Should we rethink A Fib cardioversion?

It’s well accepted that the window for acute atrial fibrillation cardioversion of atrial fibrillation ends at  48 hours post onset. We did a whole episode on that very point.  The 48 hour window is now being challenged by the biggest study to date looking at this topic.


Time to Cardioversion for Acute Atrial Fibrillation and Thromboembolic Complications was published as a letter to JAMA on August 13, 2014. Ryan Radecki sent the first FOAMed shot across the bow with this review.  You can stop now and check out Ryan’s review; he’s far more erudite than I. If you need more info, read on…

Study in a nugget: This was a retrospective study from Finland that looked at around 2500 patients with a primary diagnosis of atrial fibrillation (AF), aged 18 years or older, with successful cardioversion in the emergency department within the first 48 hours of AF onset.  The primary outcome, a thromboembolic event, was defined as a clinical stroke or systemic embolism confirmed by computerized tomography or magnetic resonance imaging, surgery, or autopsy. Time to cardioversion was determined as the difference between the beginning of arrhythmic symptoms to the exact time of cardioversion. There were 3 groups: less than 12 hours, 12 hours to less than 24 hours, and 24 hours to less than 48 hours.

Thromboembolism Rates


Under 12 hours: 0.3%

24 to 48 hours: 1.1%.

It seems like 12 hours is the inflection point when risk went up and a CHADS VASC score of greater than 1 increased risk.

I’m not sure where this leaves us, maybe risk stratification in ED cardioversion? This was observational, retrospective, and did not include post cardioversion anticoagulation as an intervention. There is no definitive answer or management change from this letter. It does raise the question of whether we should anticoagulate cardioverted AF patients with over 12 hours of symptoms, or those with a CHADS VASC over 1. However, there is no evidence that a post cardioversion anticoagulation strategy would decrease thromboembolic event rate. Also, the incidence of post cardioversion thromboembolic events in this letter is far higher than reported in other literature.

Rob’s practice changers

Cunningham Technique

Loop Abscess Drain Technique

Delayed Sequence Oxygenation

Swami’s  practice changers

Nasal Cannula Apneic Oxygenation

Tranexamic Acid for Mucosal Bleeds


Nuotio, Ilpo, et al. “Time to cardioversion for acute atrial fibrillation and thromboembolic complications.” JAMA 312.6 (2014): 647-649.

CABO CME Retreat December 5-9, 2014

Primary Care RAP


  1. clay josephy

    This comes as a bit of a surprise to hear…I guess maybe Im the oddball, but I was taught that regardless of the duration of AF, stroke risk is managed independent of rhythm management strategy and is based on risk factors. Those with risk factors get anti-coagulated regardless of duration of AF, and the atrial stunning phenomenon with cardioversion confers the same risk of thromboembolism and all that neat stuff. Basically stroke prevention is something we need to do and think about separately and irrespective of our management of the rhythm or duration of rhythm (conversion, rate control, etc). In other words, I hear you guys saying that if the AF is of less than 12 hours duration you do not anti coagulate regardless of risk factors??? That sounds strange to me, but y’all are infinitely smarter than me.

    AHA, ESC and Chest Physicians all recommend OAC in patients with risk factors for thromboembolism regardless of duration of AF whether or not you proceed to cardioversion (and chemical vs electrical has no bearing either). If I have someone with new onset AF, and a CHADS >1, I convert them and anti coagulate them assuming no contraindications. For the record I use mostly NOACs these days.

    Wann, L Samuel et al. 2011 ACCF/AHA/HRS focused update on the management of patients with AF. Circulation. 2011.

    Camm AJ et al. ESC Guidelines for the management of AF. Eur Heart J. 2010.

    You JJ et al. Antithrombotic therapy for AF. American College of Chest Physicians Giudelines. Chest. 2012

    1. Rob Orman

      Hi Clay,
      You bring up a great point, and one that causes an inordinate amount of debate and confusion! The recommendations you site are, to their own admission, based on a paucity to absence of data. The preponderance of data to date point to the opposite conclusion, namely that patients with paroxysmal a fib of less than 48 hours duration are at an extremely risk for post cardioversion thromboembolic events. There is no evidence that acutely anti coagulating this cohort is beneficial. We did a show on this very issue a few years back…

      1. clay josephy

        What a great conversation. So interesting that these topics (STEMI, SAH, AF, Sepsis etc) that we love and think about so much just have infinite corners of intricacies that are so fascinating. The things we think about the most are the most apt to confuse us. Especially when there is a gazillion papers and as you point out, consensus based guidelines to incorporate, and a busy ED to wade through.

        Thats why i love your style so much, your always looking from the perspective of docs in the trenches trying to be intellectually honest, evidence driven and rigorous, but also practical and patient oriented. Thank you for what you do. I am just an ordinary ED doc slugging it out in the trenches, so it resonates.

        I think of it as two separate questions:

        1) Is there a risk of an thromboembolic event in the peri-procedural period attributable to conversion?

        I think the answer is yes, kind of, maybe..Super low, but apparently, like the yearly stroke risk, hinges primarily on the patients intrinsic vascular risk factors. Lets say in general its 2 cohort to acute anticoagulation versus none.

        Practically speaking, the majority of my PAF, lone AF, RAFF cases are CHADS <1 I give them ASA or more commonly, nothing, but occasionally you have someone who is CHADS 2 or greater, and I think those patients are the ones who need to be anti coagulated in the ED during the procedure or sometime during the visit (I just give NOAC orally) and then I send them home on NOAC and usually tell them at least a month (duration of the longest follow up in most studies pertaining to ED conversion) and that they need to talk it over with their cardiology as to how long they stay on it.

        You are treating both a post/peri procedural risk of thrombotic event(minimal), as well as their annual risk conferred to them by nature of having AF on top of their intrinsic risk factors (maybe not so minimal).

        The real question, as you allude to above, is whether anti coagulating those patients would avoid the short term event at all, and none of the patients in this new study received anticoagulation.

        So, in other words, this study doesn't surprise me at all. ED cardioversion is generally super safe, short term thrombotic events are very uncommon in all comers, the sooner you convert the better (less time for clot to form i guess, if thats even the real cause for embolic events) and the longer you wait the more events you will see, and most importantly, predictors of a thrombotic event include CHADS VASC criteria. This last point should make one think about anti coagulating the patient, at least for the short term.

        So for me, if the CHADS VASC is greater than 2 than I give them a dose of NOAC in the ED, convert them, send them home on a NOAC. Admittedly this case is much less common than the ASA or nothing group CHADS VASC <2 so this conversation doesn't come up that frequently in real life.

        I wonder if this strategy might mitigate the events we are seeing in this study.

        For me this study doesn't really change my practice of converting people and choosing candidacy for ED conversion, but I think I will continue doing what Im doing now and anti coagulating those with risk factors…

        Great episode and discussion.

        thanks, Im sure Ill get torn to bits now…

        1. clay josephy

          that post got kind of messed up at submission:

          meant to say

          “I think the answer is yes, kind of, maybe..Super low, but apparently, like the yearly stroke risk, hinges primarily on the patients intrinsic vascular risk factors. Lets say in general its less than 1%.”

          “The real question, as you allude to above, is whether anti coagulating those patients would avoid the short term event at all, and none of the patients in this new study received anticoagulation. Someone needs to randomize the CHADS>2 cohort to acute anticoagulation versus none.”

  2. clay josephy

    the second question was;

    2) does this patient have an intrinsic risk of stroke and how is that mitigated now that they have AF?

    man, my post got all sorts of cut up and changed when i posted it…maybe i screwed something up..apologies

Awesome article, I know - please share your erudite thoughts...