When the antiarrhythmic dronedarone first made its appearance in 2009, it offered great promise: its lack of iodine, compared to amiodarone, would lead to reduced toxic effects on the thyroid and other organs while proving maintained efficacy.
In a broad range of patients with paroxysmal, persistent or permanent atrial fibrillation and heart failure, the DAFNE (2), EURIDIS (3), ADONIS (3) and ERATO (4) studies, dronedarone appeared safe and showed efficacy for rhythm control. Compared with amiodarone in DIONYSOS (5) it also had a good safety profile - albeit showing less efficacy in the latter study.
Dronedarone was then studied in a series of three large phase III studies-ATHENA (6), ANDROMEDA (7) and PALLAS (8); the latest of which found that dronedarone in permanent atrial fibrillation had a two fold increase in all-cause death, cardiovascular death, arrhythmic death, stroke, myocardial infarction, cardiovascular (CV) hospitalisation and hospitalisation for heart failure.
ATHENA & ANDROMEDA
ATHENA (6) randomised 4629 subjects with paroxysmal or persistent atrial fibrillation. Dronaderone showed a significant reduction in its primary composite endpoint of cardiovascular hospitalisations or death from any cause (24%, p<0.001). There was a consistent positive effect seen in reductions achieved, CV death (29%, p=0.03), arrhythmic death (45%, p=0.01) although all cause mortality was not significantly reduced (16%, p=0-.18). In the permanent atrial fibrillation subgroup of patients there were trends for favourable CV outcomes (26% reduction in hospitalisation and death, 20% reduction in stroke, acute coronary syndromes and CV death).
ANDROMEDA (7) evaluated 627 subjects with heart failure (without atrial fibrillation) as a possible other indication for this agent however the study was terminated early due to a two-fold increase in mortality (25 in the dronedarone group and 12 in the placebo group, HR=2.13; CI-1.07-4.25, p=0.03). The primary combined endpoint of all-cause mortality or hospitalisation for worsening heart failure was not significantly different between the two groups.
Following ATHENA, which studied subjects with paroxysmal or persistent atrial fibrillation and ANDROMEDA which studied subject with heart failure PALLAS was designed to study patients with permanent atrial fibrillation. Subjects with active decompensated heart failure were excluded. PALLAS was also halted abruptly for excess mortality. Full results were presented at the 2011 AHA scientific sessions which were also published in the New England Journal of Medicine (8). PALLAS recruited 3236 patients, and was halted at a median follow up of 3.5 months for two fold increases in the first composite primary outcome of stroke, systemic embolism, myocardial infarction or CV death (HR-2.29, C.I-134-3.94, p=0.002) and the second composite primary outcome of unplanned CV hospitalisation or death (HR-1.95, C.I= 1.45-2.64, p<0.001). All-cause death, CV death, arrhythmic death, stroke, MI, CV hospitalisation and hospitalisation for heart failure were all significantly higher in the dronedarone treated patients (Table), with a consistent hazard seen across all-predefined subgroups. The fact that 21% subjects discontinued the study drug in the dronedarone arm versus 11% in the placebo arm highlights the poor tolerance of the drug in this subset of patients.
Such different results. What happened?
In trying to assess the reasons for the diametrically opposed results of ATHENA relative to PALLAS, some thoughts come to mind:
- While no previous trials like PALLAS were carried out specifically in permanent AF patients, various previous experiences with antiarrhythmic drugs such as quinidine (10), d-sotalol (11), encainide (12), moricizine (13), flecainide (14) and amiodarone (15) have shown similar adverse results.
- The hints of problems seen in ANDROMEDA in patients with heart failure, may have been true after all. There does appear to be an issue, albeit not fully defined, that increased cardiovascular event rates of all types are now seen with this agent.
While one is left to wonder about what is electrophysiologically so different in permanent atrial fibrillation, the results of PALLAS reflected in the recommendations from the US Federal Drug Administration (16) and the European Medicines Agency (17) are the following:
- Avoid dronedarone 1) in patients with permanent atrial fibrillation 2) previous amiodarone related liver toxicity 3) current symptoms or past symptoms of HF 4) left ventricular systolic dysfunction (EF <35%)*.
- Consider dronedarone as still an option 1) for paroxysmal or persistent atrial fibrillation patients who present in sinus rhythm and are clinically stable (EMA), 2) in patients who are proposed to be cardioverted (FDA).
- Monitor patients on dronedarone every 3 months for their heart rhythm.
- Keep in mind that in permanent atrial fibrillation antiarrhythmic drugs carry significant risks with little benefit.
- Focus on rate control and adequate antithrombotic therapy.
*only in FDA, in the context of Class IV NYHA heart failure.
Table 1. Cardiovascular outcomes - PALLAS
|End point||Dronedarone versus placebo|
Hazard ratio (95% Confidence Interval)
|Co-primary end point: stroke, MI, systemic embolism, or CV death||2.29 (1.34-3.94)||0.002|
|Co-primary end point: death or unplanned CV hospitalisation||1.95 (1.45-2.62)||<0.001|
|CV death||2.11 (1.00-4.49)||0.046|
|Arrhythmic death||3.26 (1.06-10.00)||0.03|
|HF hospitalisation||1.81 (1.10-2.99)||0.02|
|HF event or hospitalisation||2.16 (1.57-2.98)||<0.001|