Experts Hash Over New Review Challenging Dronedarone's Value in AF

April 07, 2010

April 7, 2010 (Los Angeles, California) - Is dronedarone (Multaq, Sanofi-Aventis) overrated as treatment for atrial fibrillation (AF)? Yes, suggests a new review, which contends, controversially, that the drug's role in clinical practice should be primarily as an alternative when, once a rate-control strategy fails as an option, other antiarrhythmics don't suppress AF or aren't well tolerated [1].

Moreover, it says, limitations in the design and execution of the key trial supporting dronedarone's approval, ATHENA, "raise questions about the quality of [its] data" and "cast doubts on their relevance to clinical practice."

Based on its survey of the major dronedarone trials, "the available data support only limited use of dronedarone for select patient populations, mostly as a second- or third-line agent in lieu of amiodarone," according to the article published online today in the Journal of the American College of Cardiology (JACC).

Dronedarone is widely perceived, based on trial evidence and clinical experience, as safer and better tolerated but also less efficacious at preventing AF than its molecular cousin amiodarone.

In practice, clinicians take patient preferences into account when there's more than one treatment option, and "there are some patients who might consider improved short-term tolerability over reduced efficacy an acceptable trade-off," observed Dr Sanjay Kaul (Cedars Sinai Medical Center, Los Angeles, CA) for heartwire . Kaul is the senior author on the report, which is slated for the journal's April 13, 2010 issue; the first is Dr David Singh (Cedars Sinai Medical Center).

But, Kaul continued, the ATHENA trial did not show symptomatic or quality-of-life improvement among patients receiving dronedarone and that, together with "its modest antiarrhythmic efficacy, lack of well-established safety advantage over amiodarone, a huge cost disadvantage compared with amiodarone, and the lingering controversy about the primacy of rhythm- vs rate-control strategy, [makes me unable to] envision dronedarone being recommended as a first-line therapy."

Clinical practice has evolved. I am not convinced that dronedarone offers us a safer alternative.

Kaul made it clear to heartwire that his safety comparison of the two drugs refers to amiodarone given at 200 mg/day, half the maintenance dosage tested in some older trials. "Most of us are very comfortable with using 200-mg amiodarone daily," he said. At that level, "I have not seen the kind of adverse-[effect] profile that we used to see with high-dose amiodarone. . . . Clinical practice has evolved. I am not convinced that dronedarone offers us a safer alternative."

The group's report has rankled some AF experts contacted by heartwire , who attacked its line of reasoning and conclusions and were far warier of amiodarone toxicity. They said the authors missed the point of ATHENA, and several held that dronedarone, far from being an also-ran among AF drugs, is a game-changer that should help shift the focus of AF management away from "rate control vs rhythm control" to improved clinical outcomes.

An editorial [2] accompanying the JACC review also took issue with its assessment of dronedarone's relative safety, efficacy, and contribution to patient care.

"If 'safety first' is the issue, the data for dronedarone are superior to those of other antiarrhythmic drugs for low/moderate-risk patients," according to the editorialists, led by Dr Christian Torp-Pedersen (University of Copenhagen, Denmark).

When rhythm control is considered necessary or preferred by the patient, they write, "decent advice would require that patients are informed of the unknown safety of most compounds and of the common problems with long-term amiodarone therapy. With dronedarone, it is for the first time possible to provide reassurance to large groups of low-risk patients."

Time to First Recurrence: "Great for Statisticians But Not the Way We Practice"

As heartwire has previously reported, the ATHENA trial showed a 24% drop in the primary end point of death from any cause or CV hospitalization and a 26% decline in CV hospitalization on its own (p<0.001 for both differences) for dronedarone vs placebo [3]. There were no big safety issues in the 4628-patient trial, according to the investigators. Entry had required patients to have AF and at least one other CV risk factor.

"Dronedarone is not the best thing since sliced bread," observed Dr Albert L Waldo (University Hospitals of Cleveland, OH). "But the reason we have all these [many] rate- and rhythm-control drugs is because none are as good as we'd like or as safe as we'd like. And dronedarone, I think, is right there in the pack with most of them," he said to heartwire .

We're finally talking about outcomes! This drug decreases hospitalizations.

"Amiodarone is better than all of them for maintaining sinus rhythm the longest, but it also has the longest list of adverse effects, and most of them are not fun. So [for the review] to pooh-pooh dronedarone the way it does, it doesn't deserve that."

Moreover, Waldo said, "ATHENA should get credit for finally changing things from 'time to first AF recurrence' to the way we really practice. Recurrence is not failure per se. The frequency of AF is important. The symptoms when AF recurs are important. Its duration, when it recurs, is important." But the review's lackluster appraisal of dronedarone efficacy is based primarily on trials looking simply at "time to AF recurrence."

In AF drug trials, "time to first recurrence had been the way to measure efficacy for a long time. It was great for the statisticians, but it's not the way we practice," he said. "The importance of dronedarone is not only that it's a new drug [for AF], but that it's a paradigm shift in how we think about drugs [for AF]. We're finally talking about outcomes! This drug decreases hospitalizations."

Where to Draw the "Line"

Dr James A Reiffel (Columbia University, New York, NY) similarly noted for heartwire that dronedarone "was approved not for the suppression of atrial fib, but for reduction of cardiovascular hospitalizations" and so isn't like other antiarrhythmics used in AF--nor should clinicians look at it the same way. He called dronedarone "the most appropriate drug" for patients with AF plus at least one of the markers of increased risk in its labeling.

Also in agreement was ATHENA co–principal investigator, Dr Stuart J Connolly (McMaster University, Hamilton, ON), who told heartwire the trial was "an important milestone in antiarrhythmic therapy showing for the first time that an antiarrhythmic drug can have important positive effects on cardiovascular outcomes."

Connolly agreed that dronedarone isn't as efficacious as amiodarone in suppressing AF and that "sometimes emphasizing efficacy over safety is appropriate." He said he wouldn't advocate dronedarone as a first-line agent against AF in all patients.

"But if having a drug that you can use without hospitalizing your patient to initiate it, that has a good safety profile in patients who don't have heart failure, if safety is a key concern, then it could be a first-line drug for many patients."

Connolly cautioned that labeling any drug as first-line or second-line "is to some extent, a little artificial," because every case is different.

If safety is a key concern, then it could be a first-line drug for many patients.

Kaul agreed that "treatment decisions should be based on a very careful assessment of benefit vs risk of the drugs, but also, more importantly, on patient preferences."

In fact, everyone heartwire contacted for this story made a similar point, and none more strongly than Dr Peter R Kowey (Lankenau Hospital, Wynnewood, PA): "I am adamantly opposed to declaring any therapy as first-, second-, or any line. It is all about the individual patient."

Such language, several experts observed for heartwire , is more useful in guidelines documents, and the JACC review does provide an algorithm for a rhythm-control strategy in flowchart form that is based on the current guidelines, with some modifications.

For example, for patients with no or only minimal heart disease, including those with hypertension but no LV hypertrophy, it recommends flecainide, propafenone, and sotalol as first-line agents "based on their proven safety and efficacy in this population." For patients with coronary disease but no heart failure, dofetilide and sotalol are the recommended first-line agents, but "dronedarone might be a reasonable alternative to these drugs or to amiodarone."

According to Kaul, "If you look at the algorithm we gave, the majority of the use of dronedarone is confined to a third-line choice, and in some instances a second-line choice." But their review, he said, also emphasizes consideration of the individual patient.

Reiffel took issue with the JACC review's perception of dronedarone as primarily a fallback drug or alternative to amiodarone. Treatment algorithms, he said, are based on "safety first," not efficacy, and that's especially relevant for an arrhythmia that isn't life-threatening. "Dronedarone is a safer drug than amiodarone, so wherever you put it [in the treatment algorithm], it should not be on the same level as amiodarone. That's makes no sense at all."

Off-Label Use Anticipated

Kaul and his colleagues write that they anticipate dronedarone "will likely be used for a broader population and for indications beyond its approved label," which specifies the drug is for low- to intermediate-risk patients with a history of nonpermanent AF or atrial flutter but no advanced or recently decompensated heart failure. "Although we recognize that reconciling efficacy vs safety is ultimately a matter of clinical judgment and patient preference, we nevertheless caution against the indiscriminate use of dronedarone."

To heartwire , Kaul said such dronedarone overuse, should it happen, will have been encouraged by what he sees as "spin and hype" over the drug's clinical-trial performance from Sanofi-Aventis and some clinicians. He also said he thinks they are "overselling" dronedarone's relative safety and understating its only modest efficacy.

I think the article hits a good, cautionary note, not to go overboard with this drug.

Dr D George Wyse (University of Calgary, AB) seemed to agree. The authors of the JACC review "are on the right track. I suspect there's been a fair bit of hype around dronedarone after ATHENA," he told heartwire . Their paper, he said, puts "some balance into the picture that we haven't seen until now."

Amiodarone is clearly more efficacious against AF, he observed, "but almost never do I use amiodarone first."

Still, Wyse said, "if [the patient] is really symptomatic, and they've already failed on propafenone, flecainide, or dofetilide, I wouldn't go to dronedarone. I'd go to amiodarone. Because once they've failed a drug or two, dronedarone isn't going to work any better."

So, he added, "I think the [JACC] article hits a good, cautionary note, not to go overboard with this drug."

Questioning the "Quality" of ATHENA's Data

The JACC review discusses at least six major dronedarone trials in some depth, and none more so than the ATHENA trial, which was the primary focus of the March 2009 FDA advisory panel that overwhelmingly recommended the drug be approved. As a member of the panel who voted in favor of approval, Kaul nonetheless said during the proceedings, as reported then by heartwire , "I'm struggling to find a proper role for this drug in clinical practice."

Part of his hesitation about the drug, he then noted, came from doubts about the ATHENA trial's methods and results, and his belief that its population was skewed toward lower risk.

The review reiterates these concerns, and to heartwire he said that "the quality of the data in the ATHENA trial does not stand up to scrutiny."

For example, after the ANDROMEDA trial's finding of elevated mortality in dronedarone recipients with severe systolic heart failure, to design ATHENA, "the sponsor went back and carefully chose their patient population, which was essentially a low-risk population" that excluded patients with severe heart failure and a history of permanent AF, Kaul said.

"In my mind, the patient population that was selected in the ATHENA trial does not represent what we see in clinical practice."

Disagreeing, editorialists Torp-Pedersen et al describe ATHENA as "the first large trial to provide safety data in a moderate-risk population that [constitutes] a large part of the relevant patients for rhythm-control therapy."

The mechanism behind what it's doing is unclear. It's really kind of puzzling.

But Kaul, who had access to unpublished ATHENA data presented to the FDA panel, also said there were interim adjustments to the trial's patient-enrollment target. Some clinical end points that hadn't shown a significant difference at the prespecified enrollment became significantly different in favor of dronedarone when more patients were entered, he said.

Kaul slammed the trial for its reliance on clinical end points that were not adjudicated and lack of systematic documentation of symptom and quality-of-life outcomes. "And we have no idea what drove cardiovascular-hospitalization events, which was the main driver of benefit from dronedarone in ATHENA."

Wyse also wonders why dronedarone appeared to prevent cardiovascular hospitalizations. He pointed to a secondary ATHENA analysis presented at the November 2008 AHA Scientific Sessions (and reported then by heartwire ) of 473 patients who were never in sinus rhythm at any ECG conducted for the study and were classified as having "permanent" AF. Those taking dronedarone showed a 26% drop in risk of CV hospitalization or death compared with controls. It didn't reach significance, but the reduction mirrored the benefit in the overall population, he noted.

"To me, that argues strongly against the favorable impact of [dronedarone] having anything to do with preventing atrial fibrillation."

In addition, Wyse observed, the ATHENA investigators reported a significantly reduced risk of stroke among patients receiving dronedarone, as reported by heartwire . But for the degree of dronedarone's antiarrhythmic effect in the trial, he said, "it makes no sense why stroke would be reduced."

He noted that dronedarone appears to have some ancillary effects that may be contributing to the benefit, such as blood-pressure reduction, control of heart rate in patients who remain in AF, "and maybe it has some ventricular antiarrhythmic effects."

Still, "the mechanism behind what it's doing is unclear. It's really kind of puzzling."

Questioning the Quality of the Case Against Dronedarone

Several of the AF experts interviewed for this story profoundly criticized the JACC review authors' analysis of the major dronedarone trials and, especially, their reasoning that consigned dronedarone to second or third place among drugs for AF.

Reiffel said it's "scientifically invalid" to compare the efficacy of two drugs unless it's within the same trial. Issues with comparing the populations of two trials could include, for example, different AF durations, history of attempted antiarrhythmic agents, and use of rate-controlling drugs.

Therefore, according to Reiffel, a conclusion that dronedarone's efficacy against AF is modest based on trials like EURIDIS and ADONIS, for example, and therefore it should be a second-line agent, "isn't fair, because you don't know if any of the drugs you now consider first-line would have performed any better in that population."

The editorialists make much the same point: "The effect of dronedarone is moderate, as demonstrated in the review by Singh et al, but all antiarrhythmic drugs have moderate efficacy. Any attempt to compare dronedarone with flecainide, sotalol, or other drugs on the market is futile because there are no comparable trials, and comparison with historical data is often misleading."

How dronedarone's efficacy in a rhythm-control strategy for AF compares with that of flecainide, sotalol, or propafenone "needs to be tested in randomized trials," write Torp-Pedersen et al.

False Impressions

Another issue with the JACC review, according to Waldo, is that it "gives a false impression about amiodarone. It doesn't say much about its adverse effects. . . . Its interaction with warfarin is a major problem, besides pulmonary fibrosis, thyroid dysfunction, optic neuritis--the list just goes on and on." And the longer a patient is on the drug, the likelier such side effects become, he noted.

In addition, Connolly said, "It's surprising that they're recommending [before dronedarone] the use of flecainide, propafenone, and sotalol, which [represent] classes of drugs that have been shown to have fairly significant hazards in important patient groups. They are patently unsafe, whereas dronedarone has been studied in the same sorts of patients and shown to be extremely safe."

According to the editorialists, "for patients with low risk, dronedarone provides the only antiarrhythmic drug with a large safety database to prove reasonable safety. The safety knowledge of dronedarone may result in patient and physician preference of dronedarone as first-line therapy, with a possible switch to amiodarone when sinus rhythm is no longer maintained."

Kaul called that scenario "an example of enthusiasm exceeding the evidence!"

The editorialists also write that the JACC review "boldly states that flecainide, propafenone, and sotalol are safe for large groups of patients." Kaul said the review "does no such thing."

But he did tell heartwire that "in correctly selected patients, that is [those] without LV dysfunction and coronary disease, the safety record of flecainide and propafenone is excellent." He also said, "The safety of sotalol and dofetilide is somewhat more challenging, as they no doubt prolong the QT [interval] and can lead to sudden death. Patients on these drugs need to be monitored carefully."

Kaul said he has no disclosures. His coauthor Dr George A Diamond (Cedars Sinai Medical Center) reports being on a speakers' panel for Merck. Torp-Pedersen reports being on the ANDROMEDA and ATHENA steering committees and that he has consulted for Sanofi-Aventis, Abbott, Cardiome, Astellas, Merck, and Novo and has presented at meetings sponsored by Sanofi-Aventis. His coauthor Dr Ole Dyg Pedersen (University of Copenhagen) reports consulting for Sanofi-Aventis; another coauthor, Dr Lars Køber (University of Copenhagen) was the primary investigator for ANDROMEDA; has given lectures at meetings sponsored by Sanofi-Aventis; is on steering committees of trials sponsored by Novartis, Sanofi-Aventis, and Servier; and is an advisory board member for Novartis and Sanofi-Aventis. Reiffel said he has "worked with virtually all the antiarrhythmics that have come to market prior to their release"; was an "investigator or advisor" to the companies behind all antiarrhythmic drugs mentioned in this article; and was an investigator or consultant for the ADONIS and ATHENA trials of dronedarone. Connolly was co–principal investigator of ATHENA and reports receiving research grants and consultant and/or lecture fees from Sanofi-Aventis. Kowey reports being a consultant for or receiving honoraria from Sanofi-Aventis, Merck, Boehringer-Ingelheim, Astellas, GlaxoSmithKline, AstraZeneca, Johnson & Johnson, Pfizer, ARYx, Portola, and Bristol-Myers Squibb. Wyse reports being on advisory boards for Merck and Bayer and a speaker for Sanofi-Aventis and serving on data safety monitoring boards or as a grant reviewer for trials associated with Sanofi-Aventis, Boehringer Ingelheim, Medtronic, Bristol-Myers Squibb, Biotronik, and Boston Scientific. Waldo is on the data safety monitoring boards of trials associated with device companies that make treatments for AF and has previously reported receiving consulting fees or honoraria from Sanofi-Aventis, Bristol-Myers Squibb, Boehringer Ingelheim, Astellas, Portola, Daiichi Sankyo, and Ortho-McNeill-Jannsen.