Stopping oral anticoagulation after successful ablation for atrial fibrillation: reality or fiction?

This editorial refers to ‘The safety of discontinuation of oral anticoagulation therapy after apparently successful atrial fibrillation ablation: a report from the Chinese Atrial Fibrillation Registry study’, by W.-Y. Yang et al., on pages 90–99. The main indication for ablation of atrial fibrillation (AF) is for symptomatic improvement and increased patient quality of life. Current guidelines also suggest that we should base decisions regarding long-term continuation of oral anticoagulation (OAC) post-ablation on patient risk rather than success of the procedure. However, this recommendation is based on expert consensus with no strong evidence to guide such a recommendation.1 Furthermore, many patients dislike being on long-term OAC and seek ablation as a way of eliminating the need for OAC. From a physician perspective, convincing a patient with no more AF to stay on OAC is a challenging discussion. We sometimes end up discontinuing OAC in very insistent, but high-risk patients, while documenting the lack of evidence to guide such a decision. In this context, the study by Yang et al.2 offers additional data to a growing number of studies suggesting the safety of stopping OAC in moderate to high-risk patients post-successful AF ablation. Using data from the prospective CHINA AF registry, the authors report on 4512 consecutive patients who underwent ablation from 2011 to 2017. Patients were divided into those who continued OAC (n = 1363) vs. those who did not (n = 3149) 3 months after ablation. The mean follow-up was about 2 years and the mean CHA2DS2-VASc score was over 2 in both groups. The first major finding was that the ‘on treatment’ rate of thromboembolic events in both groups was much lower than what would have been predicted by the CHA2DS2-VASc score: 0.36 per 100 patient-years in the off-OAC group and 0.35 per 100 patient-years in the on-OAC group. Even in patients with the highest risk scores, the rate of thromboembolism was still low (0.67–1.11 per 100 patient-years). Furthermore, major bleeding rates were also low, ranging from 0.19 in the off-OAC group and 0.35 in the on-OAC group per 100 patient-years. Of the 4212 patients, 29% developed AF recurrences which lead to reinitiation of OAC in patients who initially were off OAC. This is a high rate of crossover, but the ‘on treatment’ and intention-to-treat analyses were not substantially different. The only risk factors identified as predicting thromboembolism was prior stroke and/or transient ischaemic attack and diabetes. In contrast to most prior studies on this topic, the current manuscript has three key strengths: a prospective, multicentre study design; large patient numbers; and inclusion of patients with higher CHA2DS2-VASc score.3,4 But does is prove that we can safely stop OAC post-successful ablation? The first question that must be raised is whether the ablations in this study were actually successful. Follow-up started 3 months post-ablation, but most recurrences occur between 3 and 12 months post-ablation with a much lower rate of recurrence beyond 12 months. Therefore, it may have been better to start follow-up at 1 year post-ablation to minimize the number of AF recurrences. Furthermore, the post-ablation monitoring strategy was very limited, with only 42% of patients receiving a follow-up Holter; many recurrences may have been missed. Having said that, the 2 year recurrence rate was reportedly low (29%) and even if it was higher, the thromboembolic event rate remains reassuringly low. But the biggest limitation to the study is the effect of selection bias. Patients who stayed off OAC tended to be younger with a smaller proportion of females, lower comorbidities, lower CHA2DS2-VASc and HAS-BLED scores, and smaller left atrial size. Thus, the off-OAC event rates represent events in a highly selected population of post-ablation patients. In the on-OAC group, we do not know what the event rate would have been off OAC. While the intermediate-risk patients on OAC had a thromboembolic rate of 0.35, the rate jumped to 1.11 in the highest risk patients. If these latter patients had not been on OAC, their event rate would most certainly have been even higher and likely close to the threshold of 1.5–1.6% which normally justifies chronic OAC. On the other hand, the bleeding risk of the patients on OAC was very low and did not substantially change according to bleeding risk profile. Major bleeding rates were 0.29, 0.46, and 0.21 in the low, intermediate, and highest risk groups, respectively, implying that these patients could receive the protective benefits of OAC with minimal downside. So, if only highly selected patients can come off OAC, while others can remain on it with minimal bleeding risk, then why not just continue it in all post-ablation patients? Furthermore, while the authors suggested that patients at the highest risk of stroke included diabetics and those with prior stroke, this finding should not be extrapolated to conclude that ONLY these patients should continue OAC. Obesity was a significant univariable predictor of events and even after multivariable analysis, it remained a borderline predictor with a P-value of 0.06. The model also suggested that OAC discontinuation was not a predictor of events, but only 59/1363 (4%) patients stopped OAC. Such a small number under powers the model’s ability to predict whether stopping OAC increases stroke or not. And finally, the model does not account for a potentially very important predictor of stroke—AF recurrence. In the no-OAC group, 824 patients developed an AF recurrence, of which 149 immediately initiated OAC and another 171 resumed OAC later. That leaves 504 patients in the no-OAC arm that had AF recurrence but did not reinitiate OAC. What was the event rate in these patients? This is not specified in the study. Even if the event rate was known, and was found to be low, what was the duration of AF recurrence in these patients? Were patients with longer durations of recurrence selected to resume OAC while patients with only brief durations were allowed to remain off OAC? What duration of AF recurrence triggered reinitiation of OAC and what duration actually predicts higher stroke rates? Neither of these questions can be answered by this study since the limited monitoring regimen limits the ability to accurately define durations of recurrent AF. There are many other reasons why we should not be so cavalier about stopping OAC post-successful ablation. We know from long-term implantable loop studies that a higher proportion of post-ablation AF recurrences are asymptomatic compared to pre-ablation.5 So without continuous monitoring, recurrent AF may be missed and patients may remain at risk of stroke. Late recurrence of AF also continues past 1 or 2 years. Data from one meta-analysis, for example, shows that late failures of AF ablation may continue at a rate of 4% per year after the first year.6 Even if such recurrences are brief, studies have shown that AF episodes as short as 6 min to 6 h may double stroke risk.7 Does this mean that stopping OAC post-successful ablation is never possible? Certainly not. We just need good evidence from prospective, randomized trials to properly answer the question. Yang et al. and many other observational studies have indicated a low risk of stroke post-AF ablation and there is a rationale to believe these findings as true. Even if AF ablation does not completely eliminate AF, a reduction in AF burden may be related to a reduction in stroke risk. In the non-Vitamin K antagonist oral anticoagulant trials, for example, patients with paroxysmal AF had a smaller risk of thromboembolism than those with persistent AF.8,9 And even though the ASSERT trial showed that 6 min of AF could double stroke risk, most of the risk was actually driven by AF episodes greater than 24 h. There is also evidence showing that the highest risk of thromboembolic stroke occurs within 5 days of a prolonged episode of AF, which suggests a temporal correlation between AF and embolic stroke.10 A properly conducted, multicentre, randomized trial is needed to compare the on and off OAC strategies post-ablation in a moderate to high-risk population post-successful AF ablation. We are hopeful that the OCEAN (Optimal antiCoagulation in Enhanced risk patients post-AF ablation, NCT02168829) trial will be able to answer this question. The trial is comparing a strategy of OAC with rivaroxaban to ASA alone in patients who are 1 year or more after a successful AF ablation(s). Patients will be followed for 3 years with baseline cerebral magnetic resonance imaging (MRI) and serial MRIs at 1 and 3 years. The primary endpoint will be the composite of stroke, systemic embolism, and covert stroke on MRI defined as a cardioembolic stroke >15 mm. Thus far, 650 of 1500 patients have been enrolled and we hope that results will finally answer whether cessation of OAC post-successful AF ablation is a reality for some patients, or pure fiction. Conflict of interest: M. Terricabras declares no conflicts of interest. A. Verma reports grants from Bayer, Biosense Webster, Biotronik, and Medtronic. The opinions expressed in this article are not necessarily those of the Editors of Europace or of the European Society of Cardiology.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.