Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, and it is a global problem [1]. Compared with normal, age-matched control subjects, AF increases the risk of ischemic stroke by 4-5 fold [2,3]. Stroke is the second most common cause of death and it is the main cause of disability [4]. The prevalence of AF and the associated stroke risk are higher among elderly patients. The stroke risk is independent of whether the AF is paroxysmal, persistent, or permanent. Paroxysmal AF often causes a cryptogenic ischemic stroke or a transient ischemic attack (TIA). Embolisms with cardiac origins cause between 17% and 30% of all ischemic strokes, but it is estimated that the causes of up to 40% of ischemic strokes are unknown [5]. Recent developments in monitoring devices have provided more robust evidence of an association between AF and stroke, especially cryptogenic stroke. Newly detected AF is identified in 10% of the patients who experience a stroke or a TIA, and an additional 11% cases of AF are newly detected when patients undergo 30 days of continuous electrocardiographic monitoring [6-8].

A key approach towards the management of AF is to reduce the risk of stroke by using oral anticoagulation (OAC) therapy. The novel oral anticoagulants are not inferior to warfarin for preventing stroke in nonvalvular AF [9-11], and they are the preferred OAC option for patients who are at high risks of ischemic stroke and have congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, and prior stroke or TIA or thromboembolism (CHADS₂) score or a CHA₂DS₂-vascular disease, age 65-74 years, and sex category (VASc) score of ≥1 or 2. Left atrial appendage (LAA) occlusion devices are not inferior to warfarin, and they help patients who have the contraindications associated with the long-term use of OAC therapy and high risks of bleeding during OAC therapy [12,13]. Furthermore, rhythm control using catheter ablation is an emerging treatment option for AF. Catheter ablation is more efficient at reducing the AF burden compared with antiarrhythmic drugs (AADs). However, the effectiveness of catheter ablation at reducing the risks of mortality and stroke is not yet evident. The current American Heart Association, American College of Cardiology, and Heart Rhythm Society guidelines suggest that catheter ablation for AF should not be performed on patients who cannot be treated with OAC therapy during and after the procedure. Catheter ablation for AF to restore sinus rhythm should not be performed with the sole intent of obviating the need for OAC therapy. There is insufficient evidence available to determine whether catheter ablation for AF reduces all-cause mortality, stroke, and heart failure.

Thromboembolic events in the control arm of the RAte Control versus Electrical cardioversion study were more common in the patients who had stopped OAC therapy or had unsatisfactory international normalized ratios [14]. Similarly, most of the patients who experienced strokes in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management study were not receiving warfarin or they had suboptimal international normalized ratios at the times of their strokes [15]. Clearly, the premature cessation of OAC therapy in one-third of the rhythm-controlled patients in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management trial rendered them vulnerable to thromboembolic consequences. The strategy underlying AF treatment to reduce the risk of stroke is evolving, especially in the era of novel OAC therapies.

AF is a major independent predictor of ischemic stroke. The estimated risk of AF during a lifetime is between 22% and 26% [16]. AF is associated with abnormal blood stasis that involves atrial hypocontractility and the loss of the atrial kick, atrial structural remodeling, and the activation of platelets and the coagulation cascade, which is known as Virchow’s triad. In addition, AF provides a prothrombotic environment through atrial dilation and stretching, a reduction in the pectinate muscle volume, the loss of a normal endocardial surface, endothelial dysfunction, and the release of coagulation factors, namely, factor Xa and thrombin. These AF characteristics promote thrombus formation and ischemic stroke. AF is a significant underlying cause of cardiac emboli in approximately 50% of patients. Stroke of an unknown cause is more frequent in younger patients. Cryptogenic stroke causes up to 40% of the strokes in patients who are <50 years of age. It is difficult to establish a treatment strategy for stroke that has a cryptogenic cause. Although a cardiac origin for emboli may be suspected, it is difficult to verify, especially when inadequate monitoring devices are used. The most commonly used devices for monitoring arrhythmias are 24 -48 hours Holter and event monitors. However, the AF detection rate for these devices is between 2.5% and 16% [17,18].

The data from recently implanted devices show that the prevalence of AF is higher in association with cryptogenic stroke. AF or atrial tachycardia (AT) is common among the arrhythmias, and this has been determined from the interrogation of the devices. The specificity of AF or AT detection using implantable devices is 98.5%. The findings from the Atrial Fibrillation Reduction Atrial Pacing Trial showed that ischemic stroke was associated with the occurrence of subclinical AF (hazard ratio [HR], 1.76; 95% confidence interval [CI], 0.99-3.11; P=0.05) [19]. The IMPACT trial was designed to test the hypothesis that remote monitoring would guide the initiation and cessation of the administration of OAC therapy to patients with implanted devices and would improve clinical outcomes by reducing the stroke risk, the development of systemic embolisms, and major bleeding [20,21]. The investigators randomized 2,718 patients with dual-chamber and biventricular defibrillators to start and stop the administration of OAC therapy using remote rhythm monitoring or ordinary office-based follow-up assessments that involved determining anticoagulation using standard clinical criteria. The primary events, including stroke, systemic embolism, and major bleeding, did not differ between the group of patients whose rhythms were remotely monitored (2.4 per 100 patient-years) and those who underwent ordinary office-based follow-up assessments (2.3 per 100 patient-years) (HR, 1.06; 95% CI, 0.75-1.51; P=0.732). Within the group of patients with AT, the thromboembolism rates were 1.0 per 100 patient-years for those whose rhythms were remotely monitored and 1.6 per 100 patient-years for those who underwent office-based follow-up assessments (relative risk, -35.3%; 95% CI, -70.8-35.3%; P=0.251). Although there was no temporal relationship between AT and stroke, the AT burden was associated with thromboembolism.

In relation to anatomical remodeling, endothelial dysfunction, which includes fibrotic changes in patients with AF, has been defined as atrial myopathy. The findings from previous studies have shown a relationship between the formation of thrombi and atrial fibrosis [22,23]. Atrial fibrosis and a decline in the contractility of the left atrial (LA) myocardium result in blood flow stasis, especially within the LAA. An inverse relationship between flow and the CHA₂DS₂-VASc score has been determined. The parameters that comprise the CHA₂DS₂-VASc score and the factors associated with a decline in the LAA flow velocity on transesophageal echocardiography often overlap, which means that LA myopathy with remodeling is associated with an increased risk of stroke in patients with AF [24-26].

Areas showing late gadolinium enhancement (LGE) on cardiovascular magnetic resonance images have been used to detect atrial fibrosis [27-29]. While T1 mapping was developed to identify diffuse myocardial fibrosis, it has recently been applied to the left atrium, and it has a correlation with the LA electrogram voltage [30]. The LGE area in the left atrium is related to a reduced LAA flow velocity and the clinical outcomes after treatment [31]. LGE cardiovascular magnetic resonance imaging of the left atrium is a noninvasive technique that assesses the structure and function of the heart. Its advantages are associated with not exposing patients to radiation, the high temporal resolution, and the functional analyses. It is also useful to determine the tissue composition of the left atrium. However, the application of and the LGE techniques have not been standardized, and there are some technical problems to be resolved. The atrial myocardial wall is too thin to accurately analyze the LGE area using cardiovascular magnetic resonance imaging. The irregular ventricular rhythms render functional analyses complex during AF. However, in the near future it has the potential to be a valuable method for predicting clinical outcomes, including the stroke risk.

Complex structures with areas of relatively low flow predispose patients to stasis, especially during AF. In patients with nonvalvular AF, 90% of the thrombi are located in the LAA. LAA enlargements and reductions in LAA flow velocities are associated with increases in thrombus formation and stroke [32-34]. Patients with AF undergo remodeling of the LAA in association with increased appendage volumes.

The findings from a study undertaken by Mügge et al. [35] showed that the group of patients with nonvalvular AF had an increased risk of thrombosis, especially in the subgroup with a low LAA emptying flow rate of <250 mm/s. The incidence of ischemic stroke and peripheral embolic events in the subgroup with a low flow rate was significantly higher (60%) than that in the subgroup with a high flow rate of ≥250 mm/s (5%).

A more specific representation of LAA morphology classifies the LAA shapes into four types, namely, the chicken wing, windsock, cactus, and cauliflower categories. The findings from previous studies have shown that the non-chicken wing types are associated with a greater risk of stroke than the chicken wing type, and that the highest risk is associated with the cauliflower type, even in patients with low CHADS₂ scores. However, the findings from other investigations have not demonstrated a relationship between these specific morphologies and the stroke risk [36-39]. A greater LAA complexity, including more extensive trabeculation and multiple lobes, has been associated with an increased stroke risk, but further studies are warranted to confirm these results.

Damaged or dysfunctional endothelium may lead to the release of prothrombotic and proinflammatory molecules, including von Willebrand factor (vWF), selectins, and asymmetrical dimethyl arginine, and an increase in the vascular tone and/or a reduction in vascular reactivity. Flow-mediated dilation, a clinical indicator of endothelial integrity, declines in AF patients, and is reversed with normal sinus rhythm maintenance after catheter ablation [40]. Flow-mediated dilation is inversely related to the serum C-reactive protein levels in chronic AF patients, which implies disruption caused by inflammation [41]. However, this continuum of endothelial dysfunction and damage appears to accentuate the highly complex pathophysiological processes involved and it perpetuates disease progression in AF. More recently, quantifying the circulating endothelial cells has been proposed as an additional measure of endothelial damage [42].

The circulating C-reactive protein level is elevated in patients with AF compared with those who do not have AF histories, and patients with persistent AF have higher C-reactive protein levels than those with paroxysmal AF [43]. An association has been reported between inflammation and the stroke risk in AF patients. Indeed, high-sensitivity C-reactive protein levels have been shown to be positively correlated with stroke risk factors, including diabetes and hypertension, in AF patients and they have also been shown to be associated with mortality and a composite outcome of ischemic stroke, myocardial infarction, and vascular death [44].

The findings from a study by Lip et al. [45] demonstrated the presence of raised vWF levels in patients with chronic AF that were independent of the underlying structural heart disease. Structural abnormalities within the LA endocardium have been associated with the circulating vWF levels. Abnormal vWF levels also correlate with the risk factors for thrombosis that were identified using transesophageal echocardiography. Indeed, Heppell et al. [46] reported that raised levels of vWF correlated with spontaneous echo contrast and the presence of LA thrombi in patients with non-rheumatic AF. An advanced age, prior cerebral ischemia, heart failure, and diabetes were independently associated with higher levels of vWF. Importantly, adding plasma vWF levels to the clinical stroke risk factors appeared to improve the risk stratification of patients with AF [47]. Asymmetric dimethyl arginine is an endogenous substance produced from proteolysis of methylated arginine-related proteins. Asymmetric dimethyl arginine exerts a competitive inhibitory effect on nitric oxide synthase and it reduces the bioavailability of nitric oxide. Elevated levels of asymmetric dimethyl arginine have been reported in patients with coronary artery disease [48], hypertension [49], diabetes [50], and atherosclerosis [51]. However, further investigations are required into the relationship between asymmetric dimethyl arginine and thrombus formation.

The cornerstone of stroke prevention in AF patient is OAC therapy. Accordingly, OAC therapy to prevent stroke is considered standard care for AF patients, particularly for those at higher risks [52]. However, a substantial proportion of AF patients are not administered OAC therapy [53]. Patient preference, complications associated with bleeding, and noncompliance are common reasons for the underuse of OAC therapy in routine clinical practice [54]. OAC therapy is often accompanied by a risk of bleeding. In clinical trials, the annual rates of major bleeding were 3.1%-3.4% in the warfarin-treated patients and 2.1%-3.6% in those treated with novel OAC agents [9-11]. Silent stroke is associated with elevated risks of clinical stroke and cognitive decline and a two-fold increase in the risk of dementia [55]. Silent stroke is common in AF patients, with an estimated prevalence of between 11% and 48% that was determined from neuroimaging studies [56-59]. Since AF is associated with cognitive impairment and dementia [60], it is conceivable that silent stroke may be the causative link [58,61].

Noninvasive rhythm control strategies to restore sinus rhythm using AADs or “pharmacologic cardioversion” and direct electrical currents or “electrical cardioversion” are widely used. They are useful for patients who do not tolerate rate control drugs or patients with refractory symptoms. Maintenance of sinus rhythm with AAD has been considered to reduce stroke in AF patients ideally. However, the findings from clinical trials have failed to show any differences between rhythm control using AADs and rate control with respect to reducing the risks of ischemic stroke or systemic thromboembolism [14,15]. Rhythm control using AADs is not effective at maintaining sinus rhythm, and the 1-year efficacy rates vary from 30% to 60% [52]. In addition, the rate of AAD discontinuation is high at 10% per year, which further diminishes the likelihood of maintaining sinus rhythm [62]. The published success rates for the elimination of AF using catheter ablation dramatically exceed those associated with AADs [63-65].

Catheter ablation has emerged as the rhythm control strategy for symptomatic AF. It was superior to AADs at reducing AF recurrences in drug-refractory patients [66]. Analyses of the data within large health-oriented databases have demonstrated lower rates of stroke or TIA in patients who underwent catheter ablation for AF compared with those who did not undergo catheter ablation for AF [67-69]. Catheter ablation is commonly used to isolate arrhythmogenic myocardial tissue via energy delivery. The catheter uses an energy source, for example, radiofrequency energy, cryo energy, or laser energy, to create a scar tissue lesion. It is generally less invasive than surgery for AF. In some cases, balloon catheters or catheters with unique designs are used for catheter ablation for AF. The findings from a previous study have shown that the pulmonary veins are a main trigger of AF [70]. Hence, catheter ablation for AF usually consists of pulmonary vein isolation and elective substrate modification.

Catheter ablation for AF for the maintenance of sinus rhythm has led to the discontinuation of OAC therapy in selected patients whose ablation procedures were successful. Some studies’ reports describe very low rates of ischemic stroke among patients whose OAC therapy was discontinued after successful catheter ablation for AF [71-77]. Most patients had low baseline stroke risks with CHADS2 scores of <2, and the risk of stroke or TIA was very low at <0.7% per year. Some studies had follow-up durations that were <3 years. Vitamin K antagonists were the OAC agents of choice. All of the studies were non-randomized and retrospective, and the follow-up durations were relatively short. None of the studies employed continuous AF monitoring using implantable loop recorders to assess late AF recurrences following evident success. Only a small percentage (<25%) of the patients were at a high risk of stroke, that is, a CHADS₂ score of 2, and asymptomatic strokes were not assessed. The findings from a recent study that analyzed Swedish health registry data from a large population, showed that compared with propensity score-matched control subjects who received medical treatment, catheter ablation can prevent ischemic strokes [78]. Although this was a nonrandomized retrospective study, catheter ablation was associated with a lower risk of ischemic stroke (HR, 0.69; 95% CI, 0.51-0.93). Interestingly, it was also associated with a lower risk of mortality (HR, 0.50; 95% CI, 0.37-0.62). The stroke reduction effect was more pronounced among patients with CHA₂DS₂-VASc scores of ≥2 (HR, 0.39; 95% CI, 0.19-0.78).

Ongoing clinical trials, namely, the Catheter Ablation Versus Antiarrhythmic Drug Therapy for Atrial Fibrillation trial and the Early Therapy of Atrial Fibrillation for Stroke Prevention Trial, should provide new information that will help to determine whether catheter ablation for AF is superior to standard therapy with either rate control or rhythm control drugs at reducing total mortality and other secondary outcome measures. These important trials will help to address whether catheter ablation provides benefits beyond quality of life improvements.

Ideally, catheter ablation to treat AF should help to prevent ischemic stroke. However, there is no robust evidence available that can assure the safety of discontinuing OAC therapy after ablation. Catheter ablation in routine clinical practice has several limitations. The substantial recurrence rate after ablation and ineffective postablation monitoring devices attenuate the effects of catheter ablation. A number of centers have reported rates of AF recurrence within 1 year of catheter ablation of between 15% and 40% [79-82]. The recurrence of AF after ablation is usually monitored using outpatient clinic-based Holter electrocardiography event recorders. These monitoring devices only disclose patients’ rhythms during extremely short timeframes relative to their entire lifetimes. Thromboembolic events often occur without evidence of AF recurrences. Although patients with implantable devices can be closely observed, they only represent a minority of the patients who require monitoring. The clinical significance of atrial high-rate episodes that are detected using device diagnostics is not yet evident, and whether we can define an atrial high-rate episode as a recurrence or whether an atrial high-rate episode increases the thromboembolic risk remains uncertain.

Innovations in ablation techniques and discovering recurrence predictors can reduce postprocedural recurrences and, hence, the thromboembolic risk. However, more sensitive monitoring methods are needed to provide the necessary assurance that OAC therapy can be discontinued after catheter ablation.

The stroke risk in AF patients is not only linked to the presence or absence of AF, but it is also related to the underlying systemic endothelial dysfunction and the presence of prothrombotic conditions. Rhythm control using AADs has failed to show benefits with respect to mortality or stroke prevention. Catheter ablation is superior to AADs in relation to the restoration and maintenance of sinus rhythm. Data are accumulating that demonstrate that patients who undergo successful catheter ablation for AF may have a low incidence of subsequent thromboembolic events. Therefore, discontinuing OAC therapy after successful ablation may result in a low thromboembolic risk and reduce the bleeding risk. OAC therapy after catheter ablation for AF should be individualized and consider not only the conventional risk score, but also the extent of the atrial fibrosis, the LAA morphology, and endothelial dysfunction.