Author + information
- Ivan Cakulev, MD* ( and )
- Albert L. Waldo, MD
- ↵*Reprint requests and correspondence:
Dr. Ivan Cakulev, Division of Cardiovascular Medicine, University Hospitals Case Medical Center, 11100 Euclid Avenue, MS LKS 5038, Cleveland, Ohio 44106-5038
Prevention of thromboembolism is a principal aim of atrial fibrillation management. Although the mechanisms underlying thrombogenesis in atrial fibrillation are clearly complex and remain only partly understood, it is intuitive that restoration and reliable maintenance of sinus rhythm is probably the best preventive strategy against thromboembolism. Over the past decade, catheter ablation has emerged as a potential cure for atrial fibrillation. Many centers worldwide have been reporting relatively high success rates with few associated complications after ablation of atrial fibrillation in selected patient populations. Particularly with increasing use of this therapy has come an obvious need finally to answer one of the most important questions related to this therapy, namely does long-term maintenance of sinus rhythm after successful catheter ablation eliminate stroke risk in patients with atrial fibrillation, thereby permitting discontinuation of oral anticoagulation therapy (OAT)? Unfortunately, few data exist to guide us in this area. The management of OAT in patients who have undergone atrial fibrillation ablation has largely been left to the individual judgment of the treating physician. However, several practice patterns have emerged based on the apparent presence or absence of atrial fibrillation, duration of recurrent episodes, and stroke risk stratification, the CHADS2(congestive heart failure, hypertension, age >75 years, diabetes mellitus, and prior stroke or transient ischemic attack) scheme being the most commonly used (1,2).
In this issue of the Journal, Themistoclakis et al. (3) present the largest and the only multicenter experience to date addressing this issue. The records of 3,355 patients from 5 well-known atrial fibrillation ablation centers were studied. In 2,692 patients, OAT was discontinued 3 to 6 months after ablation, although these patients were continued on aspirin (Off-OAT group). The decision was made on an individual-case basis according to the local institutional policy regardless of the CHADS2score, but we note that 347 patients had a CHADS2score ≥2. As a general rule, warfarin was discontinued if the patient did not experience any recurrence of atrial tachyarrhythmia, did not have severe pulmonary vein stenosis, and did not have severe left atrial mechanical dysfunction. The rest of the patients (n = 663) remained on oral anticoagulation treatment after the 3- to 6-month period post-ablation (On-OAT group). After an average follow-up of 28 ± 13 months, only 2 (0.07%) patients had experienced an ischemic stroke in the Off-OAT group, and no patient with a CHADS2score ≥2 had experienced a stroke. In the On-OAT group, after an average follow-up period of 24 ± 15 months, only 3 (0.45%) patients had experienced an ischemic stroke. One of the 3 patients experienced this event after a cardioversion. The other 2 patients had unsuccessful ablations, with 1 of them having a subtherapeutic international normalized ratio during the neurological event. On the basis of these results, the authors concluded that “ … it seems that the risk–benefit ratio favors the discontinuation of OAT after successful AF ablation even in patients at moderate-high risk of TE based on CHADS2score alone” (3). They also concluded that “… this conclusion needs to be confirmed by future large randomized trials” (3). After a careful read of the paper by Themistoclakis et al. (3), we strongly support the latter conclusion, but believe equally strongly that it is premature to accept the former conclusion.
What should we do about OAT in patients who seemingly have had a successful cure of their atrial fibrillation? The desire, of course, is to be able safely to terminate OAT in patients who, because of the presence of stroke risk factors, especially those with a CHADS2score ≥2, would otherwise be well advised to continue OAT (4–6). If there is no longer any risk in these patients because they are cured of atrial fibrillation, clearly the expectation is that there should be no need for OAT. However, before that expectation can be realized, there are several important questions that need to be addressed and answered. How can we be sure that a patient is cured? The authors certainly tried to be sure, but there are important concerns regarding their available data. Follow-up was often not face-to-face, involved long intervals between each follow-up visit, and on average was only a little over 2 years. The latter is especially worrisome in patients with stroke risks in light of recent studies that have shown that even after there has been no apparent atrial fibrillation recurrence for 1 year, there is a 5% to 13% recurrence in year 2, and an actuarial recurrence rate of 25% to 46.8% in 5 years and 54.6% at 6 years (7–10). Thus, how can one be sure that in patients at high risk for stroke (CHADS2score ≥2), it is safe to stop OAT? Moreover, there were only 347 of the latter patients in this study, of the 3,355 total patients. In fact, the great majority of patients in this study (82%) would not necessarily have warranted OAT post-ablation of atrial fibrillation because their CHADS2score was ≤1 (4–6).
Additionally, there is the issue of how much atrial fibrillation is a clinically meaningful recurrence, potentially warranting OAT (at least with a CHADS2score ≥2). Studies have shown that there is as much as a 7-fold increase in asymptomatic atrial fibrillation after radiofrequency ablation of atrial fibrillation (11). Perhaps the best illustration of the relevance of such data comes from Martinek et al. (12), who reported on patients who had had radiofrequency ablation of atrial fibrillation, but also happened to have an AT-500 full-disclosure implantable pacemaker device (Medtronic, Inc., Minneapolis, Minnesota) in place. In this small series (n = 14), 4 patients had a symptomatic recurrence (a 71% success rate), but using a 24- to 48-h Holter monitor every 6 months, they found another recurrence, decreasing the success rate to 64%. Then using a 1-week Holter monitor every 6 months, they found yet another case, reducing the success rate to 58%, and then with the continuous monitoring of the AT-500, they found 2 other recurrences, reducing the success rate to 43%. It seems the more rigorous the monitoring, the more atrial fibrillation is uncovered. Then, of course, there is the need to know how much atrial fibrillation recurrence warrants OAT. And all this should be considered in light of the CHADS2score. Furthermore, we know that a recurrence of short duration one time does not indicate that the duration of recurrence will always be short. Thus, as the authors recognized, in the presence of stroke risk factors, how long must an episode or episodes of atrial fibrillation last to impact the likelihood of subsequent stroke, and therefore, to identify the need for prophylactic OAT? They use a duration of 1 min.
Additionally, that there were strikingly few strokes in this study is of considerable interest. It may be, or probably is, explained by one or more of the following: 1) the great majority (82%) of patients had a CHADS2score of ≤1, that is, they had a very low stroke risk to begin with (13), and the guidelines consensus would not have mandated OAT in any event (4–6); 2) the follow-up surveillance was really superb (perhaps especially because patients taking their pulse daily really worked), such that they picked up clinically meaningful atrial fibrillation recurrences quickly and restarted OAT (77 patients were restarted on OAT); 3) in patients with stroke risks, the follow-up was simply not good enough to pick up all of the atrial fibrillation recurrences; and 4) strokes were simply missed.
We have already remarked that the authors observed a very low incidence of thromboembolism in the Off-OAT group. However, also curious is that the incidence of stroke was strikingly low in the On-OAT group. Warfarin reduces but does not eliminate the risk of stroke in patients with atrial fibrillation. Even in the patients with a CHADS2score of 0, the expected annual incidence of stroke should be significantly higher than the overall incidence of 0.23% that the authors reported in their patient population (14–16). However, their reported incidence of major hemorrhagic episodes in the OAT group matches the incidence reported in the literature (14,15). Paradoxically, because of the very low stroke rates observed in this study, it appears that the OAT group, in which 72% of the patients had experienced arrhythmic recurrences, 39% of the patients had CHADS2scores of 1, and 37% had CHADS2scores of ≥2, would have been better served with aspirin rather than with warfarin therapy (6). The authors themselves remind us that “patients with a thromboembolic risk of 2% per year or less do not benefit substantially from OAT, and according to the international guidelines, should not be treated with this therapy” (3). One explanation for such a low incidence of stroke may be that the arrhythmic recurrences in the OAT group were brief in duration and/or clinically not significant. Nevertheless, the surprising and discrepant paucity of thromboembolic events across the groups in this retrospective study seriously limits deriving a meaningful conclusion regarding discontinuation of warfarin therapy post-ablation, especially in patients with higher risk factors for stroke.
Secondly, only 347 (13%) of the patients in the Off-OAT group had a CHADS2score ≥2, and only 10 patients had a CHADS2score of 5 to 6. Current guidelines suggest warfarin therapy in patients with atrial fibrillation and CHADS2scores of ≥2 (4–6). To change this practice on the basis of data from such a small subgroup in a nonrandomized, observational study seems unwise, and is not recommended by the consensus statements (4,6). Again, the need for prospective studies is clear and much needed.
It would also help to understand the mechanism of atrial fibrillation in the same way we understand the mechanism of atrial flutter, atrioventricular re-entrant tachycardia, and atrioventricular nodal re-entry tachycardia. Catheter ablation of these well-understood arrhythmias approaches a 100% acute success rate, and has an insignificant recurrence rate even after a very long follow-up period. Until then, the Heart Rhythm Society/European Heart Rhythm Association/European Cardiac Arrhythmia Society atrial fibrillation ablation consensus statement recommendations should probably be followed (4):
1. Warfarin is recommended for all patients for at least 2 months after an AF ablation procedure.
2. Decisions regarding the use of warfarin more than 2 months after ablation should be based on the patient's risk factors for stroke and not on the presence or type of AF.
3. Discontinuation of warfarin therapy post-ablation is generally not recommended in patients who have a CHADS2score ≥2.
In short, although this is clearly the largest follow-up of post-atrial fibrillation ablation patients and late stroke, it is really only hypothesis generating. These data cry out for a prospective, randomized clinical trial that includes standardized methods of follow-up to assess and characterize recurrence of atrial fibrillation and to determine the incidence/prevalence of stroke.
Therefore, our conclusion: do not stop the warfarin until we have prospective, randomized clinical trials that can help guide us in providing anticoagulation therapy for our patients.
Supported in part by the Jennie Zoline, Blue Dot, and Glenstone Foundations.
Dr. Waldo is a consultant for Boehringer-Ingelheim Pharmaceuticals and Ortho-McNeil-Janssen Pharmaceuticals.
↵* Editorials published in the Journal of the American College of Cardiologyreflect the views of the authors and do not necessarily represent the views of JACCor the American College of Cardiology.
- American College of Cardiology Foundation
- Meiltz A.,
- Zimmermann M.,
- Urban P.,
- Bloch A.
- Themistoclakis S.,
- Corrado A.,
- Marchlinski F.E.,
- et al.
- Calkins H.,
- Brugada J.,
- Packer D.L.,
- et al.
- Fuster V.,
- Ryden L.E.,
- Cannom D.S.,
- et al.
- Pappone C.,
- Rosanio S.,
- Augello G.,
- et al.
- Bertaglia E.,
- Tondo C.,
- De Simone A.,
- et al.
- Hindricks G.,
- Piorkowski C.,
- Tanner H.,
- et al.
- Gage B.F.,
- van Walraven C.,
- Pearce L.,
- et al.
- Halperin J.L.