Recurrent ischemic stroke in patients with atrial fibrillation ablation and prior stroke: A study based on etiological classification

Abstract Background Different subtypes of ischemic stroke may have different risk factors, clinical features, and prognoses. This study investigated the incidence and mode of stroke recurrence in patients with a history of stroke who underwent atrial fibrillation (AF) ablation. Methods Of 825 patients who underwent AF ablation from 2006 to 2016, 77 patients (9.3%, median age 69 years) with a prior ischemic stroke were identified. Patients were classified as those with prior cardioembolic (CE) stroke (n = 55) and those with prior non‐CE stroke (n = 22). The incidence and pattern of stroke recurrence were investigated. Results The incidence of asymptomatic AF (54.5% vs 22.7%; P = .011) and left atrial volume (135.8 mL vs 109.3 mL; P = .024) was greater in the CE group than in the non‐CE group. Anticoagulation treatment was discontinued at an average of 28.1 months following the initial ablation in 34 (44.2%) patients. None of the patients developed CE stroke during a median 4.1‐year follow‐up. In the non‐CE group, 2 patients experienced recurrent non‐CE stroke (lacunar infarction in 1 and atherosclerotic stroke in 1); however, AF was not observed at the onset of recurrent ischemic stroke. Conclusions In patients with a history of stroke who underwent catheter ablation for AF, the incidence of recurrent stroke was 0.54/100 patient‐years. The previous stroke in these patients may not have been due to AF in some cases; therefore, a large‐scale prospective study is warranted to identify the appro priate antithrombotic therapy for the prevention of potentially recurrent stroke.


| INTRODUC TI ON
Atrial fibrillation (AF) increases the risk of a cerebrovascular event (CVE) up to 5 times. 1 Patients with AF and history of stroke are at higher risk of death, heart failure, and long-term disability. Catheter ablation is the standard procedure for treating patients with AF, and recent studies have suggested that it may further reduce the risk of thromboembolism. [2][3][4] For antithrombotic therapy after catheter ablation for AF, current guidelines recommend continued oral anticoagulation (OAC) therapy based on the CHA 2 DS 2 -VASc score risk profile for all patients. 5 However, in clinical practice, OAC therapy has been discontinued for many patients with a low-risk profile for thromboembolism. Very few studies have described outcomes in high-risk patients with apparently successful AF ablation following discontinuation of OAC therapy. Furthermore, these studies did not provide details of the type of stroke experienced (ie, whether past CVEs were cardiogenic embolisms). We hypothesized that different subtypes of ischemic stroke may present different risk factors, clinical features, and prognosis; therefore, the best post-procedural antithrombotic management for AF ablation may differ in patients with prior cardioembolic (CE) stroke and prior non-CE (ie, non-AF related) stroke. This retrospective, observational study investigated the incidence and mode of ischemic stroke recurrence in patients with a history of ischemic stroke that have undergone AF ablation. Retrospective information about the history of stroke (based on self-report and patient records) and risk factors were extracted from clinical medical records. Nine patients with systemic emboli, considered a risk factor in the CHA 2 DS 2 -VASc score, were excluded from this study (involvement of the peripheral or coronary artery in 6 and 3, respectively). All patients provided written informed consent before undergoing the procedures. Data collection was approved by the institutional review board of the hospital.

| Classification of ischemic stroke
The causes of ischemic stroke were classified according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification. 6 The findings specified in the classification of ischemic stroke were as follows: symptoms at onset, computed tomography (CT) or magnetic resonance imaging (MRI) findings in the acute phase, interpretation of images by radiologists or neurologists, and history of AF. If necessary, carotid artery ultrasonography, CT angiography, and magnetic resonance angiography were also performed.
When multiple causes were considered, we classified them as more reasonable causes. Etiologies of ischemic stroke were classi- and (f) TIA. Patients were divided into 2 groups: prior CE stroke (CE group) and non-CE stroke (non-CE group). The incidence and pattern of recurrence of stroke during the long-term follow-up in both groups were investigated.

| Catheter ablation protocol
Our ablation protocol has been described previously. 7 Briefly, transesophageal echocardiography was performed to exclude the formation of intra-atrial thrombus the day before the procedure. All patients underwent circumferential isolation of the ipsilateral PVs at the antrum. A 3.5-mm irrigated-tip catheter (ThermoCool Navistar or ThermoCool Navistar STSF, Biosense Webster; or Cool Path, St. Jude Medical) and a three-dimensional anatomical mapping system (CARTO, Biosense Webster; or Ensite, St. Jude Medical) were used for mapping and ablation. In cases using an ablation catheter with a contact force sensor, the contact force was controlled to achieve a >450 force-time integral for each point. From December 2012, ablation with a cryoballoon was performed for patients with paroxysmal AF except for left common trunk and patients with chronic kidney disease or contrast agent allergy. For cryoablation, a 28 mm cryoballoon (Arctic Front Advance; Medtronic) was used with confirmation of PV occlusion with contrast injection. In general, each cryoablation was performed with a target ablation time of 180 seconds.
If electrical isolation was not achieved despite up to 3 applications, touch-up ablation was performed using an irrigated-tip radiofrequency catheter.
The endpoint of PV isolation was the achievement of a bidirectional conduction block between the LA and PVs. After confirmation of the complete bidirectional block, continuous intravenous administration of isoproterenol (4 μg/min) was initiated, followed by a bolus injection of 40 mg of adenosine triphosphate to exclude reconduction or dormant conduction between the LA and the PVs.
Catheter ablation was performed to eliminate the presence of reconnection and/or dormant conduction. If non-PV foci were identified with drug infusions, catheter ablation was applied to the foci.
When non-PV foci were located in the superior vena cava (SVC), circumferential isolation of the SVC was performed. In patients with paroxysmal AF and persistent AF lasting for ≤1 year, circumferential PV antrum isolation was performed. In patients who had longstanding persistent AF (>1 year), PV isolation was performed, followed by mapping and ablation of complex fractionated atrial electrograms.
Additionally, linear ablation of a roof and mitral isthmus line or left atrial posterior wall box isolation was performed at the physician's discretion.

| Pre-and periprocedural anticoagulation
Oral anticoagulations at the time of ablation included warfarin in 53 patients and direct oral anticoagulants (DOACs) in 24 patients (Table 1). There was no significant difference in the type of OACs between the CE and non-CE groups. Warfarin was not interrupted unless there was a marked prothrombin time-international normalized ratio prolongation. All DOACs were suspended the morning of the procedure. During ablation, we administered 90 units/kg of heparin before the transseptal puncture, measured the activated clotting time every 10 minutes, and administered an additional dose of heparin to maintain activated clotting times between 350 and 400 seconds.  for recurrent AF in patients who have discontinued OAC therapy.

| Repeat ablation procedure
After confirming the absence of AF recurrence after the last session, OAC was either terminated again or continued in some cases, and the decision was left to the discretion of the physician.

| Follow-up and clinical outcomes
Following the ablation procedure, all patients were discharged from the hospital with prescription of OACs. Recurrence of atrial tachyarrhythmia was evaluated based on the patient's symptoms, resting 12-lead electrocardiogram findings at regular visits at the outpatient clinic, and 24-hour Holter ambulatory monitoring results at 1, 3, 6 months, and every 6 months thereafter the final procedure.
Recurrent atrial tachyarrhythmia was defined as a documented AF/ flutter/tachycardia lasting >30 seconds without a blanking period.
OAC discontinuation during follow-up was permitted in selected patients, and restarting of OACs was allowed in cases with recurrent atrial tachyarrhythmia according to the physician's discretion.
When patients were admitted for ischemic stroke at other institutions, medical history, and examination results were collected from the attending neurologists.
The following outcomes were compared: the primary outcome was the recurrence of ischemic stroke, and secondary outcomes were the occurrence of major bleeding events, the time to discontinuation of OACs from the time of initial catheter ablation, and recurrence of atrial tachyarrhythmias.  OACs used initially were warfarin (n = 53), dabigatran (n = 6), rivaroxaban (n = 9), apixaban (n = 7), and edoxaban (n = 2).

| Ablation procedure and periprocedural complications
Overall Log-Rank P = 0.303

| Follow-up outcomes
Despite multiple ablation procedures (1.8 per patient), the rhythm control strategy to maintain sinus rhythm was switched to rate control (followed as permanent AF) in 2 patients ( The continuation rate of OAC therapy stratified by the CHA 2 DS 2 -VASc score is shown in Table 3. Given the concern for bleeding risks, the continuation rate of OAC therapy increased from the CHA 2 DS 2 -VASc score 2 points to 5 points, but conversely it decreased at >6 points. The reasons for OAC discontinuation were prior non-CE stroke in 8 patients, bleeding event/concern in 7, rhythm monitoring by cardiac implantable electronic devices (CIEDs) in 3 and by portable electrocardiogram in 2, and patient preference in 14 (Table 4). OACs were discontinued earlier in patients with bleeding events or those at risk for bleeding events than in other patients.
The Kaplan-Meier curves for recurrent stroke in the CE and non-CE groups are shown in Figure 3. Importantly, none of the patients experienced recurrence of CE stroke in the CE group. In the non-CE group, 2 patients exhibited recurrence of CVEs (log-rank; P = .03); however, atrial tachyarrhythmia were not documented at the onset of the recurrent ischemic stroke (lacunar infarction in 1 and atherosclerotic stroke in 1 patient). One patient discontinued OAC at the time of recurrence, and another patient was receiving OAC. Case 1 was receiving aspirin, but Case 2 was not receiving aspirin. Kaplan-Meier curves for recurrent stroke in on-OAC vs off-OAC could not be drawn because there was only 1 event in each group (P = .754).

| Cases
We describe 2 cases of recurring ischemic stroke after successful AF ablation.

| Case 1
A 52-year-old woman with symptomatic drug-refractory paroxysmal AF and prior lacunar infarction was referred to our hospital for catheter ablation. Antithrombotic therapy with warfarin and aspirin was continued after initial catheter ablation. A second ablation was performed for AF recurrence, with no subsequent events over the 4-year follow-up. The patient discontinued OACs and continued only aspirin; however, 6 years after the first ablation, a new cerebral infarction recurred. The subtype of recurrent stroke was lacunar infarct, and antiplatelet therapy was changed from aspirin to clopidogrel and no recurrence was observed for 3 years.

| Case 2
A 72-year-old man with symptomatic paroxysmal AF was referred to our hospital for catheter ablation. Embolism risk factors were absent except for age and preventive therapy with AAD was started.
Although the palpitation attack disappeared after administration of drugs, a transient dysarthria appeared after 3 months. Before transfer to the stroke center by ambulance, the speech disorder improved and the MRI study revealed a small acute infarction due to atherosclerosis. Prior to this stroke, the patient had been completely unaware of palpitations and the electrocardiogram showed sinus rhythm. OAC therapy was introduced, and catheter ablation for AF was performed. The patient followed an uneventful course without complications and anticoagulation therapy with dabigatran was continued. Ten months later, speech disturbance and cerebral infarction recurred. Again, there was no evidence of AF, the cause of cerebral infarction was diagnosed as hemodynamically significant cerebral artery narrowing, and antiplatelet therapy was added.

| D ISCUSS I ON
This study showed (a) a low incidence (0.54/100 patient-years) of cerebral infarction recurrence in patients with a history of ischemic stroke who underwent AF ablation; (b) CE was the cause of approximately 70% of prior cerebral infarction, many patients had no

Proportion of free for stroke recurrence
Log-Rank P = 0.030 symptoms of AF, and the LA was enlarged compared to non-CE patients; and (c) a non-CE mechanism (lacunar infarction and atherosclerotic stroke) was the cause of stroke in 2 patients with recurrent CVEs without atrial tachyarrhythmia recurrence.

| Proportion of prior stroke patients in ablated patients and procedure safety
In reports from National Registry data of catheter ablation for AF, 5%-10% of patients had a history of ischemic CVEs, with 8.1% reported in Japan, 8

| Risk reduction of stroke after ablation and discontinuing OAC therapy
Previous observational studies have suggested an association between catheter ablation and a reduction in stroke risk. [2][3][4] Furthermore, several recent reports have indicated that patients discontinuing OACs after successful ablation may experience a low stroke rate. [13][14][15]

| Subtype of recurrent ischemic stroke and antithrombotic therapy
Hata et al showed higher recurrence rates of stroke in a Japanese community than in western populations. 22 The recurrence rates ing warfarin (P = .02), while the proportion of non-CE stroke was lower in those taking aspirin (P = .06). Most (56%) ischemic strokes occurring in AF patients taking adjusted-dose warfarin were non-CE; thus, aspirin in AF patients appeared to primarily reduce non-CE strokes.

| Study limitations
This study was a small sample, single-center, non-randomized retrospective study where interruption or continuation of OAC therapy was left to the discretion of the physician. Attempting to distinguish which stroke subtypes would respond to a particular strategy is a great problem for small retrospective sample sizes; therefore, a large prospective study is warranted. Since recurrence of cerebral infarction was not routinely screened using imaging findings such as MRI, cerebral infarction with asymptomatic, or poor symptoms may have been overlooked. An intracranial and/or extracranial arteriography by MRI or ultrasonography was not routinely performed for all patients to assess atherosclerotic involvement. Asymptomatic AF recurrence is a very important issue for patients with a high risk of embolism. Even in patients presenting with highly symptomatic AF, asymptomatic episodes may occur and significantly increase after ablation. Conventional follow-up such as 24-hour Holter monitoring is not sufficient to identify asymptomatic AF recurrences after ablation. In this retrospective study, assessment for AF recurrence was performed with rhythm monitoring by CIEDs in 3 patients and well-trained self-monitoring using portable electrocardiogram in 2 patients. We believe that this problem may be solved by advances in wearable monitoring technology.

| CON CLUS IONS
In patients who underwent catheter ablation for AF, the incidence of recurrent stroke was 0.54/100 patient-years for individuals with prior stroke. While prior stroke in some cases may not have been due to AF, careful follow-up with optimal antithrombotic therapy is needed to minimize the risk of recurrent stroke following catheter ablation.

ACK N OWLED G EM ENTS
Language editing was performed by professional editors at Editage, a division of Cactus Communications.

CO N FLI C T O F I NTE R E S T
Authors declare no conflict of interests for this article.