Stroke and thromboembolism prevention in atrial fibrillation

Abstract

Prevention of stroke and systemic thromboembolism remains the cornerstone for management of atrial fibrillation (AF) and flutter. Multiple risk assessment models for stroke and systemic thromboembolism are currently available. The score, with its known limitations, remains as the recommended risk stratification tool in most major guidelines. Once at-risk patients are identified, vitamin K antagonists (VKAs) and, more recently, direct oral anticoagulants (DOACs) are the primary medical therapy for stroke prevention. In those with contraindication for long-term anticoagulation, left atrial appendage occluding devices are developing as a possible alternative therapy. Some controversy exists regarding anticoagulation management for cardioversion of acute AF (<48 hours); however, systemic anticoagulation precardioversion and postcardioversion is recommended for those with longer duration of AF. Anticoagulation management peri-AF ablation is also evolving. Uninterrupted VKA and DOAC therapy has been shown to reduce perioperative thromboembolic risk with no significant escalation in major bleeding. Currently, under investigation is a minimally interrupted approach to anticoagulation with DOACs periablation. Questions remain, especially regarding the delivery of anticoagulation care and integration of wearable rhythm monitors in AF management.

INTRODUCTION

Atrial fibrillation (AF) is the most common cardiac arrhythmia, impacting over 33 million people worldwide. It is a leading cause of cardiovascular disease and death globally.¹ Given the association between AF, obesity, hypertension and other rapidly growing health epidemics, the incidence of AF is estimated to continue growing over the next few decades. AF disproportionally afflicts older patients who often have multiple comorbid conditions that raise the risk of various AF-related complications.²

ASSOCIATI ON BETWEEN AF AND STROKE

In general, the most concerning complication of AF is cardioembolic stroke. Due to abnormal blood flow in the left atrium associated with disorganised electrical signals and an absence of coordinated atrial contractions, in addition to endothelial dysfunction and other prothrombotic conditions, thrombus often forms in the left atrial appendage (LAA). When that thrombus ‘breaks free’, it can embolide to the peripheral or (more commonly) the cerebral arterial beds. This is particularly concerning given that patients who experience AF-related embolic stroke have worse outcomes than those who experience stroke not related to AF.^3,4

Multiple risk assessment models have been developed to estimate an individual patient’s risk of stroke or systemic thromboembolism (table 1). First developed in 2001, the CHADS₂ score was used broadly by clinicians for a decade.⁵ This score, developed by expert consensus and drawing from analysis of a nationwide registry, assigns points based on age in addition to a number of comorbid risk factors. Unlike some prior risk stratification schemes, all risk factors are considered equivalent, with the exception of prior stroke or transient ischaemic attack that carried ‘double’ the risk of each other comorbidity. It is also simple to remember and calculate, thanks to its acronym structure: congestive heart failure (with preserved or reduction function), hypertensions, age ≥75 years, diabetes mellitus and stroke ×2. Approximately 20% of patients are categorised as ‘low risk’, 60% as ‘moderate risk’ and 20% as ‘high risk’.⁶ While the CHADS₂ score was found to have a high degree of discrimination in its derivation study (c statistic of 0.82, 95% CI 0.80 to 0.84), it did not perform nearly as well in subsequent validation studies (c statistic of 0.56, 95% CI 0.45 to 0.67).^5,7

Table 1.

Stroke risk assessment models

Risk factor	CHADS2	CHA2DS2-VASc	R2CHADS2	ATRIA (no prior stroke)	ATRIA (prior stroke)
Heart failure	1 point	1 point	1 point	1 point	1 point
Hypertension	1 point	1 point	1 point	1 point	1 point
Age	1 point (≥75 years)	1 point (65–74 years) 2 points (≥75 years)	1 point (≥75 years)	0–6 points	7–9 points
Diabetes mellitus	1 point	1 point	1 point	1 point	1 point
Prior stroke or systemic embolism	2 points	2 points	2 points	n/a	n/a
Vascular disease		1 point
Female sex		1 point (only as an additive risk)		1 point	1 point
Chronic kidney disease			2 points (eGFR<60)	1 point (eGFR <45 or ESRD)	1 point (eGFR <45 or ESRD)
Proteinuria				1 point	1 point

eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease.

In 2010, Lip and colleagues⁷ published the CHA₂DS₂-VASc score as an update to the widely used CHADS₂ score.⁷ This score is also relatively easy to remember and calculate. It was designed to reduce the number of patients in the intermediate risk and to better identify those who were truly low risk for thromboembolic complications. Specifically, the CHA₂DS₂-VASc score further stratifies risk based on age, giving two points for patients aged ≥75 years and one point for patients aged 65–74. It also includes vascular disease (defined as prior myocardial infarction, aortic plaque and vascular arterial disease) as an additional clinical risk factor. Finally, it recognises the role that female sex has as an additive risk factor to the other clinical characteristics. In various analyses, the CHA₂DS₂-VASc score has been shown to have similar or modestly better predictive ability than its predecessor (the CHADS₂ score) and other proposed clinical risk scores.^8–10 Currently, it is incorporated into most major guidelines as the recommended stroke risk stratification tool.^11,12 However, there continues to be some confusion around the inclusion of female sex as a risk factor. As it was initially intended, female sex alone should not be used to make decisions about stroke prophylaxis. However, when other risk factors are present, then women should be considered at higher risk for stroke than men with similar clinical characteristics.

STROKE PREVENTION: MEDICAL THERAPY

Prior to 2000, numerous studies explored the role of oral anticoagulation with vitamin K antagonists (VKAs), such as warfarin, against placebo or aspirin for the prevention of AF-related stroke.¹³ While each study used a different design, including a variety of target INR ranges, the overall results demonstrate a significant benefit associate with VKA therapy as compared with both placebo and aspirin. However, this came at the increased risk of bleeding complications, including intracranial haemorrhage. In general, VKA therapy with a target INR range of 2.0–3.0 was recommended for most high-risk patients.¹³ Either VKA therapy or aspirin therapy was generally recommended for patients at intermediate stroke risk.

Since 2009, four direct oral anticoagulants (DOACs) have been compared with VKA therapy for stroke prevention in non-valvular AF. Each has received approval in both North America and Europe based on a favourable efficacy and safety profile. Of note, there have been concerns raised about the integrity of some data from the ARISTOTLE trial of apixaban versus VKA.¹⁴ However, after a thorough review, the US Food and Drug Administration (FDA) elected to approve this medication with a package label highlighting an overall reduction in stroke and systemic embolism as compared with VKA therapy. Multiple practice-based and claims-based analysis have largely confirmed the overall efficacy and safety of DOAC medications as compared with VKA.^15,16

While DOAC therapy has become mainstream for patients with AF, concerns remain about persistence and adherence to therapy.¹⁷ In one analysis from Ontario, Canada, up to one-third of patients were no longer taking either rivaroxaban or dabigatran 6 months following initiation of therapy.¹⁸ These patients experienced higher rates of stroke and death as compared with patients who continued taking their DOAC medication.

STROKE PREVENTION: NON-MEDICAL THERAPY

Percutaneous closure of the LAA is an alternativeoption for stroke and systemic thromboembolism risk reduction in those unableto tolerate long-term anticoagulation with non-valvular AF. In 2002, proof of concept of LAA closure was demonstrated using the PLAATO device. Sincethen, significant advances have occurred in the field. In 2009, a large randomised-study (PROTECT AF) demonstrated the non-inferiority of the WATCHMAN device tochronic warfarin for the primary composite endpoint of stroke, systemicembolism and cardiovascular death.¹⁹ The subsequent PREVAIL trial in 2014, which was pursued to demonstrate device safety, achieved non-inferiority for stroke and embolism only.²⁰ The WATCHMAN device is now approved for clinical use in both the USA and Europe and has become the best studied and most commonly used worldwide.²¹ Subsequent meta-analyses and practice-basedregistries continue to support its safety and efficacy.^22–24 Device-relatedthrombus formation, however, has been reported in 3.7% of patients-postimplantation. This complication is currently treated with short-termescalation of systemic anticoagulation and is an area of investigative interest.²⁵

Limited data are currently available regarding use of this device in patients with an absolute contraindication for systemic anticoagulation. In the above randomised studies, patients remained on systemic anticoagulation for some time to allow for device endothelialisation or indefinitely with a presence of significant blood flow around the device (>5 mm) on postimplantation transoesophageal echocardiograms. The ASAP study, published in 2013, showed that patients at moderate risk for stroke (mean CHA₂DS₂-VASc score of 4.4±1.7) with a contraindication for oral anticoagulation who underwent WATCHMAN device implantation had a reduced risk of stroke (2.3% per year) compared with the expected thromboembolic risk (7.3% per year).²⁶ The ASAP TOO (NCT02928497) randomised trial is currently enrolling patients with a contraindication to anticoagulation to further assess the WATCHMAN’s efficacy in this population. Despite this evidence, some have questioned the role of WATCHMAN and LAA occlusion for stroke prevention in AF.²⁷

Globally, the Amplatzer devices (Amplatzer Cardiac Plug and Amulet) are the second most commonly implanted, and manufacturing guidelines do not recommend systemic anticoagulation postimplantation. Both the WATCHMAN and the Amplatzer use a self-expanding mesh to occlude the LAA, but the WATCHMAN’s is parachute shaped. while the Amplatzer uses a cylindrical lobe connected to a flat disc. Although available for clinical use in Europe, these devices have not yet received FDA approval in the USA primarily due to the absence of randomised clinical trials. In the largest multicentre observational study (n=1047), Tzikas et al²⁸ found that patients who had the Amplatzer device implanted had a greater than 50% reduction in the risk of stroke compared with what was predicted by their CHA₂DS₂-VASc scores.²⁸ The introduction of the Amulet device in the USA is dependent on the current randomised trials underway (NCT02879448 and NCT02830152).

STROKE PREVENTION: GUIDELINE RECOMMENDATIONS

Both North American and European guidelines favour the use of CHA₂DS₂-VASc score to stratify patients with AF according to their risk of stroke and systemic thromboembolism.^11,12 The most recent versions also agree that stroke prevention should be achieved with anticoagulant therapy, forgoing recommendations in favour of aspirin therapy. Specifically, patients with one or more CHA₂DS₂-VASc risk factors (other than female sex) should be offered anticoagulation to reduce the risk of stroke or systemic embolism. In patients with no CHA₂DS₂-VASc risk elements (or only female sex), no antithrombotic therapy is recommended as the baseline stroke risk is sufficiently low that the net benefit of anticoagulant therapy is most likely outweighed by bleeding risk. Therapy with a DOAC is recommended over VKA in eligible patients given their favourable efficacy profile and reduced risk of serious bleeding. Aspirin monotherapy is not recommended.

For patients with a contraindication to anticoagulant use, both the European and North American guidelines provide a class IIb recommendation to consider the use of LAA occluder devices.^11,12 They both also suggest that clinicians and patients consider a surgical occlusion of the LAA at the time of other cardiac surgery in patients with AF. However, even when surgical occlusion occurs, patients who remain in AF should continue to receive anticoagulation to reduce the risk of stroke or systemic embolism. For patients who develop postoperative AF, guidelines recommend considering anticoagulant therapy but acknowledge that high-quality, prospective data are lacking for this population.

ELECTRICAL AND PHARMACOLOGICAL CARDIOVERSION

With the advent of cardioversion as the foundation for rhythm control, the need for prophylactic anticoagulation has become well established. It is presumed that the sudden mechanical recovery of atrial systole during cardioversion may cause an embolic event due to the migration of clot present in the left atrium. Furthermore, stunning of the atria increases the short-term risk for thrombi formation. Large registry data have demonstrated a 2.5-fold increased risk of thromboembolic events within 30 days following cardioversion in non-anticoagulated patients.²⁹ Although low, the incidence of thromboembolic events in adequately anticoagulated patients (0.8%–1.32%) is not negligible.^30,31 Systemic anticoagulation therapy precardioversion and postcardioversion with warfarin has been the standard of care for decades in patients with non-valvular AF of more than 48-hour duration.³² Anticoagulation should also be continued precardioversion and postcardioversion even in those with an LAA occluder device due to paucity of current data.

A series of studies have demonstrated the safety and efficacy of DOACs pericardioversion. A substudy of the RE-LY trial in 2011 demonstrated no difference in the risk of embolic events postcardioversion in patients on warfarin versus dabigatran.³³ Similar findings have been demonstrated for warfarin versus either rivaroxaban, apixaban or edoxaban therapy in post hoc analysis and in prospective studies.

Transoesophageal echocardiography (TOE) has also emerged as an important adjunct in anticoagulation management at the time of cardioversion. In a 2001 randomised study with AF of more than 48-hour duration, patients with cardioversion directed by TOE findings had similar embolic events with reduced bleeding complications compared with those who had 3 weeks of warfarin anticoagulation prior to cardioversion.³⁴ The current guidelines allow for TOE-directed cardioversion in lieu of precardioversion anticoagulation.^11,12 Of note, adequate systemic anticoagulation does not definitively exclude the presence of LAA thrombi. Up to 4% of anticoagulated patients who undergo a TOE may still have an LAA thrombus.³⁵ The benefit of TOE-guided cardioversion for higher risk patients on appropriate systemic anticoagulation is not well studied.

Controversy also remains regarding anticoagulation management in patients with acute onset (less than 48 hours duration) of AF as there is a paucity of randomized trials and limited supporting data. LAA thrombi have been shown to form even with a very short duration of AF with a reported incidence of 1.4%.³⁶ Thromboembolic events seem to primarily occur in those with an elevated baseline risk for stroke. The FinCV study in 2013 demonstrated an increased incidence (9.8%) of embolic events following cardioversion of acute-onset AF in higher risk patients (heart failure and diabetes) without systemic anticoagulation.³⁷ However, in a large national registry study of 16 274 Danish patients, first-time cardioversion without prior or subsequent anticoagulation was associated with a greater than 50% increase in risk of thromboembolic events over 1 year regardless of CHA₂DS₂-VASc score. Even those with low CHA₂DS₂-VASc benefited from pericardioversion anticoagulation.²⁹ The latest American anticoagulation guidelines for management of acute-onset AF are now primarily based on the CHA₂DS₂-VASc score, and the strength of recommendation has been downgraded given the controversy above.¹¹ Unlike the prior editions, the latest European guidelines do not differentiate based on thromboembolic risk and recommend at least 4 weeks of postcardioversion anticoagulation.^12,38 TOE-guided cardioversion in this population is likely beneficial but is yet to be well studied.

Figure 1 presents a stepwise approach to anticoagulation management for patients with AF and atrial flutter undergoing elective cardioversion. Dashed lines represent the European approach for patients with acute-onset of AF or flutter.

Figure 1.

Anticoagulation management for direct current or pharmacological cardioversion of atrial fibrillation and flutter. Dashed lines represent the European Society of Cardiology (ESC) guideline recommendations. A/C, anticoagulation; DOAC, direct oral anticoagulation; VKA, vitamin K antagonist.

ABLATION THERAPY AND ANTITHROMBOTIC MEDICATION MANAGEMENT

Stroke or TIA is relatively rare but dreaded complications following AF ablation procedures, with historical reported rates of around 0.2%–1.4%.^39–41 The majority of events appear to happen within 2 weeks following the procedure.⁴²

Periablation management of anticoagulation has evolved over the last decade as clinicians have attempted to balance the risk for thromboembolic events with procedural complications by minimising the time patients are without systemic anticoagulation. The COMPARE (2014) trial was the first randomised study to demonstrate a reduction in periprocedural thromboembolic risk in patients undergoing AF ablation without warfarin interruption. Warfarin interruption was associated with a 10-fold increased chance of stroke without any significant alteration in major bleeding complications as compared with uninterrupted warfarin therapy.⁴³ The latest guidelines currently advocate (class I) no interruption in therapeutic warfarin before ablation.⁴⁴

Subsequent studies have assessed the safety of uninterrupted DOACs for non-valvular AF ablation. The VENTURE-AF trial in 2015 was the first randomised controlled study to do so. The study demonstrated very low event rates for thromboembolism and bleeding in those randomised to uninterrupted rivaroxaban or warfarin.⁴⁵ The RE-CIRCUIT trial in 2017 and the AXAFA-AFNET 5 study in 2018 showed similar findings for dabigatran and apixaban, respectively.^46,47

The latest area of investigation is a minimally interrupted approach to anticoagulation, defined as holding only 1–2 doses prior to ablation. A large meta-analysis by Abdulhak et al in 2013 showed similar thromboembolic and bleeding risk postablation with minimally interrupted dabigatran versus uninterrupted warfarin.⁴⁸ Subsequently, the randomised ABRIDGE-J trial in 2019 found a similar thromboembolic risk but fewer bleeding complications in patients with minimally interrupted dabigatran compared with warfarin.⁴⁹ Table 3 summarises the major studies on anticoagulation management periablation.

Table 3.

Summary of randomised clinical trials for periablation anticoagulation

Randomised studies
Name	Overview	Population/design	Highlighted endpoint	Take-away
COMPARE43	Compare the safety and efficacy of uninterrupted warfarin to interrupted warfarin periablation.	▶ n=1584 patients undergoing ablation for non-valvular atrial fibrillation. ▶ Randomised 1:1 to interrupted versus uninterrupted warfarin before ablation.	▶ Incidence of stroke, transient ischaemic attacks or systemic embolism 48 hours after ablation.	▶ Reduced periprocedural risk for thromboembolic events with uninterrupted warfarin. ▶ Similar risk for major bleeding (tamponade, haematoma and retroperitoneal bleeding). ▶ Reduced risk of minor bleeding (haematoma or other bleeding not requiring intervention) with uninterrupted warfarin.
VENTURE-AF45	Compare the safety and efficacy of uninterrupted rivaroxaban to uninterrupted VKA periablation.	▶ n=248 patients undergoing ablation for non-valvular atrial fibrillation. ▶ Randomised 1:1 to uninterrupted VKA or rivaroxaban. ▶ Not powered for superiority or non-inferiority,	▶ Major bleeding (as defined by ISTH, GUSTO or TIMI criteria) within 1 month of ablation. ▶ Incidence of stroke, systemic embolism, myocardial infarction and vascular death), bleeding or procedure-related event.	▶ Similar low thromboembolic and major bleeding events in patients with uninterrupted rivaraxaban or VKA.
RE-CIRCUIT46	Compare the safety and efficacy of uninterrupted dabigatran to uninterrupted VKA periablation.	▶ n=704 patients undergoing ablation for non-valvular atrial fibrillation. ▶ Randomised 1:1 to uninterrupted dabigatran or warfarin.	▶ Major bleeding (as defined by ISTH) within 8weeks of ablation. ▶ Stroke, systemic embolism and TIA within 8 weeks of ablation.	▶ Reduced risk of major bleeding events with uninterrupted dabigatran compared with warfarin. ▶ Similar incidence of stroke and systemic embolism.
AXAFA-AFNET 548	Compare the safety and efficacy of uninterrupted apixaban to uninterrupted VKA periablation.	▶ N=767 patients undergoing first time ablation for non-valvular atrial fibrillation. ▶ Randomised 1:1 to continuous apixaban or VKA.	▶ Death, stroke or bleeding (based on BARC) at 90 days. ▶ Acute brain lesion on MRI and cognitive function.	▶ Apixaban was non-inferior to VKA for the stroke, bleeding or death at 90 days. ▶ Similar improvement in cognitive function and similar incidence of brain lesions.
ABRDIGE-J49	Compare the safety and efficacy of minimally interrupted dabigatran to uninterrupted warfarin periablation.	▶ N=504 patients undergoing ablation for non-valvular atrial fibrillation. ▶ Randomised 1:1 to minimally interrupted dabigatran or uninterrupted warfarin 4 weeks prior to procedure.	▶ Incidence of embolism and presence of LAA before ablation (on TOE or ICE). ▶ Major bleeding events (as defined by ISTH) up to 3 months postablation. ▶ Major bleeding, thromboembolic events, all-cause death up to 3 months postablation.	▶ Reduced bleeding events in minimally interrupted dabigatran compared with uninterrupted warfarin. ▶ Similar incidence of thromboembolic events.

BARC, Bleeding Academic Research Consortium; GUSTO, Global Use of Strategies to Open Occluded Coronary Arteries; ICE, intracardiac echocardiography; ISTH, International Society on Thrombosis and Haemostasis; LAA, left atrial appendage; TIMI, thrombosis in myocardial infarction; TOE, transoesophageal echocardiography; VKA, vitamin K antagonist.

Given the concern for persistent presence of LAA clot even with adequate systemic anticoagulation, TOE is routinely performed prior to catheter-based ablations, especially for those who present in AF regardless of their anticoagulation status. There is, however, limited outcome data supporting its use, and the recommendations provided in the latest guidelines are based on expert opinion only. The utility of an intracardiac echocardiography to exclude LAA clot is not yet well established but remains a possibility in expert hands.

Following ablation, regardless of the success of the procedure, there is consensus among experts to continue systemic anticoagulation for at least 2 months. Continuation is then dependent on the patient’s thromboembolic risk.⁴⁴ Figure 2 outlines a guideline-based approach for anticoagulation management before, during and after AF ablation procedures.

Figure 2.

Anticoagulation management for catheter ablation of atrial fibrillation. A/C, anticoagulation; ACT, activated clotting time; DOAC, direct oral anticoagulation; ICE, intracardiac echocardiography; TOE, transoesophageal echocardiography.

KEY OUTSTANDING RESEARCH QUESTIONS IN PREVENTION OF STROKE AND SYSTEMIC EMBOLISM

We have seen significant advances in the thromboembolic risk assessment and management of patients with AF over the last decade. The validation and appropriate incorporation of these advances in clinical care and their impact on the need for ongoing anticoagulation is of significant value and warrants future study. Simultaneously, efforts to improve the risk stratification and implementation of these risk tools into routine clinical care will help personalise therapeutic decision making.

Similar to implantable device detection of subclinical AF, the emergence of wearable heart rhythm monitoring will inevitably result in detection of AF in a significantly larger and healthier patient population. Management of these patients is not established and requires further evaluation, especially identifying a clinically relevant duration of arrhythmia necessitating treatment and cost-effective ways to mitigate the impact of large-scale false-positive detection.

In addition, the outcome of the ASAP-TOO trial (NCT02928497) and AMULET IDE (NCT02879448) studies will likely expand the indication and availability of different LAA occluding devices in the market, especially in the USA. Questions still remain regarding management of those with absolute contraindications for anticoagulation, as well as identification of those at risk for device-related thrombus formation.²⁵

Finally, protocolised management, including in dedicated anticoagulation clinics, has become the bedrock of high-quality VKA therapy. However, the introduction of DOACs challenged this treatment paradigm given their lack of frequent laboratory monitoring and dose adjustments. While some anticoagulation clinics also monitor and manage patients prescribed DOACs, this has not achieved the same broad level of utilisation as with VKA.⁵⁰ One exciting strategy for managing DOACs and other high-risk medications is a population health approach. By leveraging data already collected within the electronic medical record, dedicated nurses and/or pharmacists can rapidly assess the appropriateness of DOAC use across an entire health system population, searching for potential inappropriate indications (eg, DOAC use in patients with mechanical valves) or incorrect doses (eg, not adjusted for renal dysfunction or drug–drug interactions). Further evaluation is needed to define the effectiveness and optimal design of such dashboard tools.

CONCLUSION

In summary, stroke and systemic thromboembolism prevention remains a top priority for patients with AF and atrial flutter. While the use of the CHA₂DS₂-VASc score is widely used and guideline recommended, further efforts to improve risk prediction and stratification are needed. Anticoagulation remains first-line therapy in stroke and thromboembolism prevention, but more data are needed on the role of LAA occlusion devices in various populations, and patients undergoing cardioversion or ablation therapy are now routinely receiving uninterrupted courses of anticoagulation therapy. How new technological devices that can detect subclinical AF will impact care remains to be studied.

Table 2.

Summary of current and future left atrial appendage closure device trials and studies

Randomised Studies
Name	Device	Population/design	Anticoagulation (postdevice)	Highlighted Endpoint	Take-away
PROTECT-AF10 (2009)	WATCHMAN	▶ n=707 (non-valvular AF, tolerant of A/C). ▶ Randomised 2:1 to device or control (warfarin INR 2–3). ▶ Non-inferiority study	▶ Warfarin/ASA for 45 days postdevice. ▶ TOE-guided discontinuation of warfarin at 45 days. ASA/clopidogrel then for 6 months. ▶ Then ASA indefinitely.	▶ Stroke, systemic embolism or cardiovascular/SCD.	▶ WATCHMAN was non-inferior to warfarin for highlighted endpoint. ▶ There was increased short-term complications postimplantation.
PREVAIL20	WATCHMAN	▶ n=407 (non-valvular AF with CHADS2>2, tolerant of anticoagulation). ▶ Randomised 2:1 to device or control (warfarin INR 2–3). ▶ Non-inferiority study.	▶ Warfarin/ASA for 45 days(table 1) postdevice. ▶ TOE-guided discontinuation of warfarin at 45 days. ASA/Plavix for 6months. ▶ Then ASA indefinitely.	▶ Stroke, Systemic Embolism, or cardiovascular/SCD. ▶ Stroke or systemic embolism >7 days after randomisation.	▶ WATCHMAN had similar event rates for stroke, systemic embolism or death compared with warfarin. ▶ Non-inferiority achieved for stroke or systemic embolism 7 days after randomisation. ▶ Improved safety with implantation.
Observational and Registry Studies
Name	Device	Population/design	Anticoagulation (postdevice)	Highlighted endpoints	Results
Tzikas et al28	Amplatzer	▶ N=1047, registry, non-randomised. ▶ Retrospective analysis. ▶ Single-arm study (compared with expected cohort risk).	▶ Recommendations provided by manufacturer (but not defined by study): ASA and clopidogrel for 1–3 months and then ASA for at least an additional 3 months. ▶ Dependent on local practice patterns and patient characteristics.	▶ Stroke, TIA and systemic embolism.	▶ Lower observed thromboembolic risks compared with expected rates
ASAP26	WATCHMAN	▶ n=150 (non-valvular AF with CHADS2≥1, ineligible for anticoagulation). ▶ Prospective, non-randomised. ▶ Single-arm study (compared with expected cohort risk).	▶ Postimplantation, patients were maintained on clopidogrel or ticlopidine and ASA for 6 months and then ASA indiefintlely.	▶ Stroke, systemic embolism and cardiovascular death/SCD.	▶ Lower observed thromboembolic risks compared with expected rates.
EWOLUTION22	WATCHMAN	▶ Registry data. ▶ n=1021 (met indication for anticoagulation).	▶ Dependent on local practice pattern. Nearly 70% were not prescribed oral anticoagulation postimplantation.	▶ Procedural success, 30-day patient outcomes.	▶ Improved success rate and safety profile compared with PROTECT AF.
Currently Enrolling/Active
Name	Device	Overview	Population/design/anticoagulation (postdevice)	Highlighted endpoints	Results
ASAP-TOO (NCT02928497)	WATCHMAN	Assess the safety and efficacy of the WATCHMAN device in patients ineligible for systemic anticoagulation.	▶ n=888 (estimated; non-valvular AF with absolute contraindication for anticoagulation). ▶ Randomised 2:1 to device or control (single or no antiplatelet therapy). ▶ ASA and/or clopidogrel postimplantation.	▶ Primaiy: ischaemic stroke or systemic embolism. ▶ Secondary: all stroke, cardiovascular death/SCD or systemic embolism.
Amulet IDE (NCT02879448)	AMULET	A non-inferiority study comparing the safety and efficacy of the Amulet LAA occlude device to the WATCHMAN device.	▶ n=1878 (non-valvular AF, CHADS2>2) ▶ Randomised 1:1 to AMULET or WATCHMAN. ▶ Non-inferiority study. ▶ Manufacturer anticoagulation recommendation for WATCHMAN and AMULET.	▶ Primary: ischaemic stroke and systemic embolism. ▶ Secondary: stroke, systemic embolism and cardiovascular/SCD.
STROKECLOSE (NCT02830152)	AMULET	Assess the impact of LAA occlusion on the incidence of stroke in patients with non-valvular AF and prior haemorrhagic stroke.	▶ n=750 (non-valvular AF, CHADS2>2, recent haemorrhagic stroke (<6 months). ▶ Randomised 2:1 to AMULET or medical therapy. ▶ ASA for at least 6 months (+/− clopidogrel in first 45 days postimplantation). ▶ Anticoagulation and/or antiplatelet therapy for control group per local physicians’s discretion.	▶ Stroke, systemic embolism, life-threatening or major bleeding and all-cause mortality.

ASA - aspirin, TOE - transoesophageal echocardiogram, SCD - sudden cardiac death

ASA - aspirin, TIA - transient ischemic attack, AF - atrial fibrillation, SCD - sudden cardiac death

AF - atrial fibrillation, ASA - aspirin, SCD - sudden cardiac death, LAA - left atrial appendage

A/C, anticoagulation; AF, atrial fibrillation; ASA, aspirin; SCD, sudden cardiac death; TOE, transoesophageal echocardiography.