Robert J. Myerburg, MD: Hello. I'm Robert J. Myerburg from the University of Miami and with me today is Stuart Connolly from McMaster University, Ontario. We are here to discuss recent information on dronedarone and its implications and recommendations for clinical use.

As background on the use of antiarrhythmic drugs for the treatment of atrial fibrillation, a few years ago there was a great deal of anticipation for the release of dronedarone because of limitations of other antiarrhythmic drugs for atrial fibrillation. Stuart, tell us about dronedarone studies in the context of the limitations of other antiarrhythmic drug studies that set the stage for what we learned about dronedarone as time went on.

Stuart J. Connolly, MD: It is pretty well known that all the antiarrhythmic drugs have some serious toxicity. Amiodarone has quite a bit of noncardiac toxicity, sotalol increases the risk of torsades de pointes, and the class Ic agents flecainide and propafenone are contraindicated in all but patients who have essentially normal hearts.

Therefore, in that context there was a clear need for a better antiarrhythmic agent. Dronedarone came along more than a decade ago with its initial studies and in the initial phase 2 and early phase 3 studies it looked very good. It seemed to be well tolerated. It reduced the recurrence rates of atrial fibrillation but then there was a negative trial, the ANDROMEDA trial. The trial was not done in patients with atrial fibrillation but rather in heart failure patients. In the population with severe heart failure, dronedarone increased mortality 2-fold and the trial was stopped early. That was clearly a cause for major concern.

That led to the ATHENA trial, which was a large phase 3 trial that looked at the clinical outcome of cardiovascular hospitalization or death. This trial was designed to identify whether or not dronedarone would be effective in the population of patients with atrial fibrillation who would be most likely to receive it and, indeed, the trial showed a remarkable benefit of dronedarone in this population. There was a reduction in the primary outcome of cardiovascular death, and there was no major toxicity. Dronedarone was found to be reasonably safe in the patients with relatively mild heart failure who were treated in that study.

That led to the approval of the drug in virtually all countries and quite a considerable utilization.

Dr. Myerburg: The major hope for dronedarone was that it has a chemical property similar to amiodarone and other properties that were different but that it would have the benefits of amiodarone without the adverse effects. Do you agree with that as the driving force behind the development of dronedarone?

Dr. Connolly: That certainly was one of the key features. Chemically, the iodine molecules were taken out and the drug was made more soluble and this did, actually, get rid of the thyroid risk and it probably got rid of the pulmonary risk, although it would take longer-term studies to know that for sure.

Dr. Myerburg: Correct. Approval by the US Food and Drug Administration (FDA), as you pointed out, was really based on the results of the ATHENA trial. Can you tell us about the patient population in ATHENA versus ANDROMEDA? That is going to have relevance in terms of the more recent studies.

Dr. Connolly: The patients who went into ATHENA had to have intermittent atrial fibrillation, paroxysmal or persistent, and they needed to have some risk factors for adverse vascular events. In fact, the inclusion criteria were modeled very much on those of the AFFIRM trial. There was an attempt to find patients with some increased risk of cardiovascular hospitalization or death. They weren't a very sick population, at least according to the design of the study.

Dr. Myerburg: In terms of the actual enrollment they were very different from the ANDROMEDA study patients.

Dr. Connolly: Absolutely. The patients enrolled in ANDROMEDA were very sick heart failure patients who had mostly been hospitalized in the past month. They were a pretty extreme population.

Dr. Myerburg: Right. One other point that doesn't get discussed very much is the 1 head-to-head study of amiodarone versus dronedarone.

Dr. Connolly: That was called DIONYSOS. It was a small study that was designed to look at both the recurrence of atrial fibrillation and discontinuations for side effects. It showed that amiodarone was more potent for preventing recurrence of atrial fibrillation, which is something that most people had anticipated. It showed that dronedarone was not greatly superior to amiodarone in terms of discontinuation for side effects. The trial actually favored amiodarone.

Dr. Myerburg: But with a benefit, nonetheless, for dronedarone. FDA approval comes with the 1 major precaution about patients with heart failure and/or low ejection fractions, based on what you just told us about ANDROMEDA.

Were there differences between Europe and the United States on the initial approval?

Dr. Connolly: There were some small differences in the wording of the label. . In the United States, there was greater focus on the reduction in cardiovascular hospitalization or death. In Europe, they tended to downplay that. I think some of that, frankly, was political. Both groups, obviously, were reviewing the same data and the primary outcome of the trial was cardiovascular hospitalization or death so that is what was most clearly shown.

Dr. Myerburg: Right, and wasn't the European approval a little more extensively limiting in terms of simply a low ejection fraction versus heart failure?

Dr. Connolly: The contraindications were a little bit different, too, but everyone was struggling to find the group of patients who could take this drug safely with the negative ANDROMEDA results and the positive ATHENA results. It is not surprising that there were some small variations in the wording, but I didn't think that they were major differences.

Dr. Myerburg: Dronedarone received approval in the United States and Europe, and then some postmarketing issues surfaced, the most dramatic of which, although quantitatively not very disconcerting, at least in terms of numbers, was the hepatic issue.

Dr. Connolly: That was after the drug had been on the market. I have forgotten how long it was, but it wasn't very long – 6 or 8 months, or maybe 12 months. There were 2 almost simultaneous cases in which patients receiving dronedarone experienced severe liver failure leading to transplantation and in whom there was no other obvious cause.

Dr. Myerburg: That was against the denominator of roughly 200,000 patients.

Dr. Connolly: Something like that. The assumption was that this could be due to dronedarone and so extensive warnings were then applied to the drug and in Europe, monthly monitoring of liver function was either recommended or actually required for the first 6 months of treatment. It was a bit more flexible in the United States. Monitoring of liver function was recommended without giving a precise specification.

The interesting thing about that is that since then -- and it has now been a year and a half -- there has been no further such reports, even though the drug continues to be used, suggesting that perhaps these cases were sporadic rather than due to the drug because you would expect to see more cases. The final word is not in on that yet, and we will just need more time to know.

Dr. Myerburg: Any more information on hepatic function studies?

Dr. Connolly: The only new information comes from the PALLAS study, which I guess we are going to get to in a minute.

Dr. Myerburg: The other issues you mentioned were pulmonary, as well as some question about thyroid toxicity, which do not seem to be holding up. Is that correct?

Dr. Connolly: I don't think there is much concern that this drug causes thyroid disorder. There is a bit of concern that there may be pulmonary toxicity. These rare toxicities are difficult to sort out. You see a couple of cases of pulmonary fibrosis and ask, "Is it due to the drug or not?" There are other idiopathic causes of pulmonary fibrosis, so how do you attribute things that are occurring so infrequently to the drug or say they are not due to the drug? It is a dilemma. I would say that there is always concern with a drug that is related to another drug that can produce pulmonary fibrosis, but it is not definitive at this point whether dronedarone does or doesn't cause pulmonary fibrosis.

Dr. Myerburg: The history is that this drug gets approved and, at least in my community, is pretty well accepted. It is getting a wide play in the community and then these issues about toxicity surface -- and some of that is sort of an evolution -- and then PALLAS comes along, which is a recently published study that looked at a different population of patients. If you would, Stuart, tell us about the design and rationale for doing the PALLAS study?

Dr. Connolly: The PALLAS study was a dronedarone randomized trial in patients with permanent atrial fibrillation with risk factors for major vascular events.

Dr. Myerburg: The emphasis here is on permanent atrial fibrillation as opposed to paroxysmal or persistent atrial fibrillation.

Dr. Connolly: Yes, so the question is, "Why would we do a study in permanent atrial fibrillation with an antiarrhythmic drug?" The answer is that there are 2 sides to the rationale. One was a great desire to do another study with dronedarone with major vascular outcomes as the primary outcome because we had seen such positive results on cardiovascular death, stroke, and arrhythmic death from ATHENA, but they had not been the primary outcomes of that trial.

The second driver for the study was that there were a lot of physiologic effects of dronedarone – such as heart rate slowing of about 8 to 10 bpm in atrial fibrillation, a fairly noticeable antiarrhythmic effect in the ventricle with a reduction in sudden death, and antiadrenergic effects – unrelated to maintenance of sinus rhythm that seem to explain the benefits of the drug in coronary disease. Therefore, it was a reasonable hypothesis that this drug was having its beneficial effects independent of its effect on recurrence of atrial fibrillation.

Dr. Myerburg: What you are saying is that there were signals that were not strong enough to lead to any conclusions, but were hypothesis generating about things beyond the direct effect on preventing recurrence of atrial fibrillation, and that is what is standing behind this.

In terms of the study population of PALLAS, how did that differ from ATHENA? If you are interpreting the data, wouldn't this be an important component?

Dr. Connolly: We are now getting into all the puzzles about ATHENA and PALLAS, of which there are many. So puzzle number 1: we attempted to design PALLAS so that it would have a higher-risk population than those who had gone into ATHENA. We thought that if this drug was going to have an effect on vascular events it would do so in patients most at risk for vascular events. So we included patients with heart failure, patients with coronary disease, patients with prior stroke, and patients with low ejection fraction, hoping to get a higher-risk population. However, when you look at the actual mortality rate in PALLAS, it is lower than the mortality rate on an annualized basis that we saw in ATHENA. So we tried to get a high-risk population. On paper they looked like they should be high risk but in fact, they were not higher risk than that in ATHENA.

Dr. Myerburg: With that as a background, what about the endpoint? What were the endpoints in PALLAS and how did the study turn out?

Dr. Connolly: The primary endpoint in PALLAS was split in 2. We first had major vascular events (eg, stroke, myocardial infarction, embolism, or cardiovascular death) and then we also had a second co-primary outcome, which was cardiovascular hospitalization or death. That had been the primary outcome in the ATHENA study. With those 2 outcomes we started enrolling patients. After about 1 year we started having very serious communications with the data monitoring board, which led fairly quickly, over a couple of months, to termination of the study because of a pretty strong signal that dronedarone was harmful in this population of patients with permanent atrial fibrillation.

Dr. Myerburg: Put this in perspective. The original intent was to enroll more than 10,000 patients divided between the 2 arms of the study. How many patients had actually been enrolled at the time the study was stopped? It is an indicator of how strong that signal was.

Dr. Connolly: Only 3000 patients had been enrolled, but the other thing was the follow-up. We had originally planned to have more than 800 primary outcome events, but the study was stopped before we even had 50, so we really had only a tiny fraction of what we had thought was necessary to show a benefit and yet we had a fairly clear signal for harm. It was still a relatively weak signal in some ways in that we only had 50 primary events, but you are correct. There is enough evidence here that this drug is harmful that the study needed to be stopped and I do not think anyone has disagreed with that.

Dr. Myerburg: So what are the consequences of this, first, in terms of consequences just from the interpretation of the data as an overview and then secondly the consequences from the point of view of approvals by the various agencies around the world?

Dr. Connolly: First, I just want to briefly say what the main results were. To say there was a signal for harm does not quite capture it. There was an increase in cardiovascular death. It was an approximately 2-fold increase, which is borderline statistically significant. There was an increase in stroke, and although there were very few strokes it was nominally significant.

There was an increase in hospitalization for heart failure, and if you look at the composite outcome of hospitalization for heart failure or heart failure episodes there were quite a lot of events. The signal is, I think, a very strong indication of clear evidence for harm for that particular outcome, so those are the main findings.

Of course, these findings were reported immediately to the regulatory agencies and they reacted quite appropriately to limit the use of dronedarone in patients with permanent atrial fibrillation primarily, and also in some segments of patients with paroxysmal atrial fibrillation in whom it had previously been used. It is now pretty clear that you do not want to use this drug in patients who are at risk for heart failure. The Europeans have also issued a caution about using it in patients with coronary disease.

Dr. Myerburg: From your point of view, and I am just asking for your opinion and I know that you do not have data on this, but for patients who have a profile of paroxysmal atrial fibrillation similar to those in PALLAS, what is your opinion on the status of risk in those patients?

Dr. Connolly: We have the puzzle of PALLAS and ATHENA. You have 1 trial that shows a reduction in cardiovascular death, a reduction in stroke, and a reduction in arrhythmic death. Then you have the other trial that shows increases in all 3 of those. You ask, "How do the trials differ?" Well, all the patients had atrial fibrillation, so that is the same. Second, the patients in the PALLAS study had more risk factors for vascular events than those in the ATHENA study, but there was some considerable overlap. In fact, when you look at the raw numbers, the actual mortality and event rates in ATHENA were higher than those in PALLAS, so to argue that the PALLAS patients are sicker runs up against that observation.

What is the big difference? The one big difference that I keep coming back to, having thought about this quite a bit, is the difference between paroxysmal and permanent atrial fibrillation, which I think has turned out to be very important and, indeed, much more important than I previously had suspected it would be.

Dr. Myerburg: In terms of the approvals, the indications, the limitations, and the recommendations for follow-up, how did that come about considering that there is still probably going to be more word coming from the FDA and the European Medicines Agency (EMA)?

Dr. Connolly: The regulators have indicated that this drug should not be used in patients with permanent atrial fibrillation. The concern, of course, is that the patient who starts off in paroxysmal or persistent atrial fibrillation can progress to permanent atrial fibrillation. Therefore, the agencies are cautioning physicians to follow their patients for evidence that their atrial fibrillation has, in fact, become permanent, at which point the drug should be discontinued.

Exactly how that is done has been indicated somewhat differently in Europe and in the United States. Like everyone else, the regulators are groping a little bit to find a rational way to do this. Without getting into the details, the fundamental message for the clinician is that you cannot start this drug and then forget about the patient. You have to continue to monitor the patient, especially for the development of permanent atrial fibrillation but also for the development of heart failure or even risk factors for heart failure.

Dr. Myerburg: From the US perspective, the recommendation is to determine rhythm every 3 months. They weren't specific about it but what that really means is an ECG. In Europe, the recommendation is to determine rhythm every 6 months plus continued monitoring of cardiac, hepatic, and pulmonary status, so it becomes more difficult. There is a lot more to do.

What about the options for the practitioner? Where is the practitioner today in terms of managing patients with atrial fibrillation in whom you consider dronedarone and then you run into this problem of recurrent episodes of atrial fibrillation on a monitored ECG? Do you do a repeat cardioversion to try to stay ahead of it? Is that a rational approach?

Dr. Connolly: That is how we manage persistent atrial fibrillation, isn't it?

Dr. Myerburg: Yes.

Dr. Connolly: Almost by definition you are talking about a patient who needs to be cardioverted once in a while.

Dr. Myerburg: Right.

Dr. Connolly: I think if you are prepared to do those cardioversions, and if your patient is prepared to go along with you, then you have some rationale for continuing dronedarone if the patient is doing well on it. Of course, there are patients who do well on it. Those of us who use the drug have seen patients who do not want to come off of it even after they hear this bad news.

Dr. Myerburg: Is this going to drive us toward greater application of the AFFIRM strategy of rate control?

Dr. Connolly: In other words, is it going to enhance the rate-control approach over and above the rhythm-control approach?

Dr. Myerburg: Yes.

Dr. Connolly: I think that is probably true, and this is more bad news for antiarrhythmic drugs just after we thought we got some good news with the ATHENA results. We now get some bad news that brings us back down to earth. It certainly brought me down to earth.

Dr. Myerburg: I have a concern about the rate-control emphasis. For some people it is obviously perfectly appropriate, but I would far prefer to have patients in sinus rhythm long-term, simply thinking of the patient who has evolving heart disease and some day may need their atrial kick and I do not want to see that atrium completely remodeled over time. So I am concerned about a rate-control approach as an equivalent option except in those patients in whom it is the only option, but this may drive things in that direction.

Dr. Connolly: Just seeing as how you brought it up, my approach is more on the rate-control approach. If I have a patient who is not symptomatic from their atrial fibrillation, I have a lot of trouble getting them to take antiarrhythmic drugs or especially to go for an ablation because there is no proven benefit. If you are not going to get symptom control, which you cannot in an asymptomatic patient, we have no evidence that maintaining sinus rhythm by any means is beneficial in the long-term, so I tend to go pretty softly on that.

Dr. Myerburg: Okay, so we move in slightly opposite directions.

Dr. Connolly: We do, but that is fine.

Dr. Myerburg: What about ablations? Is this going to drive us to do more ablations?

Dr. Connolly: One of the reasons that ablation has become popular in spite of its very significant limitations, both in efficacy and in safety, is the fact that the antiarrhythmic drugs do not provide anything much better, so I think this will probably push along the ablation agenda.

Dr. Myerburg: But ablation has its own set of issues.

Dr. Connolly: The major problem with ablation is that it has not been shown to have any substantial long-term benefits in any important cardiovascular outcome other than recurrence of atrial fibrillation, and even in that it is mostly a series of very small studies.

Dr. Myerburg: A final point on antiarrhythmic drugs. Where are we in terms of anticipation of new antiarrhythmic agents coming down the pike? Do you hold out any hope for this? The second component of that is could there be a drift back to the old antiarrhythmic agents, thinking of sotalol and propafenone?

Dr. Connolly: On the latter point, I don't see that. The one thing about dronedarone that we can say is that although we have shown that it has some significant limitations, we at least know what it does, whereas with the other drugs we do not have the large trials telling us whether they are safe or not safe. It may be a case of "ignorance is bliss" with those drugs, but that is not to say that I don't use them.

Dr. Myerburg: Yes, dronedarone is the most studied drug among the antiarrhythmic agents in terms of good formal studies.

Dr. Connolly: It is, by far, the most thoroughly studied of the antiarrhythmic drugs by an order of magnitude with the possible exception of amiodarone, which has also gone through a number of large trials, and that is one of the reasons we use the drug. We have some reassurance of its overall lack of effect on mortality.

Dr. Myerburg: The issue at this point is to keep our eyes and ears open. There is probably more information coming down the pike and refinement of the initial reaction to this as time goes on, but there are new clear limitations and guidelines for the use of this drug. It behooves us to adhere to that for now.

Stuart, thank you very much for sharing your knowledge and experience with us on this topic, with which I know you have been very much involved from the earliest years. I thank you for your help.