Cardioversion for Recent-Onset Atrial Fibrillation: Early, Delayed, or Not at All?

In the study, 437 patients with symptomatic, stable, recent-onset AF were randomized to early or delayed cardioversion. In the early group, patients were pharmacologically cardioverted with flecainide. If pharmacologic cardioversion failed – or patients had a prior history of failed pharmacologic cardioversion or a contraindication to pharmacologic cardioversion – the patients were electrically cardioverted. 

In the delayed group, patients were first rate-controlled with the physician’s drug of choice and discharged. They then returned for a repeat ECG 48 hours after onset of symptoms. If still in AF at 48 hours, they were referred to the ED where physicians cardioverted patients using their preferred method, pharmacologic or electrical.

In the early group, 16% spontaneously converted prior to any cardioversion attempts. A total of 95% of patients in the early group converted to sinus rhythm. In the delayed group, 70% spontaneously converted prior to any cardioversion attempts, a rate similar to findings in prior studies (7). A total of 96% in this group were successfully converted to sinus rhythm.

Patients returned 4 weeks later for an ECG. For the primary outcome – presence of normal sinus rhythm on ECG at 4 weeks – the authors found that delayed cardioversion was non-inferior to early cardioversion: 91% in delayed group, 94% in early group (between-group difference −2.9 percentage points; 95% confidence interval [CI] −8.2 to 2.2; P = 0.005 for noninferiority).

Quality of Life

Cardioversion of AF (rhythm-control) has not been shown to improve survival when compared to rate-control (8), but studies have shown that cardioversion may abate AF-related symptoms and improve quality of life (9,10,11,12,13). This study found no difference in quality of life at 4 weeks between early and delayed cardioversion. Yet, as others have pointed out (2), this does not necessarily mean that patients had equal satisfaction during those 4 weeks. This exact question may be answered after the authors complete 1 year follow-up and report patient reported experience measures.

Duration of Visit

While the average duration of index visit was less in the delayed group, this does not consider the resources needed for the 31% of patients in the delayed group who required two ED visits in 48 hours after failure to spontaneously convert. Additionally, this finding contradicts prior observational studies which found a shorter length of stay for patients who were immediately cardioverted (14). 

This conflicting finding perhaps suggests practice variation in the Netherlands, as the presence of both cardiologists and anesthesiologists was required for the administration of IV antidysrhythmic and conscious sedation, respectively. Perhaps a shorter length of stay would be found in the US, where ED physicians are considered competent to independently perform the interventions.

Complication Rates

The authors found that the cardiovascular complication rate within the four week follow-up period was not insignificant – 4.6% in delayed and 3.6% in the early group. When broken down into ‘life-threatening’ complications – TIA/CVA (n=2) or sinus arrest requiring 30 seconds of CPR (n=1) – the complication rate was just 0.9% in the delayed group and 0.5% in the early group. There were no mortalities. 

These rates are comparable to prior studies which found an overall complication rate of 7.6%, primarily hypotension and bradycardia, and a 0.7 – 0.9% rate of serious adverse events, primarily stroke and bradyarrhythmia (4,15,16,17). 

While the collective studies demonstrate a low rate of clinically significant adverse events, patients should be made aware of these potential risks when deciding whether or not to be cardioverted. Likewise, providers should be knowledgeable of the potential risks and be ready to provide resuscitative care.

Admission Rates

A very low percentage of patients were admitted: 1.4% in delayed group and 2.3% in early group. This hospitalization rate is far below the 60-70% rate observed in the US (18). The difference in hospitalization rates is multifactorial and reflects variation in approach to AF (19).  Reasons that may explain increased admission rates in the US include: lack of familiarity and comfort with AF management (20), lack of reliable access to follow up (18), more conservative approach of US cardiologists (19), and absence of ED-specific guidelines for AF management (21,22,23).

Given the low complication rates and low return ED visit rate (7% in both groups) and considering rising healthcare costs (24), we in the US may want to rethink our practice in hopes of safely discharging more patients with recent-onset AF in order to avoid the costs (and complications) related to unnecessary hospitalization.

The study was overall carried out fairly well. Patients were adequately randomized, however, physicians and patients were not blinded to the group allocation. Patients were analyzed based on the groups to which they were randomized. Follow-up was very good – only 10 patients (2.3%) were lost to follow up.

One source of bias was that both groups were not treated equally in terms of the initial method of cardioversion. A higher percentage of patients in the delayed group underwent initial electrical cardioversion without an attempt with drugs. Length of stay would be shorter for those patients who only required one method of cardioversion. This may have been a source of bias, explaining a shorter average length of stay for those in the delayed group.

For a deeper dive into the quality of study, listen here (start at 14:20 for discussion of bias). Some points raised by Glaser and Milne (3) include 1) skepticism over whether patients were chosen consecutively as to avoid bias, 2) greater number of patients in the delayed group had prior MI (24% vs 13%) 3) arbitrary selection of 10% for difference to determine non-inferiority 4) changes in secondary outcomes from those originally posted on www.Clinical Trials.gov, and 5) questionable use of intention to treat (ITT) analysis. As Glaser and Milne (3) point out, ITT is typically used in superiority trials as a conservative approach as it will bias results “to accepting the null hypothesis.” However, in a non-inferiority study, ITT may actually introduce bias in favor of “the conclusion of non-inferiority” and therefore per-protocol may have been a more appropriate analysis.

In the appendix of the Pluymaekers et al (1), the authors list additional end-points that will be reported after completion of follow up at 1 year. These include: 1) the number of recurrent paroxysms, 2) ED visits for dysrhythmia, and  3) utilization of antidysrhythmic drugs, rate control and ablation. Thus, while the authors found no difference between the two groups at 4 weeks, it is possible that there may be statistically significant differences in the long term.

Interestingly, the authors suggest that they will look at other patient characteristics including genetic analysis, variance in outcomes based on gender and time of day of presentation, differences in patients with and without heart failure, determinants of spontaneous conversion, and triggers of paroxysmal AF. These findings could have significant implications, as we could further identify those patients who would most and least benefit from early cardioversion. In light of increasing healthcare costs, the authors also will look at the potential cost savings between the two interventions and will calculate yearly totals in health care consumption, expenditure as well as “total societal costs.”