JC: Should we rapidly cardiovert AF in the ED? St Emlyn’s

If I develop AF then I reckon I’d be able to spot it pretty quick, and I’d get myself down to ED pronto so that I could get myself cardioverted having read the excellent work of Stiell et al​1​. Why? Well I quite like to do cardioversions and so it would be nice to give someone else the opportunity, but more than that, it’s because I think it’s a good idea. But is it?

My belief is that the risks of cardioversion are low, and that the risks of complications are higher if we wait to get it done. In other words my ‘belief’ is that earlier is better, but in truth that may not be the case. The data that’s out there suggests that cardioversion is low risk up until 48 hours​2​ and so what’s the rush? Perhaps it’s because of this thought…..

There are two indications for any procedure. Either the patient needs it, or you want to do it. Only one of these is valid and honourable.

St Emlyn (with HT to John Hinds)

We might just want to stop and think about how urgent new onset AF really is. Maybe we can wait and see if the patient will spontaneously convert back to sinus rhythm to avoid sedation, electricity or drugs? If so, then how long should andcan we wait?

Thankfully, there is a paper that might help us in this dilemma (ably reviewed on a newly found EM blog called the Breach​3​ which is well worth a look​4​). An RCT of early vs delayed cardioversion.

The abstract is below, but as we always say please read the evidence for yourself​3​ and come to your own conclusions

What kind of paper is this?

This is a randomised controlled trial which is an appropriate design for an intervention trial. Specifically, it is described as a non-inferiority trial which seeks to demonstrate that the difference between two treatments is no more different than a prespecified amount (in this case a 10% difference). The 10% difference is arguably a rather large one. In such a common condition we should perhaps be looking for a rather more robust analysis of equivalence. Patients were randomised in a 1:1 ratio to either a wait-and-see policy or for chemical cardioversion.

Who was studied?

Patients with new onset AF. Aged >18 and who were cardiovascularly stable. They had to be within a 48 hour window of onset of symptoms, so pretty much the group that we could either watch and wait or proceed to cardioversion in the ED.

This paper enrolled 327 patients in 15 hospitals. Specifically this was in the cardiology departments but that might reflect how things are organised in the Netherlands. In the ED most of these patients would be in the ED or under the care of the acute medical teams (our cardiologists in Virchester are selective in what type of heart problems they take).

What did they do?

In the wait-and-see group, patients were administered a variety of rate-control medications, so not really a true wait-and-see approach.

The cardioversion group were preferentially treated with pharmacological cardioversion (flecainide where possible) or with electrical cardioversion if that was not possible.

So, this was not really a trial of nothing vs cardioversion, but rather a rate-control approach vs. cardioversion.

Primary outcome

The primary outcome was whether the patient was in AF at 4 weeks post intervention. To me this is not really the outcome measure that matters acutely. I’m more interested in whether the patient gets out of AF in the early phase of the disease.

Main results

In terms of the 4 week outcome there was no real difference between the groups. 91% were in sinus rhythm in the delayed group vs. 94% in the cardioversion group.

What is interesting is the number of people who spontaneously cardioverted in the watch-and-wait group. 69% of patients got themselves back into sinus rhythm with just rate control medications. That’s a lot of patients who essentially sorted themselves out.

So what does this mean for practice?

When I first read the title and some of the buzz around it, I did think that this paper would change my practice, but now I’m not so sure. In Virchester, when a patient with new onset AF arrives we don’t usually go straight for electrical cardioversion. If there are no contraindications we start pharmacological interventions first whilst making arrangements for DC cardioversion to take place within the 48 hours from onset window. If the patient cardioverts within that time then of course we cancel the DC cardioversion.

Does this new evidence mean that we should not give pharmacological agents in this group? I don’t think it does, but it certainly takes any pressure off us to DC cardiovert as soon as the patient arrives. It also means that in those patients who are unsuitable for pharmacological intervention (flecainide is contraindicated in many patients), that a delayed approach to DC cardioversion is appropriate, for example by planning to do this after a period of observation rather than straight away.

Just be careful with timings though – remember that you do want to cardiovert your patient within 48 hours if possible. So, if your patient presents at 40 hours, and it’s nearly midnight, then you need to be super slick to make sure that they have the opportunity for cardioversion within the remaining 8 hours. System issues may mean that on balance a DC cardioversion right now might be the best option for that patient.

This will disappoint some clinicians. DC cardioversions in the ED are fun (for us, but not the patient) and in procedure-based systems there may be resistance to not ‘doing stuff’.

What this study cannot tell us (as it was not designed to and is not powered enough to) is whether there are advantages to early cardioversion in terms of complications such as thrombosis, sedation issues and others. Is there an argument that the earlier you cardiovert the better it is? My reading of the data out there is that this has not been shown within the 48 hour window, but please do get in touch and post in the comments section if you’ve seen differently. The authors clearly thought the same as they excluded those presenting at >36 hours from symptom onset.

What I don’t get from this paper is any information that might guide me in predicting who is likely to spontaneously convert. Data from 1998 suggests that early presentation is a predictor of conversion​5​, but that does not really help me (Ed- interestingly that paper also found that roughly two thirds of patients spontaneously converted as they did in this trial). For example, my approach to a 65 year old overweight, hypertensive, smoker might be very different from that of a 22 year old who ends up in AF after a weekend of alcohol and drug excess.

Also remember that you must manage the stroke risk in these patients post cardioversion. In this study they used the EMERG-AF​6​ approach but still had 2 patients suffer strokes. You may use something different, but you should be using something. As the accompanying editorial suggests​7​, it may be a sprint race to get the patient back into sinus rhythm, but it’s often the start of a long term treatment marathon for the patient.

This is a pretty well-conducted study that certainly adds to the literature on AF management so well done to the authors on putting this together.

The bottom line

In patients who present to the ED within 48 hours there is probably no panic to cardiovert the patient. It’s fine to delay DC cardioversion to try a period of either rate control, or (as we will continue to do) an attempt to pharmacologically cardiovert them.

If you want to go straight for a DC cardioversion then that’s probably also fine, but just make sure you balance the risks of the procedure against time, space and convenience.




  1. 1.
    Stiell I, Clement C, Rowe B, et al. Outcomes for Emergency Department Patients With Recent-Onset Atrial Fibrillation and Flutter Treated in Canadian Hospitals. Ann Emerg Med. 2017;69(5):562-571.e2. https://www.ncbi.nlm.nih.gov/pubmed/28110987.
  2. 2.
    Weigner M, Caulfield T, Danias P, Silverman D, Manning W. Risk for clinical thromboembolism associated with conversion to sinus rhythm in patients with atrial fibrillation lasting less than 48 hours. Ann Intern Med. 1997;126(8):615-620. https://www.ncbi.nlm.nih.gov/pubmed/9103128.
  3. 3.
    Pluymaekers NAHA, Dudink EAMP, Luermans JGLM, et al. Early or Delayed Cardioversion in Recent-Onset Atrial Fibrillation. N Engl J Med. April 2019:1499-1508. doi:10.1056/nejmoa1900353
  4. 4.
    Stevenson B. should we cardiovert everyone with recent onset AF. The Breach. https://the-breach.com/should-we-cardiovert-everyone-with-recent-onset-fast-af/. Published 2019. Accessed 2019.
  5. 5.
    Danias P, Caulfield T, Weigner M, Silverman D, Manning W. Likelihood of spontaneous conversion of atrial fibrillation to sinus rhythm. J Am Coll Cardiol. 1998;31(3):588-592. https://www.ncbi.nlm.nih.gov/pubmed/9502640.
  6. 6.
    Coll-Vinent B, Martín A, Sánchez J, et al. Benefits of Emergency Departments’ Contribution to Stroke Prophylaxis in Atrial Fibrillation: The EMERG-AF Study (Emergency Department Stroke Prophylaxis and Guidelines Implementation in Atrial Fibrillation). Stroke. 2017;48(5):1344-1352. https://www.ncbi.nlm.nih.gov/pubmed/28389612.
  7. 7.
    Healey JS, McIntyre WF. The RACE to Treat Atrial Fibrillation in the Emergency Department. N Engl J Med. April 2019:1578-1579. doi:10.1056/nejme1902341

Cite this article as: Simon Carley, "JC: Should we rapidly cardiovert AF in the ED? St Emlyn’s," in St.Emlyn's, May 11, 2019, https://www.stemlynsblog.org/should-we-rapidly-cardiovert-af-in-the-ed-st-emlyns/.

7 thoughts on “JC: Should we rapidly cardiovert AF in the ED? St Emlyn’s”

  1. Simon, thanks for the fantastic review. There’s so much variation in the management that depends on the clinician, the patient and, the health system. This changes nothing for me and I’m a convert in the ED guy (electrical or pharmacologic). My issue, and the one that many American docs will share, is that I can’t get 48 hour follow up accomplished. Many of our cards/primary care colleagues don’t want it because their clinics are already full and there’s nowhere to squeeze in the patients. Additionally, if the patient is still in AF at 48 hours, they’re gonna be sent back to the ED so business at time of first presentation doesn’t really matter too much to me – really what I’m often going to be doing is punting the procedure to another EM colleague at 48 hour follow up.
    The biggest thing to me isn’t just that they’re in AF or not or how long exactly it’s been present but how it’s affected the patient. Many of the patients I see come in within an hour or two of onset because they feel terrible and in those patients, I’ve got no reluctance to cardiovert immediately. For those who don’t have much in the way of symptoms, I’m less convinced that it’s critical to convert (though I’m often still doing it).

  2. Thanks for a great review Simon. We’ve discussed this paper at work as well and actually decided that we are going to have a talk with our cardiology colleagues about it and see what they think and if this should change our current method. Just to clarify, in Denmark this is not an ED/EM procedure but a joined venture between anaesthesia and cardiology mostly carried out in the cardiology ward . What we often see is that while patients wait for us from the anaesthesia dept to find time to sedate for the procedure they tend to convert anyway just as they reported in the paper. I see Swami’s points to not change and are happy that our system allows for a much closer and easier follow up. I think this paper might change our practise at the least for some of our patients and spare them some sedation and cardioversion.

  3. Pingback: AFib: Wait-and-See or Early Cardioversion to Obtain Normal Sinus Rhythm? - REBEL EM - Emergency Medicine Blog

Thanks so much for following. Viva la #FOAMed

Scroll to Top