in those patients. They are not the ones who are most symptomatic. Until we have data that AF ablation can actually decrease stroke risk in a randomized control study, again, although it is useful in certain patients in terms of decreasing stroke risk, for most patients it is not the modality that we look at right now.

Dr Stone: We do not have randomized trial data. It is not being used in high-risk patients with the greatest clinical need. It is operator dependent. The techniques are evolving, etc, so not quite ready for prime time yet. Let's get to what has been really exciting. We have watched this evolution now for the past almost 10 years of device-based LAA closure. Of course, this started with surgical closure of the LAA. Now we have ways to close the LAA from either the inside or the outside. Brian, please provide us with some of the alternatives, and tell us about the history of LAA closure.

Dr Whisenant: From the outside we have those surgical devices. We have the Atriclip^® (AtriCure Inc) and the TigerPaw^® (Terumo Cardiovascular Systems Corp) devices, which surgeons generally use in a thorascopic approach. These are both 510(k) (FDA Premarket Notification) approved without good stroke data. We have the transcatheter Lariat^® (SentreHEART Inc) approach, also 510(k) approved, although they are talking about trying to generate some more data with this. This is a combined hybrid endocardial-epicardial approach, but the Lariat certainly comes from an epicardial approach by tying off the appendage. Then we have many devices that are either in development or in use for endocardial LAA closure.

Clearly, the leader in this field has been the Watchman™ device (Boston Scientific Corp). The Watchman device has generated a tremendously powerful data set, primarily in PROTECT AF, and now in PREVAIL, which we will discuss. It has demonstrated not only the power of the Watchman device but also proof of concept for LAA closure in stroke reduction. We have the Amplatzer™ Cardiac Plug device (St Jude Medical Inc), which also has CE mark approval and has quite a bit of traction outside the United States. Then we have several other devices in development.

Dr Stone: For the most part, these devices are implanted through a transseptal puncture. These are self-expanding nitinol-framed devices with fabric coatings that have little grippers or anchors, and they are placed in the LAA to exclude the appendage. Because that is where 90% of the thrombi that cause AF-related strokes tend to occur, or arise from, this makes a lot of sense. We have 1 large randomized trial, here at the American College of Cardiology meeting we had a second randomized trial, for which the data were disclosed. We have gone from the PROTECT AF trial now to the PREVAIL trial. In between we have learning curve experience.

Dr Whisenant: PROTECT AF, almost 800 patients, randomized 2:1, device vs drug, in a noninferiority trial. The device clearly met noninferiority in terms of efficacy. The primary efficacy end point was a combined end point of stroke, systemic thromboembolism, and cardiovascular unexplained death.

Vivek just published the latest data set in Circulation a few weeks ago. It was an incredibly powerful data set, and very well run.^[3] Vivek, congratulations. It showed clear noninferiority. Frankly, it is trending toward statistical superiority as we now follow these patients up to 5 years. This is a very large trial. This is 800 patients, and 440 patients randomized to receive the device were followed up for 5 years. In my own practice, I have seen many of these patients coming in at 5 years now who have been off warfarin and they are doing well.

Dr Stone: These were warfarin-eligible patients?

Dr Whisenant: Warfarin eligible.

Dr Stone: These were CHADS₂ scores of 1 or greater. The device closure was noninferior, which means as good or better, and now with extended follow-up to 5 years. Why was this device not FDA approved?

Dr Whisenant: This went to the FDA panel and the panel voted for approval on a narrow vote. The FDA, of course, is not bound by the recommendations of the panel. They chose not to approve the device but rather to request a confirmatory trial, and that is the PREVAIL trial. The primary reasons why they did not approve the Watchman device after PROTECT AF were the periprocedural complications. The most important of those was pericardial effusion requiring drainage and surgery in some patients. There were 5 patients who had periprocedural strokes. Interestingly, between PROTECT AF, the CAP registry, which came next, and now PREVAIL, there has never been a procedural death in those patients. That is important. There have been, I believe, in that entire data set of now 800 patients, 5 procedural strokes, which have almost been eliminated after the early PROTECT AF data set.^[4,16]

The second is questions about efficacy. They were critical of PROTECT AF, that it had a reasonable number of patients with CHADS₂ scores of 1. It is easier to demonstrate noninferiority with a lower-risk population. They suggested that we move into a higher-risk population. They were concerned about just the magnitude of this disease, the fact that there are so many millions of patients with AF, and said are we ready to approve this device after 800 patients or do we need just a little more data, a little more follow-up? They were a little concerned about patients taking warfarin and clopidogrel for a period afterward.

Dr Stone: Now we have a new randomized trial, PREVAIL, with 407 patients, and there were 3 end points: 1 safety and 2 efficacy end points.^[16]

Let's focus right now on the safety end point because that, again, was the major concern. The efficacy follow-up is not even complete yet. How much safer was it with the learning curve, with time, with new ways to teach operators? They brought new operators into this trial who had never used the device before. Has it put most of the safety concerns to rest in your estimation?

Dr Whisenant: "To rest" is a strong statement. I would suggest that it has brought the safety concerns into an acceptable realm. It cut the major adverse event rate from approximately 8.5% in half to the low 4% rate. Stroke, which is the most important adverse event, was essentially eliminated. The need for surgery was essentially eliminated. There were still occasional adverse events. Adverse events are never acceptable, but it brought it into a reasonable perspective.

The FDA mandated that a portion of the patients come from new operators and new sites. In fact, the new operators and low sites had a lower complication rate than those with experience. This comes as a little bit of a paradox, as a surprise at least. The hypothesis is that the experienced sites were not turning away anyone. They felt that this device was appropriate for everyone, where maybe the newer centers were a little bit more cautious in their approach to this therapy. It showed that it can be rolled out to cautious new operators in a very safe manner. There was clearly an institutional learning curve that came into this therapy, where not only did each operator learn how to do the procedure, but as a community of cardiologists we have learned how to do this procedure. We can do it much safer.

Dr Stone: Ted, the Watchman has been CE Marked, meaning available in Europe, since 2005, 8 years. Now PREVAIL did not meet its safety end point, which was a prespecified upper bound. Of course, we do not speak for the FDA. We do not know what they will do.

Let's assume that this device does get approved. We now have warfarin. We now have new oral anticoagulants. We now have AF ablation. Now we have the SentreHEART device, which is already on the market as a 510(k) without that much data. Now we have the Watchman with all of these data.

Let's talk about, quickly, the patient who is anticoagulation eligible vs the patient who is anticoagulation ineligible. How do you weigh all these different options? What is the role of patient preference?

Dr Feldman: One of the things we are going to learn when Watchman does come to FDA review in the next several months is what the 4-year data looked like. As we expect, the patients taking warfarin in the PROTECT AF trial continue to have events, whereas the patients using the Watchman device will have, as we have seen, events at a much lower rate. If we just look at the curves that Vivek published recently, if they continue to move as they have, we may be recommending to patients who have a CHADS₂ score of 2 or greater that their long-term outcomes are going to be better with the device than with drug therapy. That is one possibility on the horizon.

With the information we have today, and the kind of conversation that I think we are all having with patients we see in PREVAIL and, going forward, the CAP portion of PREVAIL, it is an individualized risk/benefit discussion. If we have the idea that these therapies are approximately equal, or that the device is not inferior, all the factors that weigh against warfarin that are subjective play heavily.

Patients do not like to take warfarin. As I mentioned, there is a quality-of-life assessment in the PROTECT AF trial that is, I think, underdiscussed. I chose 2 things that all patients taking warfarin have a lot of quality-of-life problems because of therapy.^[5] All patients using the device, who are not taking warfarin, actually have a measurably better quality of life. That is an interesting discussion.

I can give you one extreme patient example that captures this: a 42-year-old guy who is kind of a patient with a CHADS₂ score of 1. This was a PROTECT AF patient, and his life and passion is martial arts. He went into AF. He is taking warfarin, and he cannot go out and break cement blocks anymore. It is horrible. For him, there is no decision here. That is an extreme of what is involved in that decision for many, many people.

Dr Stone: Many general cardiologists and electrophysiologists, AF specialists, are going to say we have data in more than 100,000 patients who have been treated with oral anticoagulation. That is probably approaching 300,000 or 400,000 now. We have randomized data in 1200 patients with this device with a relatively broad noninferiority margin. In patients who can tolerate these agents, and who are willing to tolerate these agents, that should still be the standard until we get much larger trials. Let's not even argue that point for a minute. What about the relatively warfarin-contraindicated patient? We do not have a big definitive randomized trial in that patient population. Are you going to be comfortable extrapolating these data to that patient and say we should be closing those appendages?

Dr Feldman: That is where the term "warfarin contraindicated" takes on a new meaning with an approved device. All the patients in the PROTECT AF trial have been treated with 6 weeks of warfarin after device implantation. Many patients who are relatively warfarin intolerant, either because of bleeding or other preference factors, are going to be fine with 6 weeks of warfarin therapy. I do not think that is a real obstacle. Again, what do you call warfarin intolerant?

We also have data from a European trial that looked at aspirin and clopidogrel therapy after device implantation rather than warfarin therapy. Small, underpowered, but no problems. I think we are all seeing, as we become involved with this therapy, more and more patients who have had repeated bleeding episodes when receiving anticoagulant therapy, and multiple strokes when not, who are truly contraindicated. In practice, we may see those patients as ideal for aspirin and clopidogrel therapy after device implantation. Again, as I have said, many of them will tolerate 4 to 6 weeks of warfarin therapy.

Dr Stone: This is going to be an evolving field. I can guarantee you that there are going to be a lot of future studies done because this applies to millions of Americans and other people around the world. It is a very, very important health issue. I want to thank the three of you. It has been my privilege really to be with 3 of the truly world's experts on this topic for the treatment of patients with AF.

I would like to thank you out there for participating in this activity. Now you can take the CME posttest by clicking on the Earn CME Credit link. Please also take a moment to complete the program evaluation that follows. That is very important to us. We do really look at them very carefully.