Date: June 20th, 2019

Reference: Pluymaeker et al. Early or Delayed Cardioversion in Recent-Onset Atrial Fibrillation. NEJM 2019

Guest Skeptic: Dr. David Glaser, emergency physician from a community teaching hospital in Denver and faculty member for the annual Emergency Medicine and Acute Care course series. Dave is also residency-trained and boarded in internal medicine.

Case: A 62-year-old woman with a history of hypertension presents with four hours of palpitations, described as a racing heart. Her vital signs: BP 148/90, HR 135, RR 16, T 37oC, O2 Sat 96%. Lungs are clear and cardiac exam shows a tachycardic and irregularly irregular rhythm without murmur. The ECG shows atrial fibrillation without ischemic changes. How do you proceed?

Background: New-onset atrial fibrillation is a common occurrence in the emergency department, and practitioners differ on whether to take a primary rate-control approach versus a rhythm-control approach utilizing either electrical or pharmacological cardioversion. We are not going to settle this debate on this podcast.

In the United States especially, these patients are often admitted to the hospital with rate control and cardiology decides on cardioversion.

Dr. Ian Stiell

In Canada, these patients are often cardioverted and discharged home. We covered the Ottawa Aggressive Protocol on SGEM#88. That episode reviewed a 2010 cohort study done by the Legend of Emergency Medicine, Dr. Ian Stiell. The results from this observational study was 92% of patients were electrically cardioverted, 97% discharged home with 93% in sinus rhythm at discharge.

Things are starting to change in the US. A study published in AEM showed that implementing a new atrial fibrillation algorithm decreased hospital admissions from 80% to 67% and cardioversion increased from 17% to 21% (SGEM#222).

Clearly there is a difference in the management of patients with recent onset of rapid atrial fibrillation depending on your practice location.

Clinical Question: In adult patients who present with hemodynamically stable, symptomatic, recent-onset atrial fibrillation without signs of myocardial ischemia, is a wait-and-see approach, inferior to an immediate cardioversion strategy.

Reference: Pluymaeker et al. Early or Delayed Cardioversion in Recent-Onset Atrial Fibrillation. NEJM 2019

  • Population: Adults(18 years and older) who presented to the emergency department of 15 hospitals in the Netherlands (3 academic, 8 non-academic teaching, and 4 non-teaching hospitals) with hemodynamically stable, symptomatic, recent-onset (< 36 hours), first-detected or recurrent atrial fibrillation, without signs of myocardial ischemia or a history of persistent atrial fibrillation (defined as lasting > 48 hours).
    • Exclusions: Signs of myocardial infarction on ECG, hemodynamically unstable, presence of pre-excitation syndrome, history of sick sinus syndrome, history of unexplained syncope, history of persistent AF (episode of AF lasting more than 48 hours), acute heart failure or deemed unsuitable for participation by attending physician.
  • Intervention: Delayed cardioversion (“wait-and-see” approach). This was defined as administration of a rate-control medication, including intravenous or oral beta-blockers, nondihydropyridine calcium-channel blockers, or digoxin, given in increasing doses to obtain relief of symptoms and a HR of 110 BPM or less. Patients were discharged when their condition was judged to be clinically stable. A cardiology out-patient clinic visit was planned for the next day, as close as possible to 48 hours after the onset of symptoms. If atrial fibrillation was still present at this visit, patients were referred back to the emergency department for delayed cardioversion.
  • Comparison: Early cardioversion. This was performed at the initial emergency department visit, either pharmacologically (and preferentially with flecainide) or electrically in those with contraindications to pharmacologic cardioversion and in patients with previous or current unsuccessful pharmacologic cardioversion. Patients were discharged when their condition was determined to be clinically stable.
    • Patients with a high risk of stroke not already on anticoagulation had anticoagulation initiated before or immediately after cardioversion. Long-term anticoagulation was continued in accordance with current guidelines based on the patient’s CHA2DS2-VASc score.
  • Outcome:
    • Primary Outcome: The presence of sinus rhythm on ECG recorded at the four week cardiology out-patient visit.
    • Secondary Outcomes: There were some differences between those in the publication and those listed on (NCT02248753). Specifically, they did not mention duration of the index emergency department visit (which included a next-day visit as needed in the wait-and-see patients) in the trial registry, but it was mentioned the supplemental index. Here are the ones listed on the trial registry website:
      • Time to conversion to sinus rhythm (Holter monitor) in the intervention group only.
      • Quality of life (SF-36) measured at baseline, 4 weeks, 6 months, and 12 months
      • One-year follow-up of Major Adverse Cerebrovascular or Cardiovascular Events
      • Time to first recurrence of Atrial Fibrillation within 1 month
      • Total health care and societal costs within 1 year
      • Quality of Life (AFEQT) at baseline, 4 weeks, 6 months, and 12 months

Authors’ Conclusions: “In patients presenting to the emergency department with recent-onset, symptomatic atrial fibrillation, a wait-and-see approach was noninferior to early cardioversion in achieving a return to sinus rhythm at 4 weeks.

Quality Checklist for Randomized Clinical Trials:

  • The study population included or focused on those in the emergency department. Yes
  • The patients were adequately randomized. Yes
  • The randomization process was concealed. Yes
  • The patients were analyzed in the groups to which they were randomized. Yes
  • The study patients were recruited consecutively (i.e. no selection bias). Unsure
  • The patients in both groups were similar with respect to prognostic factors. Unsure
  • All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  • All groups were treated equally except for the intervention. No
  • Follow-up was complete (i.e. at least 80% for both groups). Yes
  • All patient-important outcomes were considered. Yes
  • The treatment effect was large enough and precise enough to be clinically significant. No

Key Results: There were 437 patients included in the study. The mean age was 65 years and 40% were female.

Sinus rhythm 91% in delayed vs. 94% in early cardioversion

  • Primary Outcome: The presence of sinus rhythm at four weeks occurred in 193 of 212 patients (91%) in the delayed-cardioversion group and in 202 of 215 (94%) in the early-cardioversion group (between-group difference, −2.9% (95% CI −8.2 to 2.2; P=0.005 for non inferiority
  • Secondary Outcomes: 
    • Time to conversion to sinus rhythm (Holter monitor) in the intervention group only: In the wait-and-see cardioversion group, spontaneous conversion to sinus rhythm occurred in 150 of 218 patients (69%) within 48 hours of symptom onset.
    • Quality of life (SF-36) measured at baseline, 4 weeks, 6 months, and 12 months: No data provided in document or supplemental material. However, they do provide AFEQT data at 4 weeks (see below).
    • One-year follow-up of Major Adverse Cerebrovascular or Cardiovascular Events. Cardiovascular complications were reported within four weeks after randomization (including during the index visit), 10 cardiovascular complications occurred in the delayed-cardioversion group (including one patient with ischemic stroke and three with acute coronary syndrome or unstable angina) and eight in the early-cardioversion group (including one patient with transient ischemic attack and three with acute coronary syndrome or unstable angina). There were no deaths during follow-up. One-year data is to be reported after one-year follow-up is completed. Both patients with a cerebral ischemic event were on anticoagulation at the time (one occurred 5d after spontaneous cardioversion having had dabigatran started on the index visit, and the other 10d after early electrical cardioversion, having started rivaroxaban on the index visit.)
    • Time to first recurrence of Atrial Fibrillation within 1 month: Median time was 12 days (range, 3 to 18) in the delayed group and 8 days (range, 2 to 18) in the early group.
    • Total health care and societal costs within 1 year: To be reported later.
    • Quality of Life (AFEQT) at baseline, 4 weeks, 6 months, and 12 months: The mean AFEQT global scores were 72±19 in the delayed-cardioversion group and 73±19 in the early-cardioversion group (difference, −1 point; 95% CI, −5.3 to 4.0). The other data was not reported and it is unclear if it will be reported based on the supplemental information.
    • Total time at the emergency department: The total median duration of the index visit (including delayed cardioversion if necessary) was 120 minutes (range, 60 to 253) in the delayed-cardioversion group and 158 minutes (range, 110 to 228) in the early-cardioversion group.The Hodges–Lehmann estimate for the difference in medians between the two groups was 30 minutes (95% CI, 6 to 51).

1) Selection Bias: They don’t explicitly state it in the manuscript that consecutive patients were recruited. Their published methods said; “all eligible patients”. However, two sites had 1,125 eligible patients with 954 not enrolled (366 declined, 361 had administrative reasons and 227 had spontaneous conversion). The other 13 sites had 266 patients added with no systematic screening process. Physicians could also exclude anyone they “deemed unsuitable for participation”. This call into question whether or not these were truly consecutive patients and could have introduced selection bias.

2) Unbalanced groups: The delayed cardioversion group had nearly twice as many patients with a previous MI than the early cardioversion group (24% vs. 13%). It is unsure if this would have an impact on the results.

3) Power Calculation:It seemed somewhat arbitrary having the difference being set at 10%. Is this what patients or clinicians would consider non-inferior?

“A difference of 10% is considered acceptable, given the natural variation in presence or durability of sinus rhythm, the generally low impact of the absence of sinus rhythm on the wellbeing of the patient and the availability of good treatment options should treatment be necessary.”

4) Secondary Outcomes: There seemed to be some changes in the secondary outcomes. There were some differences than those listed on (NCT02248753). Specifically, the originally published study design posted to the trial registry does not mention duration of index emergency visit duration but the supplemental index does mention this as a pre-specified secondary outcome.

Nearly everyone was discharged home following their initial emergency department visit (only 3 of 218 in the delayed-cardioversion group and 5 of 219 in the early cardioversion group were admitted).

Of the 335 patients who underwent telemetry monitoring, ~ 30 % had a documented recurrence of atrial fibrillation within 4 weeks of the index visit

The other thing to recognize is that this secondary outcome, the prolonged LOS in the ED observed in the immediate cardioversion group might have been due either to their preference to perform pharmacologic cardioversion or to their local practice of having cardioversion performed by cardiology and anesthesiology rather than the emergency physician.

5) Intention-To-Treat (ITT) Analysis: They performed an ITT analysis in this study. Normally we look for the authors to perform an ITT analysis as a quality indicator. That is because we are often reviewing a superiority trial.

The ITT principle is to include all randomized patients irrespective of post-randomization occurrences. This will tend to bias the results towards having no effect (accepting the null hypothesis) and is a more conservative approach. Using a per-protocol (PP) analysis can increase the effect size and bias the results to rejecting the null hypothesis.

Things are flipped around in a non-inferiority trial. Because ITT analyses bias towards the null this would more likely result in falsely accepting the null hypothesis and supporting the conclusion of non-inferiority. It is a quality indicator in non-inferiority studies to do a PP analysis to mitigate this potential bias.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We generally agree with the authors’ conclusions.

SGEM Bottom Line: Both delayed and early cardioversion of acute onset atrial fibrillation achieve high rates of sinus rhythm in their patients at the time of a 4-week follow up.

Clinical Application: New-onset atrial fibrillation is a common occurrence. Patients can be managed either with immediate cardioversion or with rate-control and a recheck the next day and cardioversion at that time if they are in the minority who remain in atrial fibrillation. Immediate cardioversion obviates the need for a visit the next day, which may be unavailable in many systems and inconvenient for patients. A wait-and-see approach, however, obviates the need for many cardioversions.

Dr. David Glaser

What Do I Tell the Patient? Because both approaches are reasonable you do some shared decision making. You tell your patient that returning them to normal heart rhythm is appropriate and could be achieved either with immediate cardioversion or during a next-day visit—if one can be arranged—and the latter approach would also give them the chance to spontaneously convert to a normal rhythm, which likely happens about two-thirds of the time. If they choose this approach, they will need to stay for a time to get their symptoms and heart rate controlled. Either way, for patients at high risk of stroke based on a CHA2DS2-VASc score >2, you begin anticoagulation. You tell your patient that regardless of the approach chosen, there’s no need to admit them to the hospital.

Case Resolution: The patient chooses immediate cardioversion, which is successfully performed using electrical cardioversion. Given her CHA2DS2-VASc score of 2, you begin a DOAC and discharge her from the ED 1.5 hours after arrival.

Keener Kontest: Last weeks’ winner was Adam Miller a critical care paramedic and firefighter from Grand Rapids, Michigan. He knew John McLaughlin, a Canadian chemist and pharmacist, created dry ginger ale that became known as Canada Dry.

Other FOAMed:

  • EM Nerd: The Case of the Irregular Irregularity Continues
  • REBEL EM: Wait-and-See or Early Cardioversion to Obtain Normal Sinus Rhythm?
  • St. Emlyn’s: Should we Rapidly Cardiovert AF in the ED?
  • The Breach: Should We Cardiovert Everyone with Recent-Onset Fast AF?

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.