Anticoagulation After Catheter Ablation of Atrial Fibrillation: An Unnecessary Evil?

A Systematic Review and Meta-analysis

Riccardo Proietti MD, PhD; Ahmed AlTurki MD; Luigi Di Biase MD, PhD; Paolo China MD; Giovanni Forleo MD; Andrea Corrado MD; Elena Marras MD; Andrea Natale MD; Sakis Themistoclakis MD


J Cardiovasc Electrophysiol. 2019;30(4):468-478. 

In This Article


Owing to the high risk of thromboembolism in the early postprocedural period, there is a general consensus that OAC should be continued in the first 2 months after AF catheter ablation, regardless of the CHA2DS2-VASc score.[9] The use of anticoagulation after this period is controversial and there are no recommendations based on controlled trials evaluating how best to reduce CVE or how to minimize the risk of bleeding after successful AF ablation. Current guidelines and expert consensus statements advise long-term continuation of OAC after successful ablation, following the general anticoagulation recommendations based on the thromboembolic risk score, regardless of the rhythm outcome after AF ablation.[2,9] The reasons for continuing anticoagulation after ablation mainly stem from studies on drug therapy and chiefly concern the risk of long-term recurrence and the risk of asymptomatic recurrences. Several reports published in recent years have indicated that patients who discontinue OAT after successful ablation may have a low rate of stroke. By contrast, a high rate of bleeding has been observed when OAC is continued, determining a risk-benefit ratio in favor of OAC discontinuation. However, none of these reports have provided randomized data, and patients at high risk of CVE have been relatively few in the various studies.

The present study is a comprehensive meta-analysis to compare the efficacy and safety of OAC discontinuation with long-term OAC continuation after successful AF catheter ablation. The overall analysis of 25 177 patients undergoing catheter ablation from 16 observational studies showed no difference in the occurrence of CVE between patients off OAC and those on OAC after AF ablation. Notably, however, the risk of major bleeding was seen to be considerably higher in patients on OAC, with a relative risk reduction of 83% in patients off OAC. Our results suggest that the risk-benefit ratio favored the suspension of OAT after successful AF ablation even in patients at moderate-high risk of TE. These results need to be carefully evaluated. First, it appears from the baseline characteristics of the studies examined that off-OAC patients have lower CHADS2 scores than on-OAC patients; this probably reflects the reluctance of clinicians to discontinue OAC in high thromboembolic risk patients and the influence of current guidelines. However, in our analysis, patients with CHADS2 or CHA2DS2-VASc score of 1 were analyzed together with patients with CHADS2 or CHA2DS2-VASc score of 0, even though the thromboembolic risk profile is different and the recent European Guidelines, in their general anticoagulation recommendations for OAC, suggest that this treatment should be considered in male AF patients with a CHA2DS2-VASc score of 1.[30] Moreover, the subanalysis carried out on patients with CHADS2 and CHA2DS2-VASc score included 948 and 1358 patients, respectively, and no statistically significant difference in the occurrence of CVE emerged between patients on and off OAC. These findings seem to suggest the safety of OAC discontinuation after successful ablation, even in patients with higher thromboembolic risk scores.

Of note, the papers analyzed did not provide sufficient data to predict the risk of bleeding in the population studied. These data could, however, be crucial for patients at high risk of bleeding, in whom the discontinuation of OAC could prevent serious adverse events without increasing the thromboembolic risk.

"Real-life" evidence[31] has suggested that, after successful ablation, the main driving criterion for stopping anticoagulant therapy is the absence of arrhythmias recurrences and is immediately restarted in case of AF recurrences. Regrettably, the studies did not directly assess the impact of AF recurrence on the incidence of CVE. However, the results of the regression model may suggest that recurrence of AF could represent a clinical significant variable to take into account in guiding OAC after ablation. In addition, the use of aspirin in this subset of patients need to be further explored.

Overall, our analysis showed a 1.3% incidence of CVE (321 on 25 117), a value that is half of that previously reported in a group of patients who had undergone AF ablation.[32] Reynolds et al conducted a propensity-score matching analysis of the MarketScan Research database and reported a 1.6% annual stroke rate in patients treated with catheter ablation vs 2.6% in those not ablated.[33] Having analyzed the Intermountain Healthcare Database, Bunch et al reported an incidence of stroke in patients who had undergone catheter ablation of AF that was similar to that of the general population without AF.[34] Hunter et al reported that maintenance of sinus rhythm was the most significant predictor of stroke-free survival in patients who underwent AF ablation.[21] Although large-scale, prospective, randomized clinical trials have not been performed to evaluate the impact of AF ablation on thromboembolic risk, other more recent evidence seems to show the reduction of this risk after successful ablation.[35,36] In line with these literature data, in our meta-analysis the percentage of patients with CVE after ablation was low and did not increase following OAC discontinuation.


There are several potential limitations in our meta-analysis. First, there was a high degree of heterogeneity among the studies included. This can probably be explained by the types of studies available: all were cohort studies, a considerable proportion were retrospective, and populations, design, follow-up duration, and endpoints differed. Naturally, this increases the heterogeneity of the baseline characteristics of the populations included and outcome assessment. Moreover, a very small number of studies reported the outcomes according to the class of risk in the group and were able to be included in the stratified analysis. The Egger test revealed a publication bias, mostly regarding studies with smaller sample sizes. However, the Egger test could also reflect true heterogeneity due to the intensity of treatment or underlying differences in risk. Second, as a meta-analysis of observational studies, the differences between the two groups might be explained by unobserved confounders. Third, owing to the absence of patient-level data, we were unable to adjust for patient characteristics, which may be potential confounders. Hence, the nonsignificant CVE incidence in the on-OAC vs the off-OAC group needs to be considered with caution. In addition, given the lack of data on the bleeding risk (eg, HAS-BLED scores) in the papers included in the meta-analysis, the comparison of bleeding events on and off OAC becomes more difficult to interpret.

Of note, all the studies but two used warfarin as OAC treatment. Whether our results can be extended also to non-vitamin K antagonist oral anticoagulants warrants further investigations. Given the relevance of this issue, randomized trials to assess the impact of thromboembolic risk stratification are needed to establish the indication for postablation OAC. With this clinical equipoise in mind, two randomized trials (NCT01959425 and NCT02168829) are currently underway to assess whether OAC discontinuation after successful catheter ablation is safe in patients at high risk of thromboembolic events.

Finally, it should be acknowledged that an overlap between patients included in two different published studies might not be excluded.[27,11]