OUP user menu

Peri-ablation warfarin vs. dabigatran: different but the same

Roger A. Winkle, Melissa H. Kong
DOI: http://dx.doi.org/10.1093/europace/eut295 1385-1386 First published online: 26 September 2013

This editorial refers to ‘Safety and efficacy of interrupted dabigatran for peri-procedural anticoagulation in catheter ablation of atrial fibrillation: a systematic review and meta-analysis’ by A.A. Bin Abdulhak et al., on page 1412 and ‘Safety and efficacy of dabigatran etexilate during catheter ablation of atrial fibrillation: a meta-analysis of the literature’ by S.H. Hohnloser and A.J. Camm, on page 1407.

Warfarin has been the traditional oral anticoagulant used to prevent thrombo-embolic events in high-risk patients with atrial fibrillation (AF). The introduction of novel oral anticoagulants (NOACs) for stroke prevention in patients with non-valvular AF represents a major advance in the treatment of these arrhythmias. The NOACs provide an efficacy and safety profile equal to or better than warfarin.13 For the patient, they provide many advantages, including a rapid onset and offset of action and freedom from frequent blood testing to monitor the anticoagulant effect.

Once these NOACs were approved by regulatory agencies, electrophysiologists began using them for thrombo-embolic prophylaxis in patients with non-valvular AF. This prompted the dilemma of how to manage oral anticoagulation in the peri-procedural AF ablation period, during which patients have an increased risk for thrombo-embolic events beyond each patient's intrinsic risk because of prolonged left heart catheterization, the risk of sheath thrombi, and the potential for thrombo-embolic events related to energy application during tissue ablation. Our group was an early adopter of utilizing dabigatran in the peri-ablation period and reported a small observational study suggesting that briefly interrupted dabigatran appeared safe and efficacious.4 We chose to interrupt dabigatran 36 h pre-ablation in patients with normal renal function because of the lack of an antidote should major bleeding occur during the procedure. Patients were given one dose of enoxaparin 0.5 mg/kg immediately at the end of the procedure, again 12 h later and the dabigatran was resumed the morning after the procedure. We chose to cover patients with enoxaparin in the early hours following the procedure because of reports in the orthopaedic literature that dabigatran was not well absorbed following surgery.5 Shortly after the publication of our article, Lakkireddy et al.6 reported a comparison of dabigatran with uninterrupted warfarin. The dabigatran was held the morning of the procedure and resumed shortly post-procedure. This publication came to the opposite conclusion of our study and suggested that dabigatran resulted in significantly more thrombo-embolic events and more bleeding than uninterrupted warfarin. In that study, the increased risk of bleeding seemed largely attributable to the near lack of interruption of dabigatran; however, the increased rate of thrombo-embolic events is more difficult to explain. The disparate conclusions of these two early reports on the use of dabigatran in the peri-ablation period sparked the publication of a number of other studies comparing warfarin with dabigatran in the peri-ablation period, several of which directly commented upon the different conclusions of these early studies as justification for publishing more data on this subject.

This issue of the Journal contains two excellent meta-analyses of all currently available studies in the literature comparing warfarin with dabigatran in the peri-ablation period.7,8 The included studies utilized both uninterrupted and interrupted warfarin therapy and all the dabigatran studies used a brief period of dabigatran interruption ranging from holding the dabigatran the morning of the procedure to 24–30 h pre-procedure. In general, dabigatran was resumed shortly after the procedure once haemostasis had been achieved. Both meta-analyses found that compared with warfarin, dabigatran had a similar rate of thrombo-embolic events and bleeding complications. These meta-analyses add clarity to the use of dabigatran in the peri-ablation period. It would appear that a very brief period of interruption of dabigatran prior to the ablation procedure, with resumption as soon as haemostasis has been achieved, is safe. There is neither need to switch a patient on long-term dabigatran therapy prior to an AF ablation back to warfarin just to undergo the procedure nor does there appear to be a need for enoxaparin bridging.

While answering the issues about the safety and efficacy of dabigatran in the peri-AF ablation period, these meta-analyses still leave other questions unanswered. They provide little or no guidance as to how to use dabigatran in patients with moderate renal dysfunction, as most ablations are undertaken in patients with normal renal function. Warfarin may be a safer alternative in these patients. These meta-analyses provide no information as to how many patients on warfarin or dabigatran had left atrial appendage clot or sludge found pre-ablation and thus had their procedures cancelled. We still perform a transoesophageal echocardiogram under general anaesthesia prior to AF ablation in all patients to rule out left atrial appendage thrombus. The current literature also provides no information to guide us on ‘mixing or matching’ one oral anticoagulant pre-ablation and another post-ablation. In the real world, there are many reasons why a patient may be on one oral anticoagulant pre-ablation and another post-ablation. Using interrupted warfarin, many of our patients on warfarin pre-ablation elect to transition to one of the NOACs immediately post-ablation. Similarly, patients on dabigatran pre-ablation, who have suffered with low-grade gastrointestinal symptoms, often choose to transition to rivaroxaban or apixaban post-ablation. Other patients preferred the once a day dosing of rivaroxaban to the twice daily dosing of dabigatran or apixaban and chose to switch drugs at the time of their ablation. In the USA, there are insurance coverage issues which may make it financially advantageous for a patient to change from one NOAC to another at the time of ablation.

While it is reassuring that dabigatran can be safely used in the peri-procedural AF ablation period, we must not extrapolate these findings to the other NOACs. It will be important for the AF ablationists to report their experiences with rivaroxaban and apixaban in the scientific literature, just as they have done with dabigatran. In the ROCKET-AF (Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation) trial, when rivaroxaban was transitioned to warfarin, there was an increased risk of thrombo-embolic events.9 While this is most probably due to lack of anticoagulant coverage rather than a prothrombotic state, this will need to be addressed if rivaroxaban is utilized in an interrupted manner for an AF ablation. Once successful antidotes are developed for the NOACs, then shorter periods of interruption, or possibly uninterrupted NOAC therapy may be feasible.

Both of these meta-analyses provide the usual, politically correct if not perfunctory, statement that what is really needed is a randomized controlled trial. It seems unlikely, however, that such a trial will ever be conducted for dabigatran and in reality, it may not be necessary. The data from these meta-analyses make it clear that whatever differences in safety or efficacy may exist between dabigatran and warfarin, the differences are extremely small. If one assumes the worst-case scenario that in the meta-analysis by Hohnloser and Camm,7 the trend towards fewer thrombo-embolic events on warfarin was a real difference and the meta-analysis was unable to detect the difference because of too few subjects, it would require a randomized trial sample size of 5488 patients for this difference to be statistically significant. It is unlikely that the pharmaceutical company associated with dabigatran would be willing to spend the money required for such a large study with a goal of showing warfarin was superior to dabigatran, especially when most patients with a successful AF ablation will discontinue dabigatran a few months post-ablation. The best we could hope for would be a smaller equivalency trial. Although the scientific community should encourage pharmaceutical companies to evaluate the NOACs in all clinical settings in which they will be used, the data in these meta-analyses may in some ways be better than data from a randomized trial. Randomized trials frequently exclude many patients who are included in the real-world experience studies combined into the meta-analyses. Once a randomized trial result has been reported, clinicians must still figure out how to treat all the patients who were excluded from participation in the randomized trial.

In the absence of a large randomized trial, the scientific process appears to have worked well in the case of dabigatran in the peri-AF ablation period. A large number of individual investigators from around the world kept careful records of their patients on this new anticoagulant—at times even performing small randomized trials —and reporting these findings in peer-reviewed medical journals. Hohnloser and Camm7 and Abdulhak et al.8 combined all these smaller studies into two meta-analyses to give us an insight into the safety and efficacy of dabigatran compared with warfarin in this setting. Hopefully, the same process will occur with rivaroxaban and apixaban.

Conflict of interest: none declared.


  • The opinions expressed in this article are not necessarily those of the Editors of Europace or of the European Society of Cardiology.


View Abstract