Amiodarone to prevent recurrence of atrial fibrillation. Canadian Trial of Atrial Fibrillation Investigators. Roy D, Talajic M, Dorian P, Connolly S, Eisenberg MJ, Green M et al.
Overall Study Question
To determine whether amiodarone is superior to sotalol or propafenone in converting and maintaining normal sinus rhythm in patients with a recent history of symptomatic atrial fibrillation (AF).
Patients: Subjects who had an electrocadiographically confirmed episode of symptomatic atrial fibrillation (lasting 10 minutes or longer) within the preceding six months for which long-term antiarrhythmic drug therapy was planned.
Outcome variables: primary end point was length of time to a first electrocardiographically confirmed recurrence of symptomatic atrial fibrillation (lasting 10 minutes or more).
Intervention: Amiodarone was compared to the aggregate of Sotalol and Propafenone (patients randomized in 2:1:1 ratio for A:S:P)
Amiodarone: 10 mg per kilogram of body weight each day for 14 days, followed by 300 mg per day for 4 weeks, then daily maintenance dose of 200 mg.
Sotalol: 80-160 mg bid, depending on age, sex and renal function.
Propafenone: 150mg q8h – 300mg q12h, depending on weight and age. All patients who did not convert to NSR within several days of starting drug therapy were electrically cardioverted. All patients were anticoagulated.
Are the Results of the Study Valid?
Was assignment of patients randomized?
Yes. Patients were randomized to amiodarone or to either sotalol or propafenone (2:1:1 randomization).
Were all patients who entered the trial properly accounted for and attributed at its conclusion?
Yes. All patients were accounted for and an intention-to-treat analysis was used. The investigators reported that 2.5% of patients were lost to follow-up. Eleven sotalol and 13 propafenone patients received another antiarrhythmic drug during the trial period.
- Were patients, their clinicians, and study personnel ‘blind’ to treatment?
- No. This study was open-label. Primary events were subjective (eg. symptoms of AF) and only resulted in an ECG evaluation if spontaneously reported by the subjects. Thus, clinician/patient biases about the drugs may have influenced the likelihood of reporting of symptoms. Treatment assignment was unaffected by the patients’ presentation, and therefore was unaffected by the open-label design.
- Were the groups similar at the start of the trial?
Yes. Except for the fact that more patients in the sotalol/propafenone groups had left ventricular hypertrophy (LVH) (21% vs. 13%). No adjustment was reported for this factor in the main results. However, a subgroup analysis presented in the paper shows that patients with LVH were associated with a higher risk of AF recurrence (i.e. treatment failure). Thus, it has been hypothesized that if the incidence of LVH was the same in both groups, the difference in efficacy between amiodarone and sotalol/propafenone would have been smaller.
- Aside from the experimental intervention, were the groups treated equally?
Yes. Except for the fact that amiodarone patients received more chest X-rays and thyroid function tests than other patients for the purpose of monitoring for adverse effects.
- Overall, are the results of the study valid?
What were the Results?
How large was the treatment effect?
Over the 468-day follow-up period, 35% of amiodarone and 63% of sotalol or propafenone-treated patients experienced recurrence of atrial fibrillation (NNT with amiodarone vs. sotalol/propafenone for 1.3 years to achieve one additional recurrence-free patient = 4 (p<0.001)). Because a minority of patients in the amiodarone group experienced the primary endpoint, the median time to recurrence could not be calculated. For the sotalol/propafenone groups, the median time to recurrence was 98 days. There was no difference between sotolol and propafenone. There were no differences in overall mortality or arrhythmic death. Thirty-four percent of amiodarone and 46% of sotalol/propafenone patients discontinued due to adverse effects (NNH for one additional drug discontinuation over 1.3 years = 9 (p<0.01)).
How precise was the estimate of the treatment effect?
The hazard ratio for recurrence of atrial fibrillation in the amiodarone group was 0.43 (95% CI 0.32 – 0.57). These confidence intervals are fairly narrow and even at their upper bound, represent a clinical effect which is likely to be clinically significant.
Will the Results Help Me in Caring for My Patients?
- Can the results be applied to my patient care?
- Yes. These results represent the first compelling data of the superiority efficacy of amiodarone versus commonly used antiarrhythmics for chronic prevention of AF. The study also demonstrated that low-dose amiodarone is better tolerated than the other drugs and does not cause serious adverse effects over the length of this study.
Were all clinically important outcomes considered?
- Are the likely treatment benefits worth the potential harms and costs?
- Yes. Serious toxicities (e.g. pulmonary fibrosis, thyroid dysfunction, ventricular arrhythmias, oculotoxicity) were rare and not more common than observed in the sotolol/propafenone groups.
This Medical Research Council of Canada-funded and well-designed study provides convincing support for the preferential use of amiodarone in patients with AF for the chronic suppression of symptoms. However, the open label design may have resulted in an expectation bias on the part of investigators and/or patients who expected amiodarone to be more effective for whatever reason. This may have resulted in an overestimation of the benefits of amiodarone. There was a high degree of initial conversion to NSR in all three groups during the first 3 weeks of therapy, but this was not maintained as well in the sotalol/propafenone groups. Only 12% of patients had left ventricular dysfunction and the applicability of these results to such patients is questionable, despite the commonly-encountered clinical scenario of AF in the presence of congestive heart failure. Despite this, amiodarone is likely the safest of the three drugs studied in patients with left ventricular dysfunction. Longer-term and larger trials are required to determine whether the benefits seen in this trial confer other morbidity (e.g. stroke, cardiovascular events) and/or mortality benefits.