2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines

Contents

Top 10 Take-Home Messages

PREAMBLE

Since 1980, the American College of Cardiology (ACC) and American Heart Association (AHA) have translated scientific evidence into clinical practice guidelines with recommendations to improve cardiovascular health. These guidelines, which are based on systematic methods to evaluate and classify evidence, provide a foundation for the delivery of quality cardiovascular care. The ACC and AHA sponsor the development and publication of clinical practice guidelines without commercial support, and members volunteer their time to the writing and review efforts. Guidelines are the official policy of the ACC and AHA. For some guidelines, the ACC and AHA collaborate with other organizations.

Intended Use

Clinical practice guidelines provide recommendations applicable to patients with or at risk of developing cardiovascular disease. The focus is on medical practice in the United States, but these guidelines are relevant to patients throughout the world. Although guidelines may be used to inform regulatory or payer decisions, the intent is to improve quality of care and align with patients’ interests. Guidelines are intended to define practices meeting the needs of patients in most, but not all, circumstances and should not replace clinical judgment.

Clinical Implementation

Management, in accordance with guideline recommendations, is effective only when followed by both practitioners and patients. Adherence to recommendations can be enhanced by shared decision-making between clinicians and patients, with patient engagement in selecting interventions on the basis of individual values, preferences, and associated conditions and comorbidities.

Methodology and Modernization

The ACC/AHA Joint Committee on Clinical Practice Guidelines (Joint Committee) continuously reviews, updates, and modifies guideline methodology on the basis of published standards from organizations, including the Institute of Medicine,^1,2 and on the basis of internal reevaluation. Similarly, presentation and delivery of guidelines are reevaluated and modified in response to evolving technologies and other factors to optimally facilitate dissemination of information to health care professionals at the point of care.

Numerous modifications to the guidelines have been implemented to make them shorter and enhance “user friendliness.” Guidelines are written and presented in a modular, “knowledge chunk” format, in which each chunk includes a table of recommendations, a brief synopsis, recommendation-specific supportive text and, when appropriate, flow diagrams or additional tables. Hyperlinked references are provided for each modular knowledge chunk to facilitate quick access and review.

In recognition of the importance of cost–value considerations, in certain guidelines, when appropriate and feasible, an analysis of value for a drug, device, or intervention may be performed in accordance with the ACC/AHA methodology.³

To ensure that guideline recommendations remain current, new data will be reviewed on an ongoing basis by the writing committee and staff. Going forward, targeted sections/knowledge chunks will be revised dynamically after publication and timely peer review of potentially practice-changing science. The previous designations of “full revision” and “focused update” will be phased out. For additional information and policies on guideline development, readers may consult the ACC/AHA guideline methodology manual⁴ and other methodology articles.^5–7

Selection of Writing Committee Members

The Joint Committee strives to ensure that the guideline writing committee contains requisite content expertise and is representative of the broader cardiovascular community by selection of experts across a spectrum of backgrounds, representing different geographic regions, sexes, races, ethnicities, intellectual perspectives/biases, and clinical practice settings. Organizations and professional societies with related interests and expertise are invited to participate as partners or collaborators.

Relationships With Industry and Other Entities

The ACC and AHA have rigorous policies and methods to ensure that documents are developed without bias or improper influence. The complete policy on relationships with industry and other entities (RWI) can be found online. Appendix 1 of the guideline lists writing committee members’ comprehensive and relevant RWI; for the purposes of full transparency, comprehensive and relevant disclosure information for the Joint Committee is also available online.

Evidence Review and Evidence Review Committees

In developing recommendations, the writing committee uses evidence-based methodologies that are based on all available data.^4,5 Literature searches focus on randomized controlled trials (RCTs) but also include registries, nonrandomized comparative and descriptive studies, case series, cohort studies, systematic reviews, and expert opinion. Only key references are cited.

An independent evidence review committee is commissioned when there are ≥1 questions deemed of utmost clinical importance and merit formal systematic review to determine which patients are most likely to benefit from a drug, device, or treatment strategy, and to what degree. Criteria for commissioning an evidence review committee and formal systematic review include absence of a current authoritative systematic review, feasibility of defining the benefit and risk in a time frame consistent with the writing of a guideline, relevance to a substantial number of patients, and likelihood that the findings can be translated into actionable recommendations. Evidence review committee members may include methodologists, epidemiologists, clinicians, and biostatisticians. Recommendations developed by the writing committee on the basis of the systematic review are marked “^SR”.

Guideline-Directed Management and Therapy

The term guideline-directed management and therapy (GDMT) encompasses clinical evaluation, diagnostic testing, and both pharmacological and procedural treatments. For these and all recommended drug treatment regimens, the reader should confirm dosage with product insert material and evaluate for contraindications and interactions. Recommendations are limited to drugs, devices, and treatments approved for clinical use in the United States.

Joshua A. Beckman, MD, MS, FAHA, FACC

Chair, ACC/AHA Joint Committee on Clinical Practice Guidelines

1. Introduction

1.1. Methodology and Evidence Review

The recommendations listed in this guideline are, whenever possible, evidence based. An initial extensive evidence review, which included literature derived from research involving human subjects, published in English, and indexed in MEDLINE (through PubMed), EMBASE, the Cochrane Library, the Agency for Healthcare Research and Quality, and other selected databases relevant to this guideline, was conducted from May 2022 to November 2022. Key search words included but were not limited to the following: atrial fibrillation; pregnancy; heart defects, congenital; smoking; cardiomyopathy, hypertrophic; alcohol; caffeine; sleep; diet; fitness; obesity; anticoagulants; diabetes; rhythm monitoring; heart failure; genetics; heart valve diseases; rate control; catheter ablation; social determinants of health; chronic kidney disease; sinus rhythm; chronic coronary syndromes; left atrial appendage occlusion; left atrial appendage exclusion; cardiac surgical procedures; amiodarone; electrical cardioversion; thromboembolism; rhythm control; Wolff-Parkinson-White syndrome. Additional relevant studies, published through November 2022 during the guideline writing process, were also considered by the writing committee and added to the evidence tables when appropriate. The final evidence tables are included in the Online Data Supplement and summarize the evidence used by the writing committee to formulate recommendations. References selected and published in the present document are representative and not all-inclusive.

1.2. Organization of the Writing Committee

The writing committee consisted of cardiologists, cardiac electrophysiologists, surgeons, pharmacists, and patient representatives/lay stakeholders. The writing committee included representatives from the ACC and AHA, ACCP, and HRS. Appendix 1 of the current document lists writing committee members’ comprehensive and relevant RWI.

1.3. Document Review and Approval

The Joint Committee appointed a peer review committee to review the document. The peer review committee comprised individuals nominated by the ACC, AHA, and the collaborating organizations. Reviewers’ RWI information was distributed to the writing committee and is published in this document (Appendix 2).

This document was approved for publication by the governing bodies of the ACC and the AHA and was endorsed by ACCP and HRS.

1.4. Scope of the Guideline

In developing the “2023 ACC/AHA/ACCP/HRS Guideline for the Diagnosis and Management of Atrial Fibrillation” (2023 atrial fibrillation guideline), the writing committee reviewed previously published guidelines. Table 1 contains a list of these publications deemed pertinent to this writing effort and is intended for use as a resource, thus obviating the need to repeat existing guideline recommendations.

1.5. Class of Recommendations and Level of Evidence

The Class of Recommendation (COR) indicates the strength of recommendation and encompasses the estimated magnitude and certainty of benefit in proportion to risk. The Level of Evidence (LOE) rates the quality of scientific evidence supporting the intervention on the basis of the type, quantity, and consistency of data from clinical trials and other sources (Table 2).

Table 2. Applying American College of Cardiology/American Heart Association Class of Recommendation and Level of Evidence to Clinical Strategies, Interventions, Treatments, or Diagnostic Testing in Patient Care* (Updated May 2019)

1.6. Abbreviations

2. BACKGROUND AND PATHOPHYSIOLOGY

2.1. Epidemiology

Atrial fibrillation (AF) is the most sustained common arrhythmia, and its incidence and prevalence are increasing in the United States and globally (Figures 1 to 3).^1,2 The increasing burden is multifactorial; causes include the aging of the population, rising tide of obesity, increasing detection, and increasing survival with AF and other forms of cardiovascular disease (CVD). The estimated global prevalence was 50 million in 2020.^2,3 Although the prevalence of undiagnosed AF in the community is unknown, using back-calculation methodology, investigators have estimated that, in 2015, about 11% (591 000 cases) of the >5.6 million AF cases in the United States were undiagnosed.⁴

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

AF is associated with higher health care utilization and costs.² Using US data from Optum (an administrative claims database for commercially insured [United Healthcare] patients in the United States), compared with patients without AF, patients with incident AF had an increased risk of inpatient visits and more cardiovascular-related emergency department visits (relative risk [RR], 2.41 [95% CI, 2.35–2.47]).⁵ AF is costly. Examining Optum data, individuals with AF have annual health care costs of $63 031, which is $27 896 more than individuals without AF.⁵ Investigators examining public and private health insurer data estimated that in US dollars in 2016, AF accounted for $28.4 billion (95% CI, $24.6 billion–$33.8 billion) in health care spending.⁶

2.1.1. Prevalence, Incidence, Morbidity, and Mortality

AF prevalence in the United States was estimated to be 5.2 million in 2010, with an expectation to rise to 12.1 million in 2030.¹ Corresponding estimates for US incidence was 1.2 million cases in 2010, with an expectation to rise to 2.6 million cases in 2030.¹ The rate of AF diagnosis varies by education, income,² clinical,^3,4 and genetic³ factors. Overall lifetime risk is about 30% to 40% in White individuals,^2–4 about 20% in African American individuals,² and about 15% in Chinese⁵ individuals.

AF is associated with a 1.5- to 2-fold increased risk of death^6,7; studies suggest that the mortality risk may be higher in women than in men.⁶ In meta-analyses, AF is also associated with increased risk of multiple adverse outcomes, including a 2.4-fold risk of stroke,⁷ 1.5-fold risk of cognitive impairment or dementia,⁸ 1.5-fold risk of myocardial infarction (MI),⁹ 2-fold risk of sudden cardiac death,¹⁰ 5-fold risk of heart failure (HF),⁷ 1.6-fold risk of chronic kidney disease (CKD),⁷ and 1.3-fold risk of peripheral artery disease (PAD).⁷ In Medicare beneficiaries, the most frequent outcome in the 5 years after AF diagnosis was death (19.5% at 1 year; 48.8% at 5 years)¹¹; the next most common diagnosis was HF (13.7%), followed by new-onset stroke (7.1%), gastrointestinal hemorrhage (5.7%), and MI (3.9%).¹¹

2.1.2. Risk Factors and Associated Heart Disease

In Table 3, we present the evidence for the most widely reported and validated factors for AF from single studies, meta-analyses, and Mendelian randomization studies. Risk factors include demographic, anthropometric, and cardiovascular risk factors, CVD, noncardiac conditions, biomarkers (eg, electrocardiographic, imaging, circulating), and genetic markers.¹ Models predicting risk of AF onset are presented in Section 4.1 (“Risk Stratification and Population Screening”). Most studies of AF risk factors and outcomes have been reported from high-income countries and in individuals of European ancestry.

2.2. Atrial Arrhythmia Classification and Definitions

2.2.1. AF Classification

The previous classification of AF, which was based only on arrhythmia duration, although useful, tended to emphasize AF once it was diagnosed and focused mainly on therapeutic interventions. The new proposed classification using stages aims to correct the deficiencies of the previous classification by recognizing AF as a progressive disease that requires different strategies at the different stages, from prevention to screening, to rate and rhythm control therapies. The different stages better define AF as a progressive disease and highlight the need to address it at the earliest stages, especially emphasizing the importance of prevention, risk factor management, and timing for screening in those patients at the highest risk. The stages are not mutually exclusive (eg, risk factors should be managed through multiple stages) (Figure 4).

• Download figure
• Download PowerPoint

AF is the most common arrhythmia in the world and accounts for significant morbidity and mortality. Over the past decade, evidence has consistently shown that the best treatment of atrial fibrillation requires multiple stakeholders committed to providing comprehensive patient-centered care. In addition, as emphasized in this guideline, AF should be thought of in a more holistic sense over an individual patient’s lifetime.

The foundation of optimal AF management is the treatment of risk factors and implementing lifestyle changes to decrease the likelihood of developing AF (Figure 5). Once AF develops, patient care should focus on assessing the risk of stroke and implementing any necessary treatment, continued optimization of all modifiable risk factors, and managing potential symptoms of AF, with an initial focus on evaluating and minimizing AF burden. However, as outlined in this guideline, access to all aspects of health care to all patients is necessary for any true improvement to be realized.

• Download figure
• Download PowerPoint

When AF develops, holistic and optimal care of the patient at risk for AF, or who has developed AF, can be simply modeled using a building. The foundation of care is treatment of comorbidities and risk factors and implementing behavioral change in all individuals to decrease the likelihood of developing AF and reducing its burden (Screening for all risk factors from HEAD 2 TOES). Once AF develops, there are 3 important care processes that must be specifically addressed with all patients and aligned with their goals of therapy: Stroke risk assessment and treatment, if appropriate, Optimizing all modifiable risk factors, and Symptom management using rate- and rhythm-control strategies that consider AF burden in the context of an individual patient’s needs (SOS). The overarching principle for AF management is Access to All Aspects of care to All (4 As).

2.2.2. Associated Arrhythmias

Other atrial arrhythmias are often encountered in patients with AF.

Atrial Tachycardia (AT): It is a regular atrial rhythm at a constant rate of >100 beats per minute (bpm) with discrete P waves and atrial activation sequences originating outside of the sinus node.¹ The mechanism can be automaticity, triggered activity, or a microreentry circuit. Focal ATs arise from a single discrete site within the left or right atrium, in contrast to macroreentrant atrial arrhythmias and AF, which involve multiple sites or larger circuits. In multifocal AT, the atrial activation sequence and P-wave morphology vary.

Atrial Flutter (AFL) and Macroreentrant AT: They occur in many of the same situations as AF. Typical AFL, also known as “typical AFL” or “cavotricuspid isthmus (CTI)-dependent AFL,”² involves a macroreentrant circuit around the tricuspid annulus traversing the CTI on the right side of the heart (Figure 6). This is the arrhythmia associated with the classic electrocardiogram (ECG) finding of sawtooth flutter waves in the inferior leads when the circuit goes in the counterclockwise direction. The same circuit in the clockwise direction is called “reverse typical AFL.” If the flutter involves a different circuit than tricuspid valve/isthmus, then it is called “atypical” AFL, which is also known as “noncavotricuspid isthmus–dependent macroreentrant AT.”² AFL was previously classified as either type I or type II. That terminology is no longer used.

• Download figure
• Download PowerPoint

2.3. Mechanisms and Pathophysiology

AF is a chaotic, rapid (300-500 bpm), and irregular atrial rhythm. Although normal rhythms are conducted through the atria in smooth waves initiated by the sinoatrial node, AF is the result of either electrophysiological abnormalities that underlie impulse generation and/or structural abnormalities of cellular connections that typically facilitate rapid and uniform impulse conduction. AF often stems from waves of electrical activity originating from ectopic action potentials most commonly generated in the pulmonary veins (PVs) of the left atrium (LA),¹ or in response to reentrant activity promoted by heterogeneous conduction due to interstitial fibrosis.² Atrial ectopy can generate runs of tachycardia, while persistent AF requires a substrate that is either sufficiently large or conduction that is sufficiently heterogeneous to enable reentrant activity to persist. The electrical abnormalities evident on the ECG during AF likely represent a shared phenotype of a condition with many distinct etiologies (genetic, environmental, and metabolic) (Figure 7).

• Download figure
• Download PowerPoint

2.3.1. Electrophysiological Mechanisms and Electrical Remodeling

The observation by Haissaguerre et al¹ that ectopic firing from PVs triggers AF revolutionized treatments for AF. PV features that increase vulnerability for initiating ectopy include a higher resting membrane potential, stretch-activated channels,² and pattern of cross myofiber orientation.

Electrical remodeling, which can contribute to and result from AF, includes perturbations that culminate in abnormal Ca²⁺ handling and shortened, proarrhythmic triangulated action potentials (eg, from decreased L-type Ca²⁺ current³ and increase in I_K1).⁴ Heterogeneity in I_K1 between left and right atria further promotes arrhythmogenicity.⁵ Downregulation of connexin results in decreased gap junctions, leading to slow heterogeneous atrial conduction velocity and repolarization,⁶ promoting regional functional conduction block that can support reentry. Connexin remodeling can be due to genetic⁷ or acquired factors, such as inflammatory state,^8,9 older age, or sleep-disordered breathing (SDB).¹⁰

Calcium mishandling from remodeling increases calcium load in the sarcoplasmic reticulum and dysfunction of ryanodine receptors, which regulate intracellular calcium release. Remodeling in AF of sarcoplasmic reticulum calcium ATPase underlies sequestration of intracellular calcium between beats, and/or altered regulation of sarcoplasmic reticulum calcium ATPase by phospholamban and sarcolipin.¹¹ Increased intracellular sodium from calcium-calmodulin II-induced increased late sodium current or cardiac glycosides can also increase sarcoplasmic reticulum calcium through the sodium-calcium exchanger.^12,13 Several upstream mechanisms (eg, oxidative stress, inflammatory signaling) promote calcium-calmodulin II -activation. CaMKII-mediated and hyperphosphorylation of ryanodine receptor 2 promotes spontaneous diastolic Ca²⁺ leak by increasing ryanodine receptor 2 channel open probability, leading to higher intracellular Ca²⁺ levels and the milieu for delayed afterdepolarizations, the most likely trigger for AF initiation.¹⁴ Electrogenic action of sodium-calcium exchanger drives afterdepolarizations, and expression of sodium-calcium exchanger is increased in HF and AF. Action potential alternans related to Ca²⁺ mishandling can be noted to precede AF onset and increases with age.

2.3.1.1. Triggers of AF

The atria of patients with AF tend to have both shorter effective refractory periods and slower conduction, which enhances dispersion of repolarization and favor reentry. This substrate is sensitive to AF initiation, frequently after premature atrial contractions (PACs).¹ PAC burden is associated with development of AF.² Larger LA volume, increased NT-proBNP levels, and impaired LA emptying are associated with increased PAC burden.³ In a study of 100 patients undergoing PV isolation, PACs induced 41 episodes of AF in 22 patients, with most episodes originating in the PVs.⁴ Earlier studies also reported an LA gradient of background potassium currents, resulting in shorter LA than right atrial effective refractory period in patients with AF.⁵ Mapping studies of AF electrograms in canine models documented a left-right gradient of high frequency sources (drivers), with the highest frequency regions located near the PVs in the LA.⁶

2.3.2. Atrial Structural Abnormalities, Remodeling, and Atrial Myopathy

Atrial cardiomyopathy has been identified as “any complex of structural, architectural, contractile or electrophysiological changes affecting the atria with the potential to produce clinically relevant manifestations.”¹ Atrial cardiomyopathy is common, associated with aging and other comorbidities with metabolic or hemodynamic stress, and often leads to or results from AF. Structural and electrical remodeling promote AF and include interstitial changes, increased myofibroblast activity, and collagen deposition, fibrofatty deposits, altered ion channel expression, calcium signaling and contractility, and inflammatory infiltrates.¹ Myeloperoxidase, an oxidase in neutrophils and macrophages, is associated with fibroblast activation, interstitial fibrosis, and inducibility of AF in a mouse model.² Prothrombotic changes are often evident in the LA, including increased endocardial expression of von Willebrand factor, vascular cell adhesion molecule-1 and monocyte chemoattractant protein-1 changes, which may increase risk of thrombus formation and stroke.¹

Experimental hemodynamic overload, such as reversible aortopulmonary artery shunting promotes atrial dilation, atrial myocyte hypertrophy, interstitial fibrosis, alterations in extracellular matrix composition and vascular dysfunction, and increased vulnerability to atrial ectopy and AF inducibility.³ Electrical remodeling similar to that seen during AF is observed, with characteristic loss of calcium currents. Shunt closure reverses electrophysiological changes.³ Clinically, the hemodynamic structural changes are difficult or impossible to reverse.

2.3.2.1. Upstream Pathways

Upstream pathways include inflammatory, oxidative stress, fibrosis, calcium handling, genetic, metabolic, obesity, and other mechanisms implicated in increasing susceptibility to or progression of AF. The renin-angiotensin-aldosterone system (RAAS), oxidative stress, inflammatory signaling, and calcium overload are discussed here and in Section 2.3.1 (“Electrophysiological Mechanisms and Electrical Remodeling”). Upstream therapies are discussed in Section 8.3.4 (“Upstream Therapy”).

The RAAS regulates blood pressure (BP) and is activated in hypertension and obesity. RAAS activation promotes vascular smooth muscle constriction (increasing BP), activates fibroblasts (increasing atrial interstitial collagen), and increases reactive oxygen species in the sympathetic nervous system. BP and weight control, including treatment with angiotensin-converting enzyme (ACE) inhibitors or angiotensin-II receptor blockers, attenuate pathologic changes.

Oxidative stress is associated with: (1) activation of calcium-dependent calmodulin kinase II^1,2 and jun kinase 2,³ which promotes ryanodine receptor phosphorylation and leads to spontaneous calcium release⁴; (2) increased late sodium current, I_Na,L, and lower calcium current amplitude; (3) formation of reactive lipid products (isolevuglandins) in hypertension⁵ and AF⁶; (4) activation of the redox-sensitive transcription factor nuclear factor-kappa B and the NLRP3 inflammasome⁷; (5) increase in mitochondrial and metabolic stress; (6) inflammatory changes; and (7) myofilament protein degradation, impairing atrial contractility, which may increase thromboembolism risk.⁸

Systemic inflammatory activation, first documented with AF after cardiac surgery,⁹ has also been associated with nonsurgical AF.¹⁰ NLRP3 knockdown prevented AF inducibility in a mouse model, implicating the NLRP3 inflammasome in AF pathophysiology.¹¹ NLRP3-blocking drugs are in development.

2.3.2.2. Persistence of AF

In general, persistence of AF reflects the substrate for AF. A suitable substrate for AF has a wavelength (wavelength = refractory period × conduction velocity) that is shorter than the dimensions of the tissue, with heterogenous conduction velocity and/or repolarization duration. Thus, an individual with large, fibrotic, and/or fatty atria is more likely to have persistent AF than one with a normal-sized atria with little interstitial fibrosis or adipose infiltration. Electrophysiologic remodeling is typically a response to persistent AF rather than the trigger.

2.3.3. Role of the Autonomic Nervous System

The autonomic nervous system (ANS) has an important role as trigger and substrate (Figure 8).

• Download figure
• Download PowerPoint

ANS triggers AF

The ANS as AF trigger is detailed in several reviews.^1–6 Sympathetic efferent stimulation releases noradrenaline, stimulating G-coupled β-adrenergic receptors, enhancing L-type calcium channels, and increasing inward current (automaticity/early afterdepolarization). Delayed afterdepolarization occurs via calcium overload and ryanodine-2 receptor dysfunction. Parasympathetic stimulation shortens atrial effective refractory period by increasing I_KACh (acetyl-choline receptor mediated inward rectifying potassium channel) activity. Atrial effective refractory period heterogeneity follows the pattern of autonomic innervation. Sympathetic and parasympathetic activity, alone or combined, can trigger AF.

ANS maintains AF

Atrial sympathetic and parasympathetic hyperinnervation and spatial heterogeneity, coupled with electrical fractionation and altered atrial electrophysiology, contribute to substrate.^7–10 Modifiable AF risk factors promote ANS dysfunction.³ AF produces autonomic afferent reflex deficiencies elicited by decreased cardiac volume^11,12 and increases sympathetic activity.¹³ Afferent abnormalities disrupt blood volume and pressure homeostasis. Similar abnormalities to AF were identified in HF.¹⁴ Afferent ANS dysfunction could link autonomic with anatomic remodeling (atrial dilatation¹⁵), contributing to AF self-perpetuation.

2.4. Genetics

Both common forms and familial AF are heritable.^1–3 Multiple recent genome-wide association studies have documented >100 loci specific for AF.^4,5 Numerous AF loci appear consistent across multiethnic groups,⁵ with some population variation.^6,7 With large genome-wide genotyped cohorts such as the UK Biobank and the US National Heart, Lung, and Blood Institute’s (NHLBI’s) Trans Omics for Precision Medicine, the genetic architecture of AF is now emerging. A UK Biobank study identified TTN loss of function variants in 0.44% of participants, 14% of whom had AF.⁸ In a Trans Omics for Precision Medicine study of nearly 1300 participants <66 years of age with AF, 10.1% harbored a disease-associated genetic variant in genes associated with inherited cardiomyopathy or arrhythmia syndromes (most common were TTN, MYH7, MYH6, LMNA, and KCNQ1), and 62.8% had variants of undetermined significance. Disease-associated variants were more prevalent at younger age of AF onset, 16.8% in those <30 years.⁹ A smaller study of persons of Hispanic or African American descent reported 7% of persons with AF onset at ≤66 years of age harbored rare likely pathogenic or pathogenic sequence variants, mostly in myocardial structural proteins and ion channels.¹⁰

2.5. Addressing Health Inequities and Barriers to AF Management

Synopsis

Inequities in AF care and outcomes in individuals who are women, from underrepresented racial and ethnic groups (UREGs),² or who have adverse SDOH have been documented.^3,4 Sex differences in AF treatment have been described with respect to anticoagulation^6,7 and rhythm control therapy approaches.^7–9 Racial and ethnic differences in clinical presentation, management, and prognosis, including stroke, HF, and death, in patients with AF are widely reported.^2,10

To avoid guidelines having the unintended consequence of widening inequities in clinical care and outcomes in individuals with AF, it is essential to longitudinally measure the receipt of AF GDMT and outcomes at the clinical practice and health system levels stratified by specific populations who have historically experienced inequitable care. If inequities are identified, barriers to GDMT should be eliminated. Data are needed to assess the impact of addressing SDOH in patients with AF on process measures, health care utilization, costs, and clinical outcomes.¹¹ In other health contexts, there are observational and randomized data¹² that screening and addressing SDOH leads to improved medication adherence,¹³ risk factor control,¹⁴ and clinical outcomes.¹⁵

Recommendation-Specific Supportive Text

1. Despite the elevated risk of stroke in women and several UREGs, many are less likely to be treated with stroke risk reduction therapies.^6,7 Although women and individuals from UREGs with AF are more symptomatic and report worse QOL than their counterparts, they also are less likely to be referred to an electrophysiologist⁷ and receive catheter ablation.^2,7–9,16 Women are referred for ablation later in the disease course and at older ages than men.^7,17,18 These differences or delays in therapy may result in worse outcome given early rhythm control of AF improves outcomes^1,19 in select patients. In addition, in the Catheter Ablation versus Antiarrhythmic Drugs in AF trial, individuals from UREGs treated with catheter ablation had a 72% relative reduction in the all-cause mortality rate.²⁰ Therefore, ensuring timely and equitable referral of women, individuals from UREG, and those with adverse SDOH for rhythm control therapy is important. In patients with AF, indicators of lower socioeconomic status were associated with lower oral anticoagulation rates,²¹ lower rate of adherence during direct oral anticoagulant (DOAC) initiation,²² specialty care,^21,23 and less use of cardioversion^4,21 and catheter ablation.^4,21,24 Indicators of socioeconomic disadvantages, such as increased risk of hospitalization,²⁵ stroke,^4,21 HF,^4,21,26 and death, were also associated with complications in patients with AF.^4,21,27

3. SHARED DECISION-MAKING (SDM) IN AF MANAGEMENT

Synopsis

There are wide variations in how SDM is implemented in clinical care settings.^5,6 Decision aids may provide standardization of SDM approaches for better informing patients about stroke reduction therapies and improve patient-reported measures but to date have not consistently been developed with recommended frameworks, have rarely been tested in systemically disadvantaged populations (low health literacy, UREGs, low socioeconomic status, rural geography, older adults), and have had variable impact on adherence and clinical outcomes.^1–8 Ongoing work will measure health and digital literacy and strengthen the evidence for the impact of decision aids on decisional quality, adherence to treatment, and health outcomes.⁹

Symptom severity strongly correlates with QOL; thus, minimizing symptoms is an essential component of patient-centered AF management decisions. Rhythm control strategies improve QOL, particularly when maintenance of sinus rhythm or low AF burden is achieved.¹⁰ Notably, few SDM decision aids are focused on rate or rhythm control treatment options, and few have measured QOL as an outcome.^3,5

Recommendation-Specific Supportive Text

1. Recently, 2 comprehensive reviews of decision aids for stroke reduction therapies were conducted to determine the impact of these tools on patient-reported measures of decisional quality, while considering other important outcomes including oral anticoagulant (OAC) uptake, medication adherence, and the effect on bleeding and stroke.^3,4 Most decision aids focused on patient- reported measures, and few underwent rigorous pilot testing or correlated the aid with clinical outcomes, such as stroke and bleeding. Decision aids consistently demonstrate improvements in patient knowledge. The pooled analysis by Song et al noted lower decisional conflict using decision aids and enhanced OAC uptake (risk ratio, 1.03 [95% CI, 1.01-1.05]).⁴ Decision aids have historically shown marginal improvement in 3-month measures of adherence, and the 2 largest randomized trials to date showed no improvement in adherence between decision aids and usual care at 1 year.^1,2 There is a paucity of data on the impact of decision aids on stroke, thromboembolic events, or bleeding, and when assessed the benefit has been minimal or neutral.^4,11 Despite the US Centers for Medicaid & Medicare Services coverage decision requirement for SDM for percutaneous left atrial appendage occlusion (LAAO), only 1 tool was identified that incorporated this option (Table 5).⁵

4. CLINICAL EVALUATION

4.1. Risk Stratification and Population Screening

There are >20 risk prediction models for incident AF in the community.¹ The most widely replicated risk prediction model for newly diagnosed AF is CHARGE-AF (Cohorts for Heart and Aging Research in Genomic Epidemiology model for atrial fibrillation; Table 6),² while the C₂HEST score (coronary artery disease or chronic obstructive pulmonary disease [1 point each]; hypertension [1 point]; elderly [age ≥75 years, 2 points]; systolic HF [2 points]; thyroid disease [hyperthyroidism, 1 point]) was derived and validated in Asian cohorts (Table 7).^3,4

Screening for AF has been investigated, mostly in patients >65 years of age, using various protocols that include both 1-time electrocardiographic recordings, recurring intermittent ECGs (including consumer-based devices), or continuous electrocardiographic external monitors. Most screening trials have shown higher AF detection using intermittent or continuous electrocardiographic recordings⁵ and higher AF detection in patients with higher predicted risk for AF.⁶ A recent study also showed that an AI algorithm able to risk-stratify a relatively uniform population (eg, older adults at risk for stroke) to detect undiagnosed AF during short-term cardiac monitoring was associated with increased AF.⁷ Conversely, mass population screening with a smartwatch app only rarely detected a new diagnosis of AF.⁸ Ultimately, for risk stratification models and screening programs to be useful, they would need to improve outcomes and be cost-effective.⁹ It is not yet established that patients at high risk of developing AF by a validated risk score benefit from screening and interventions to improve rates of ischemic stroke, systemic embolism, and survival.

4.2. Basic Evaluation

4.2.1. Basic Clinical Evaluation

Synopsis

The initial clinical evaluation of the patient with newly diagnosed or suspected AF should be focused on confirming the diagnosis and identifying relevant clinical factors that will impact management. A targeted history and physical examination should be performed at the initial assessment and repeated during periodic follow-up, especially given the evolving risk of thromboembolism and the cadence of symptoms in response to therapy (see Section 11, “Future Research Needs”). An ECG can assess other electrical abnormalities, including possible substrates such as Wolff-Parkinson-White (WPW) syndrome, coexisting atrial arrhythmias, as well as abnormalities that may affect decision-making in pharmacological management (eg, bradycardia, QT duration). Basic laboratory tests should be performed to determine if other clinically relevant disorders are present and impact on management, particularly with respect to stroke and bleeding risk. A transthoracic echocardiogram provides information on chamber size, thickness, function, and the presence of valvular pathology. Additional testing, including multimodality advanced imaging and further ambulatory electrocardiographic monitoring, may be pursued based on the results of these initial evaluations. AF itself does not increase the likelihood of myocardial ischemia, ACS or PE, and therefore routine testing for these disorders in the absence of signs or symptoms is of no benefit.

Recommendation-Specific Supportive Text

1. A transthoracic echocardiogram is essential to evaluate chamber size and function, valve function, and right ventricular (RV) pressure. Left ventricular ejection fraction (LVEF) impacts decisions for antiarrhythmic drug therapy and whether to prioritize other rhythm control therapies, including catheter ablation. Additionally, strain imaging may suggest an underlying infiltrative cardiomyopathy, such as amyloidosis.¹ Echocardiography also provides information on LA size and function. Altered LA compliance is known to be associated with AF¹¹ and progression toward persistent-type AF.² In a meta-analysis, AF recurrence after ablation was associated with a lower LA strain,³ while LA volume was a stronger predictor of recurrence after ablation than the characterization of AF as paroxysmal or persistent.⁴ Laboratory testing can detect other medical conditions that are associated with AF and would impact therapeutic decision-making, such as CKD,⁵ liver dysfunction,⁶ and hyperthyroidism.^7,12 Laboratory testing may also reveal electrolyte abnormalities, including from medications such as diuretics. Laboratory testing is also needed to determine stroke risk and bleeding risk factors, which will guide management decisions. When clinical suspicion exists, additional testing might be needed to evaluate for potentially related conditions, such as significant valvular disease.

2. The presence of AF itself should not prompt routine protocolized testing for myocardial ischemia, ACS, or PE, in the absence of signs or symptoms to suggest those diseases. A retrospective analysis of asymptomatic patients with AF compared with age- and sex-matched controls that were referred for myocardial stress imaging found no difference in mean summed stress score or rate of abnormal studies.⁸ A retrospective analysis of 1700 asymptomatic AF patients (no chest pain or dyspnea) found that 4.6% had >5% ischemic myocardium, and the yield to detect ischemia that resulted in revascularization was only 0.4%.⁹ Among patients suspected of PE, a retrospective analysis showed that the presence of AF did not increase the probability of PE.¹⁰ Certainly, this would not preclude from evaluating patients with signs and/or symptoms of ischemia and PE.

4.2.2. Rhythm Monitoring Tools and Methods

Synopsis

Monitoring options for AF include a standard 12-lead ECG, continuously recording or loop-recording electrocardiographic monitors using separate electrodes or as patches, implantable cardiac monitors (sometimes referred to as implantable loop recorders), cardiac rhythm management devices with an atrial lead (eg, pacemakers and defibrillators), handheld ECGs, and smartwatches. Photoplethysmography has been used to infer AF from irregular pulse patterns using a variety of devices, predominately smartphone cameras¹⁴ and smartwatches.^15–18 Electrocardiographic monitors often deploy automated algorithms for AF detection, but due to variable accuracy,^1–5 the initial diagnosis should rely on a health care professional’s examination of the electrocardiographic tracing. Although photoplethysmography monitors may alert individuals to obtain an electrocardiographic tracing, it is not sufficiently reliable to establish an AF diagnosis.^14–18 AF detected from an atrial lead has been validated versus surface ECGs^6,7 and independently predicts stroke.^8,9 RCTs have demonstrated that implantable cardiac monitors exhibit the highest sensitivity in detecting AF compared with external ambulatory monitors, likely related to the longer duration of monitoring.^10,11 Automated algorithms to analyze electrocardiographic devices have generally been found to be sufficiently reliable to infer the frequency, duration, and burden of AF among those with an AF diagnosis.^1,4,5,12 A randomized trial showed that a handheld electrocardiographic monitor resulted in earlier detection of recurrent AF.¹³

Recommendation-Specific Supportive Text

1. Although automated algorithms in various devices are generally reliable, health care professional overread of electrocardiographic tracings remains necessary given the imperfect test characteristics of those algorithms.^1–5 Similarly, while algorithms utilizing photoplethysmography signals (derived using smartphones or smartwatches) to infer irregular heart rates can discriminate AF from normal sinus rhythm, these are not sufficiently reliable to establish an AF diagnosis.^14–18

2. Cardiac rhythm devices with an atrial lead have been shown to detect AF validated against conventional surface electrocardiographic tracings.^6–8 In addition, both the presence and duration of AF detected solely by these devices predict stroke in a manner that would be expected of AF.^8,9 It is still essential that the intracardiac tracings are reviewed for confirmation because false-positives are possible. The duration of AF that mandates intervention with anticoagulation will be discussed in Section 6.4.1 (“Oral Anticoagulation for Device-Detected Atrial High-Rate Episodes Among Patients Without a Previous Diagnosis of AF”).

3. The more frequent and longer monitoring for AF is deployed, the greater the sensitivity in detecting AF.^10,19 Randomized trials, predominately among cryptogenic stroke patients, have revealed that implantable cardiac monitors exhibit the highest sensitivity in detecting AF in view of extended monitoring periods compared with external monitors.^10,11

4. It is often not feasible to manually review all electrocardiographic strips from various monitoring devices, either due to inaccessibility or time and resource constraints on health care professionals. Although variability in accuracy across different devices may be present, the validity demonstrated in automated algorithms is generally sufficient to infer frequency, duration, and burden of AF using electrocardiographic devices such as continuously wearable monitors, implantable cardiac monitors, and cardiac rhythm devices with an atrial lead.^1,4,5,12

5. Cardiac monitoring may be advised to AF patients for various reasons, such as for detecting recurrences, screening, or response to therapy. Among patients with AF who are undergoing cardioversion or AF ablation, a single-center, randomized trial demonstrated that use of a self-administered handheld ECG resulted in earlier detection of recurrent AF¹³ and possibly improvement in survey-determined AF-related QOL²⁰ compared with usual care.

5. LIFESTYLE AND RISK FACTOR MODIFICATION (LRFM) FOR AF MANAGEMENT

5.1. Primary Prevention

Synopsis

The clinical, family history, and genetic risk factors for AF are well established (Table 3), and risk prediction models (Section 4.1, “Risk Stratification and Population Screening”) for AF have been reported and replicated.^7–9 Multiple reports have established that maintenance of optimal risk factors and ideal cardiovascular health are associated with substantially reduced risk of AF^10,11 onset and complications (Section 5, “Lifestyle and Risk Factor Modification [LRFM] for AF Management”). To reduce risk of AF onset, individuals in the general population, particularly those at increased risk of AF, should receive comprehensive integrated LRFM, including maintenance of ideal weight and weight loss if overweight or obese^1,12; pursue a physically active lifestyle,² particularly if sedentary; receive smoking cessation counseling and/or medications⁴; moderate (≤1 standard alcoholic drink/day) or abstain from alcohol and avoidance of binge drinking³; control diabetes⁵; and control BP in accordance with GDMT.^6,13 There is also an association of cannabis, cocaine, methamphetamine, or opiate use with increased incidence of AF.¹⁴

Recommendation-Specific Supportive Text

1. Most cardiovascular risk factors are associated with increased risk of new-onset AF. Observational studies have demonstrated that obesity and physical inactivity each independently increase the risk of newly diagnosed AF.^{1,2,12,15–20} However, caution should be considered in pursuing years of regular, high-volume (≥3 h/day) high-intensity endurance training given observational data linking it with increased AF risk^21–23 in men and similar “J” curve risk curve observed for high or vigorous activity in both men and women in another study.²⁴ Alcohol consumption enhances the risk of AF in a fairly linear fashion, with clear evidence that binge drinking heightens the risk.^3,25–29 Uncertainty persists regarding harms or benefits of no more than 1 regular drink per day.^28,29 Self-reported,⁴ biomarker-verified,³⁰ and genetically predicted³¹ smoking is associated with increased risk, and smoking cessation is associated with decreased risk of incident AF.⁴ The presence of either type 1 or type 2 diabetes increases AF risk,^5,32 with evidence that worse glucose control correlates with a higher probability of developing AF.³³ Hypertension is the risk factor with the highest attributable risk for AF¹⁰; intensive BP control lowers the risk of incident AF in observational and randomized data.⁶ Effective strategies to manage risk factors and prevent CVD have been reported elsewhere.^12,13

5.2. Secondary Prevention: Management of Comorbidities and Risk Factors

5.2.1. Weight Loss in Individuals Who Are Overweight or Obese

Synopsis

Obesity is associated with the development and progression of AF.⁵ It results in direct changes to the atrial myocardium forming the substrate for AF.^6–8 In addition, obesity is also associated with several comorbidities that have been independently associated with the development of AF.^9,10 Obesity has a significant adverse impact on attempts to maintain sinus rhythm, with each 5-unit increase in BMI being associated with a 10% and 13% greater risk of postoperative and postablation AF, respectively.^5,11,12 Management of weight is important in the prevention and treatment of AF.

Recommendation-Specific Supportive Text

1. In an RCT in overweight and obese individuals with BMI >27 kg/m² and AF, weight loss, as part of a comprehensive LRFM program, was associated with reduction in arrhythmia symptoms, recurrence, and burden.¹ Observational studies demonstrated graded responses commensurate with the degree of weight loss, with achievement of at least 10% weight loss associated with greater maintenance of sinus rhythm,² improved ablation outcomes,³ and reversal of AF type.⁴ In observational studies, bariatric surgery in Class III obese individuals (BMI ≥40 kg/m²) with AF was associated with improved sinus rhythm maintenance after catheter ablation^13,14 and reversal of AF type.¹⁵ The greater number of risk factors managed associated with likelihood of maintaining sinus rhythm.¹⁶ However, a small observational study in individuals with obesity (BMI ≥30 kg/m²) with long-lasting persistent AF observed no difference in symptoms or sinus rhythm maintenance despite significant weight loss, suggesting that there may be extreme substrates in which a weight loss strategy may not be effective.¹⁷

Structured programs with regular review of progress facilitate achievement of weight loss and appear essential, as demonstrated by inability to reduce AF burden in a small RCT that achieved only 4.5% weight loss in the intervention arm.

5.2.2. Physical Fitness

Synopsis

Randomized trials provide evidence that prescribed aerobic exercise interventions may reduce arrhythmia burden in those with nonpermanent AF² and improve functional capacity and health-related QOL in both permanent^1,6 and nonpermanent AF.² In the ACTIVE-AF (An Exercise and Physical Activity Program in Patients With Atrial Fibrillation) study, an exercise intervention combining home and supervised aerobic exercise over 6 months resulted in greater freedom from arrhythmia recurrence, reduced burden, and improved QOL.³

Recommendation-Specific Supportive Text

1. In patients with nonpermanent AF, aerobic exercise training may contribute to a reduction in AF burden and improve maintenance of sinus rhythm.^2–5 Aerobic exercise interventions reduce severity of AF symptoms, increase functional capacity, and improve health-related QOL among patients with both nonpermanent and permanent AF.^1,3,6 Initiation of or continuing regular exercise after development of AF was associated with a lower risk of HF and mortality in a population-based cohort study.⁷ Exercise training should be moderate- to vigorous-intensity aerobic exercise, with a target of 210 minutes per week, and should be prescribed to reduce the frequency and duration of AF episodes, while improving cardiorespiratory fitness and symptom severity.³ Exercise prescription may be further modified to patient comorbidities, such as obesity, hypertension, and diabetes.⁴ Caution should be applied to ensure adequate ventricular rate control during exercise training and absence of atrial myopathy related to excessive exercise training. To date, there is little evidence that high-intensity aerobic exercise may be favorable over moderate-intensity activities,⁸ and extreme levels of exercise have been associated with higher incidence of AF.^9,10 Exercise training may be delivered as a stand-alone intervention or as a component of multidisciplinary cardiac rehabilitation. Exercise-based cardiac rehabilitation improves QOL and functional capacity among patients with AF undergoing ablation,^11,12 although available studies have not been adequately powered to assess AF-specific outcomes, such as arrhythmia recurrence.¹³ There is mixed evidence for a reduction in hospitalization or all-cause mortality with exercise training or exercise-based cardiac rehabilitation.^12,14,15

5.2.3. Smoking Cessation

Synopsis

Observational data support that cigarette smoking in individuals with AF is associated with worse cardiovascular outcomes and death and that individuals with AF who quit smoking are less likely to develop stroke or die. Despite the benefits of smoking cessation, individuals with AF are less likely to receive smoking cessation management than counterparts without AF.^7,8 Patients with AF should be strongly advised to quit cigarette smoking⁹ and should receive GDMT for tobacco cessation, including behavioral interventions^10,11 and pharmacotherapy.²

Recommendation-Specific Supportive Text

1. In individuals with AF, cigarette smoking was associated with poorer outcomes. For individuals who have undergone AF catheter ablation, current cigarette smoking was associated with increased risk for AF recurrence.^12,13 Cigarette smoking was associated with less time in therapeutic range for patients on warfarin.¹⁴ Although some heterogeneity exists, the preponderance of studies have reported that individuals with AF who smoked cigarettes had increased risks of stroke,^3,4 HF,⁵ hospitalization,⁶ and death.^6,15 Of concern, studies have reported that individuals with AF (versus those without) were less likely to receive smoking cessation interventions.^7,8 Observational studies suggest that compared with current smokers, individuals with AF who quit smoking after AF diagnosis were less likely to experience CVD,³ stroke,^3,16 and all-cause mortality.¹⁶

5.2.4. Alcohol Consumption

Synopsis

Among patients with a diagnosis of AF, randomized trials have demonstrated a reduction in AF burden on assignment to abstinence,¹ that alcohol acutely changes human electrophysiology in a fashion that renders the atria more prone to fibrillate,² and suggest that avoiding alcohol can reduce the risk of a near-term AF event.³

Recommendation-Specific Supportive Text

1. Among patients with AF, a case-crossover study revealed a higher risk of a discrete AF episode hours after objectively confirmed alcohol consumption.⁴ In the context of a structured comprehensive management of risk factors, alcohol abstinence¹ or reduction to ≤3 standard drinks per week has been demonstrated to reduce AF symptoms,⁵ AF burden,^6,7 and progression of AF from paroxysmal to persistent.⁸

5.2.5. Caffeine Consumption

Synopsis

A randomized N-of-1 trial,² as well as longitudinal, observational studies, have generally found that caffeine consumed within normal limits is either associated with no heightened risk^3–6 or a reduced risk of incident AF.^3,5,7–9 Patients often report caffeine as a trigger of AF,¹⁰ although this has not been supported by objective data.

Recommendation-Specific Supportive Text

1. A randomized trial of caffeine failed to show any difference in postoperative AF risk,¹ and paroxysmal AF patients experienced no detectable difference in AF episodes when exposing themselves to versus avoiding caffeine in randomized N-of-1 trials.² Longitudinal, observational studies have generally found that caffeine consumed in usual amounts is either associated with no heightened risk^3–6 or a reduced risk of incident AF.^3,5,7–9Also, Mendelian randomization studies examining caffeine metabolism-related genetic variants as instrumental-variable surrogates of caffeine consumption have not shown either a harmful or protective effect related to incident AF.^8,11 Several case reports have described a relationship between excessive consumption of caffeine (involving overdoses or highly caffeinated energy drinks)^12,13 and AF in young, healthy individuals. Individuals should not begin or increase their caffeine consumption with the intent of reducing their AF risk.

   Patients often report that caffeine triggers their AF,^10,14 although this has neither been supported by nor extensively studied in an objective manner. Current studies cannot exclude the possibility of individual-level idiosyncratic relationships between caffeine and AF. It is also possible that caffeine exacerbates symptoms of AF, or causes similar symptoms of palpitations, or enhances heart rhythm awareness.

5.2.6. Diet and Dietary Supplementation

Promoting a healthy diet is an effective strategy for prevention of cardiovascular disease. The evidence pertaining to the prevention of AF using dietary supplements is inconsistent, complicated by substantial misclassification and difficulties controlling for potential confounding factors associated with dietary interventions.

Several studies have evaluated the role of omega-3 fatty acids and AF, demonstrating an inverse relationship between plasma omega-3 polyunsaturated fatty acid levels and prevalent AF.^1,2 However, with supplementation there has been no effect or a potential for greater AF.^3–7 Although vitamin D supplementation is not useful on a population basis,^8,9 its use in the perioperative period in deficient individuals reduces AF.¹⁰ Similarly, ascorbic acid has been beneficial in reducing postoperative AF in some studies, although not uniformly.¹¹

The importance of weight management on AF symptoms, burden, and recurrence is increasingly recognized,^12,13 leading to various dietary interventions. Although evidence on diets is still evolving, analyses of the ARIC (Atherosclerosis Risk in Communities) study caution using low-carbohydrate diets, which were found to increase the risk of incident AF, regardless of the type of protein or fat used to replace carbohydrate.¹⁴

5.2.7. Diabetes

AF and diabetes are associated with increased risk of cardiovascular mortality and sudden death.¹ In patients with diabetes, vascular mortality was lower in patients treated with DOACs compared with warfarin.² In patients with AF and diabetes undergoing catheter ablation, optimal glycemic control preablation may lessen the risk of AF recurrence postablation.³

5.2.8. Treatment of Hypertension

Synopsis

Renal denervation and mineralocorticoid receptor antagonists (MRAs) have been associated with decreased AF burden in clinical trials.^1,2 The only randomized trial singularly targeting BP control as a mechanism to reduce the recurrence of AF⁸ showed no significant difference in the primary outcome of recurrent symptomatic atrial arrhythmia beyond 3 months’ postablation among the 2 groups. However, clinical trials of integrated comprehensive LRFM programs in conjunction with rhythm control therapies for AF have resulted in longer arrhythmia-free survival and decreased AF burden.^3,5,9 Among patients with AF, treating elevated BP to guideline-directed goals reduces major cardiovascular events.⁹

Recommendation-Specific Supportive Text

1. In patients with AF, treatment of hypertension should aim for current BP guidelines to reduce stroke, bleeding, and other adverse outcomes.^5,10 An RCT of patients with paroxysmal AF and hypertension noted fewer recurrences among participants treated with renal denervation and pulmonary vein isolation (PVI) compared with PVI alone.² Randomized studies of mineralocorticoid receptor antagonists for hypertension have been shown to reduce AF burden, and ACE inhibitors and angiotensin II receptor blockers have been associated with lower AF incidence in secondary analyses of RCTs (see Section 8.3.4, “Upstream Therapy”). In addition, several integrated LRFM programs that have resulted in a decrease in BP have been associated with a decreased recurrence of AF.^3,11

   The SMAC (Substrate Modification With Aggressive Blood Pressure Control) AF trial randomized hypertensive patients scheduled for AF ablation to standard or aggressive BP treatment with a systolic BP target of 120 or 140 mm Hg.⁸ No significant difference was shown in recurrent symptomatic atrial arrhythmia beyond 3 months postablation (median follow-up, 14 months), and this is the only RCT to date singularly targeting BP control as a mechanism to reduce AF recurrence.⁸ Post hoc analyses of DOAC clinical trials to reduce the risk of stroke in patients with AF have consistently found lower rates of stroke in patients with controlled BP.^6,7

5.2.9. Sleep

Synopsis

SDB is a risk factor for the development of incident AF, and more severe SDB acutely increases the risk of a discrete AF episode.^14–18 Independent of SDB, poor sleep quality is associated with an increased risk of developing incident AF.^19–21 When formal testing for SDB is used, the disease is frequently observed in >20% of patients with an AF diagnosis.^1–5 Observational studies suggest that treatment of SDB may reduce the risk of AF recurrence and AF burden, but RCTs have not been adequately powered to reveal a relationship between treating SDB and reduced AF.^6–13

Recommendation-Specific Supportive Text

1. The prevalence of SDB is remarkably and consistently high in patients with AF (generally substantially >20% and sometimes >50% regardless of the type of AF population assessed), is often undetected, and, as conventional symptoms of SDB may be absent, may require formal study to establish a diagnosis of SDB.^1–5,11 Multiple observational studies have described fewer recurrences of AF and a reduction in AF burden, including after cardioversion and catheter ablation procedures, among those undergoing treatment for SDB compared with those not treated for their SDB.^6–11,22 However, small trials randomizing patients with AF and SDB treatment versus usual care have not shown significant differences in AF burden or AF recurrence, including after cardioversion or PVI ablation,^12,13,22 although studies tended to exclude patients who might benefit most from SDB treatment (eg, those who had more symptomatic sleep apnea [Epworth Sleepiness Scale score >10 or 15], severe cardiovascular disease [LVEF <40% or 45%], or severe obesity).

5.2.10. Comprehensive Care

Synopsis

Almost all patients with AF have multiple conditions that either increase AF risk or are exacerbated by AF. Patients with AF are also at risk of developing thromboembolism, stroke, and HF, so a comprehensive approach tailored to the needs of the individual patient should improve outcomes. Randomized trials have shown that interventions aimed at AF risk factors, such as alcohol use, overweight, and HF, reduce AF burden. Clinical care pathways and algorithms for AF management, including LRFM, have shown promise in randomized trials through coordinating care and facilitating comprehensive AF management.

Although comprehensive care of the multiple conditions associated with AF is logical, it is unclear whether integrated care by a multidisciplinary team leads to better outcomes than comprehensive care by a single clinician applying an evidence-based AF clinical care algorithm or pathway. Nevertheless, coordination of care among multiple clinicians should also improve care, even though it has not been evaluated rigorously.⁷

Recommendation-Specific Supportive Text

1. Randomized trials have shown the efficacy of many individual components of patient-centered care for AF, as discussed earlier in this section. Several randomized and nonrandomized studies have utilized comprehensive programs for patients with AF.^2,8,9 The RACE 3 (Rate Control versus Electrical cardioversion for persistent atrial fibrillation) trial³ found that multifaceted treatment for patients with AF and early HF (with mineralocorticoid receptor antagonists, statins, ACE inhibitors and/or angiotensin receptor blockers [ARBs], and cardiac rehabilitation) improved maintenance of sinus rhythm.

2. Various clinical care pathways, algorithms, and electronic clinical decision support systems for the care of patients with AF have been tested in clinical trials, with mixed results. Care provided by specialty nurses using a clinical support system aimed at enhancing adherence to guideline-directed therapies improved cardiovascular outcomes in a single center study⁴ but not in the subsequent multicenter RACE 4 trial, where results were associated with experience of the center, although adherence to guideline-based recommendations was still greatly improved.¹⁰ In a cluster randomized trial of elderly patients with AF where integrated care intervention consisting of (a) quarterly AF check-ups by trained nurses in primary care, also focusing on comorbidities, (b) monitoring of anticoagulation therapy in primary care, and finally (c) easy-access availability of consultations from cardiologists and anticoagulation clinics, the patients assigned to integrated care experienced a 45% reduction in all-cause mortality when compared with usual care.¹¹ In another study (SAFETY [Standard versus Atrial Fibrillation specific management study]), a posthospital discharge management program specific to AF was associated with proportionately more days alive and out of hospital but not prolonged event-free survival relative to standard management.¹² On the other hand, data on technology-based clinical decision support systems are mixed, with a reduction in stroke and thromboembolism in the mAF-App 2 (Mobile Health Technology for Improved Screening and Optimized Integrated Care in Atrial Fibrillation) trial⁵ and improved anticoagulation use in the CDS-AF (clinical decision support tool for stroke prevention) trial,⁶ but no significant effect on primary outcomes in 3 other randomized trials.^13–15 Educational interventions for both providers and patients can improve use of anticoagulation.¹⁶

6. PREVENTION OF THROMBOEMBOLISM

6.1. Risk Stratification Schemes

Synopsis

Patients with AF have an increased risk of stroke that varies widely among individuals. Several risk scores based on clinical factors have been developed.^1–3 Risk scores should be evaluated using accepted criteria¹¹: their ability to discriminate between high- and low-risk individuals (eg, as assessed by the c-index), their accurate calibration to actual risk levels, and their validation in independent populations. A patient’s absolute risk of stroke is central to recommendations about anticoagulation and can be characterized as low (∼<1%/y), intermediate (∼1 to ∼2%/y), and high (∼>2%/y). Currently used risk scores discriminate moderately well between patients at higher and lower risk yet can greatly overestimate or underestimate absolute risk levels⁴ in different populations. Current risk scores may be inaccurate because they omit other factors that alter risk of stroke, especially characteristics of AF^12–14; consequently, risk scores should be calibrated against the actual annual rates of stroke in the target population to assure an accurate, unbiased risk prediction. Newer risk scores may modestly improve risk discrimination (c-index) compared with CHA₂DS₂-VASc and may offer potential advantages in specific populations (Table 8). Online calculators are available for the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation),¹ CHA₂DS₂-VASc,² and GARFIELD-AF³ (Global Anticoagulant Registry in the Field-Atrial Fibrillation) risk scores.

Anticoagulation increases the risk of bleeding, so patients with AF are generally evaluated for bleeding risk as part of SDM about anticoagulation. Currently used bleeding risk scores—HAS-BLED (hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalized ratio [INR], elderly [age ≥65 years], drugs/alcohol concomitantly]),⁵ HEMORR₂HAGES (hepatic or renal disease, ethanol abuse, malignancy, older age [≥75 years], reduced platelet count or function, re-bleeding risk, hypertension [uncontrolled], anemia, genetic factors, excessive fall risk, stroke),⁶ and ATRIA (anemia, renal disease, elderly [age ≥75 years], any prior bleeding, hypertension)⁷—discriminate poorly between patients with and without bleeding¹⁵ and include many nonspecific factors that predict an increased risk of stroke as well as an increased risk of bleeding (eg, age, hypertension, renal disease, and previous stroke). Assessment of factors that specifically predict an increased risk of bleeding without predicting an increased risk of stroke is more helpful when balancing risks and benefits of anticoagulation.

Recommendation-Specific Supportive Text

1. Several risk scores based on clinical factors have been developed^1–3; in general, risk discrimination is improved by including more predictors (eg, CHA₂DS₂-VASc improved on the original CHADS₂² [congestive heart failure, hypertension, age >75 years, diabetes, stroke/transient ischemia attack/thromboembolism] by adding additional risk factors and age categories). However, the absolute risk associated with any particular score level varies widely across populations; among 15 cohort studies, patients with a CHA₂DS₂-VASc score of 2 had annual rates of stroke that ranged from low to high: <1% in 4 cohorts, 1% to 2% in 6 cohorts, and >2% in 5 cohorts,⁴ although higher scores were associated with higher stroke risk in each cohort (Figure 9). The CHA₂DS₂-VASc score is considered the most validated score, most therapies have used that score to prove efficacy, and thus CHA₂DS₂-VASc is generally the preferred score. Yet, despite its extensive use, CHA₂DS₂-VASc has shown suboptimal performance in selected populations, such as those with renal disease, prompting the creation of other scores. Newer risk scores, such as the ATRIA^1,16,17 and GARFIELD-AF^3,18 scores, modestly improve risk discrimination (c-index) compared with CHA₂DS₂-VASc, but their calibration and risk reclassification performance has not been as rigorously evaluated. A recent meta-analysis reported that 19 risk scores, 329 external validations, and 76 risk score updates have been conducted to predict ischemic stroke in patients with AF.¹⁹ Potential differences of the most studied risk scores are listed in Table 9. Some potential advantages of other scores, for example, the GARFIELD-AF score, includes mortality and bleeding risk to facilitate discussion with patients in a more comprehensive way. Also, when uncertainty exists, a score such as GARFIELD-AF adds additional variables to consider, such as smoking status, renal disease, and dementia. This guideline does not intend to preclude the future development of more accurate scores for stratifying patients.

2. Clinical decisions surrounding stroke prevention therapy in patients with AF must balance the risks of ischemic stroke, the risks of bleeding with treatment, net clinical benefit, and patient preferences. The most studied bleeding risk scores (HAS-BLED,⁵ HEMORR₂HAGES,⁶ and ATRIA⁷) discriminate poorly between patients with and without bleeding²⁰ (c-index 0.58 to 0.59 in a French nationwide study²¹). These scores are problematic to use in clinical decision-making because they incorporate several clinical factors that increase the risks of both stroke and bleeding (eg, age, hypertension, renal disease, and previous stroke), which makes it difficult to balance the benefits and risks of anticoagulation because the same risk factors also predict higher risk of stroke. Consideration of factors that specifically indicate a higher risk of bleeding without predicting higher risk of stroke (eg, previous bleeding, anemia, and certain medications) may better inform decision-making about the balance of benefits and harms expected from anticoagulation. Assessment of risk factors specific for bleeding may suggest interventions to reduce bleeding risk, such as discontinuing antiplatelet medications or nonsteroidal anti-inflammatory medications or the use of LAAO devices.¹⁰

3. The decision to treat with OACs for stroke prevention in patients with a CHA₂DS₂-VASc of 1 (CHA₂DS₂-VASc of 2 in women) may at times require additional discussion with patients as the strength or recommendation is less robust (2a), and often patients might be more reluctant. Also, the 1-point-concept of risk estimation in the subgroup of patients with a CHA₂DS₂-VASc score of 1 has shown to be simplistic because the magnitude of risk for each factor is heterogeneous and because data have shown a wide range of risk depending on the studied cohort.²² Thus, as part of SDM discussions with patients, other factors, such as AF burden, can be considered when interpreting a stroke risk score.^12–14,23 Additional factors, such as degree of hypertension control, can also affect the risk of stroke; in the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial, a single elevated BP measurement during the study was associated with a 50% increased risk of stroke.^23,24 Additional patient-specific risk factors, such as certain biomarkers (eg, proBNP), LA or left atrial appendage (LAA) function and anatomy, or electrocardiographic features, among others, have been demonstrated to influence stroke risk, yet it is unclear how to incorporate them into clinical practice.¹² Antithrombotic treatment with DOACs seems to offer a superior net-benefit with regard to prevention of thromboembolic events and the risk of major bleeding compared with vitamin K antagonists (VKAs). Consideration of these additional factors may inform decision-making for patients with an intermediate risk of stroke who remain uncertain on deciding anticoagulation.

4. Unless an absolute contraindication to anticoagulation is present, bleeding risk scores have limitation in clinical decision-making because the most commonly used scores (HAS-BLED,⁵ HEMORR₂HAGES,⁶ and ATRIA⁷) are based on several clinical factors that indicate higher risks of both stroke and bleeding, and patients with higher risk of bleeding also tend to have a higher risk of stroke. Furthermore, a bleeding risk score cannot be interpreted in isolation because it does not assess the net clinical benefit of anticoagulation or balance the risk of bleeding against the risk of stroke. Population-based studies suggest that the benefits of stroke prevention with oral anticoagulation generally outweigh the risks of bleeding, even in patients determined to be at high risk for bleeding.^8,9 Decision-making about oral anticoagulation should be based on consideration of both benefits and harms, not by using bleeding risk scores in isolation, and the best utility of these scores may be identifying potential modifying risk factors.^8,9

• Download figure
• Download PowerPoint

6.2. Risk-Based Selection of Oral Anticoagulation: Balancing Risks and Benefits

Synopsis

Prevention of stroke is important in patients with AF to maximize survival, health, and QOL. Selection of stroke risk reduction therapy should be guided by the patient’s risk of stroke, risks of bleeding with therapy, and their individual preferences. When considering stroke prevention therapy, the risk of stroke should inform the decision regardless of the pattern of AF (paroxysmal, persistent, long-standing persistent, or permanent). All decisions regarding stroke prevention therapy should be periodically reassessed since a patient’s risk, eligibility, and preferences can change over time.

Recommendation-Specific Supportive Text

1. Although the risk of stroke and systemic embolism and all-cause mortality are increased in persons with more persistent forms of AF, selection of stroke prevention therapy should be based on the risk of stroke and not the pattern of AF. Treatment effects with oral anticoagulation are consistent across AF patterns (paroxysmal, persistent, long-standing persistent, or permanent) in trials of stroke prevention therapy.^1–3

2. AF is a lifelong condition and, thus, patient characteristics, risk factors, and net clinical benefit can and often will change over time.⁴ In long-term follow-up, stroke risk increases due to age and accumulation of other risk factors.⁶ Physiologic factors that impact stroke prevention therapy also change over time and can have important implications for proper medication dosing and patient safety.⁵ Typically, periodic assessment should be performed once a year but might need to be performed more frequently in the context of changes in clinical status, such as reduction in renal function or development of additional risk factors.

6.3. Oral Anticoagulants

Vitamin K Antagonists

Since the 1950s, warfarin was used as a first-line therapy until DOACs came into practice. A narrow therapeutic window based on international normalized ratios (INRs), frequent monitoring, more frequent drug interactions (mainly through CYP2C9), dietary restrictions, and low clinical safety profile affected the routine use of warfarin in practice. Because of affordability issues of DOACs for some patients with AF, warfarin is still an appropriate OAC due to its lower cost for patients who cannot afford DOACs. About 21% of patients with nonvalvular AF were still receiving warfarin, while the rest received DOACs in the first quarter of 2017.¹ Warfarin remains the first-line therapy in patients with AF and moderate-severe rheumatic mitral stenosis or mechanical heart valves. Clinical studies show that the target INR is between 2 and 3, and risk of bleeding becomes mostly apparent when INR exceeds 4. OAC use in special populations will be discussed in separate sections.

Direct oral anticoagulants

DOACs were developed to address the disadvantages of warfarin and are currently recommended as the first-line therapy over warfarin in patients with AF (except moderate to severe mitral stenosis or mechanical heart valve recipients) in this guideline. All 4 pivotal clinical trials comparing individual DOACs (apixaban, dabigatran, edoxaban, and rivaroxaban) with warfarin showed superiority or noninferiority to warfarin for the prevention of stroke or systemic embolism in patients with AF except for moderate to severe mitral stenosis or mechanical heart valve.^2–5 They also showed significantly lower risks of major bleeding in the apixaban, dabigatran 110 mg twice daily group, and edoxaban 30 mg or 60 mg daily dose groups compared with the warfarin group or nonsignificant differences in major bleeding between dabigatran 150 mg twice daily group or the rivaroxaban group and warfarin. All DOAC groups showed significantly lower risks of intracranial hemorrhage (ICH) compared with warfarin. Gastrointestinal bleeding risks were significantly higher in the dabigatran 150 mg twice daily, edoxaban 60 mg once daily, and rivaroxaban groups compared with the warfarin group. However, the apixaban group did not significantly increase the risk of gastrointestinal bleeding compared with the warfarin group.

6.3.1. Antithrombotic Therapy

Synopsis

A high risk for stroke or systemic embolism is about 2% per year, and all the DOAC trials (Re-LY [Randomized Evaluation of Long-Term Anticoagulation Therapy]; ROCKET AF [Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation]; ARISTOTLE; and ENGAGE AF-TIMI 48 [Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation – Thrombolysis in Myocardial Infarction 48]) included patients with this level of risk.^1–4 Patients at intermediate risk (1%-2%/y) can also benefit from anticoagulation, and the RE-LY¹ and ARISTOTLE³ trials included this population. Stroke risk scores applied to cohorts give different stroke rates, and therefore any score should be viewed as only an estimate of true risk; in addition some scores used stroke, while others used thromboembolic events. Nonetheless, it is practical to use a validated risk score, such as CHA₂DS₂-VASc, ATRIA, or GARFIELD-AF. Future research may yield improved risk scores that refine how to incorporate risk modifiers, such as female sex¹² and other parameters such as AF burden. Anticoagulation has also been shown to be superior to antiplatelet therapy to reduce stroke risk.^8,9 Recommendations for antithrombotic selection in valvular heart disease (VHD) are provided in Section 6.8.5 (“AF in VHD”) and for AFL in Section 6.8.6 (Figure 10).

• Download figure
• Download PowerPoint

Recommendation-Specific Supportive Text

1. Randomized trials published in the 1990s established the superiority of anticoagulation, at that time limited to warfarin, to reduce stroke.¹¹ Since that time, management has advanced considerably, and the randomized trials that compared DOACs with warfarin^1–4 are more relevant to current antithrombotic management. In these trials^1–4 and metanalyses,^5–7 DOACs were favored for lower risk of stroke, systemic embolism, and ICH.^1–7 These trials reported the CHADS₂ score. The recently studied ATRIA and Swedish cohorts^13,14 reported lower stroke rates, by CHADS₂ score, compared with that previously reported.¹⁵ A CHADS₂ score of 1 gave a stroke rate of 1.20 in ATRIA¹⁴ and 2.4 in a Swedish cohort.¹³ A CHADS₂ score of 2 gave a stroke rate of 2.59 in ATRIA¹⁴ and 3.5 in the Swedish cohort.¹³ Therefore, the stroke risk of patients in the rivaroxaban² and edoxaban⁴ trials was likely >2% given a required minimum CHADS₂ score of 2, and lower in the apixaban³ and dabigatran¹ trials that included scores of 0 and 1. A Markov state transition decision model¹⁶ concluded that anticoagulation was preferred for a stroke rate of 1.7%. Therefore, a stroke and systemic embolism risk threshold of 2% is likely to yield a benefit that far exceeds risk. In some conditions, such as hypertrophic cardiomyopathy, the risk of stroke is high enough independent of risk score to indicate anticoagulation.¹⁷

2. In the randomized DOAC trials, all DOACs achieved noninferiority^1–4 and in 2 trials (for dabigatran, RE-LY, and for apixaban, ARISTOTLE) were superior to warfarin. With warfarin, regular assessment of the INR is necessary to maintain a therapeutic value. In the RCTs, the mean time in therapeutic range was only 55% in ROCKET AF² and highest at 66% in ARISTOTLE.³ In a meta-analysis,⁶ DOACs, compared with warfarin, had a relative risk of stroke of 0.81 (95% CI, 0.73-0.91; P<0.0001), a relative risk of mortality of 0.90 (95% CI, 0.85-0.95; P=0.0003), relative risk of ICH of 0.48 (95% CI, 0.39-0.59; P<0.0001), and increased relative risk of bleeding of 1.25 (95% CI, 1.01-1.55; P=0.043). Although patients with moderate to severe mitral stenosis or mechanical valve were excluded from DOAC trials (and subsequently shown they may have worse outcomes with DOACS), other forms of VHD were allowed, such as aortic stenosis or regurgitation or mitral regurgitation. Bioprosthetic valves and valve repair were allowed in the edoxaban (ENGAGE AF) and apixaban (ARISTOTLE) trials, and valve repair in the rivaroxaban (ROCKET AF) trial. A systematic review¹⁸ of patients with VHD (other than mitral stenosis or mechanical valve) concluded that DOACs were safe.

3. As the DOAC trials^1–4 demonstrated improved safety compared with warfarin, the threshold of using a DOAC might be different than for warfarin.¹⁶ A Markov decision model found the tipping point for warfarin to be at 1.7%/year stroke risk.¹⁶ Considering the improved ICH and mortality risk of DOACs compared with warfarin in meta-analyses of the DOAC trials,^5–7 it is appropriate to designate a lower stroke risk threshold if a DOAC is utilized. The dabigatran¹ and apixaban³ trials included lower risk patients and about one-third of patients had a CHADS₂ score of 0 or 1. This likely corresponds to an estimated annual stroke risk of about 1%, considering more recent cohorts.^13,14 This is further supported by meta-analyses that showed a consistent benefit of DOACs across a broad range of vulnerable patients, including as classified by CHADS₂ scores.^6,7

4. A meta-analysis of AF trials from the 1990s and early 2000s found that while antiplatelet therapy (APT), most commonly aspirin, reduced stroke and systemic embolism compared with placebo, APT was inferior to warfarin in patients with AF.¹¹ Aspirin was studied compared to apixaban in the AVERROES trial (Apixaban Versus ASA to Prevent Stroke In AF Patients Who Have Failed or Are Unsuitable for Vitamin K Antagonist Treatment)⁹ for patients unsuitable for VKAs; this trial was stopped early due to the benefit of apixaban over aspirin to prevent stroke or systemic embolism, while major bleeding was similar between the 2 arms. The combination of clopidogrel and aspirin was compared to VKAs in ACTIVEW (Atrial fibrillation Clopidogrel Trial with Irbesartan for prevention of Vascular Events),⁸ and this trial was stopped prematurely due to the superiority of anticoagulation with VKAs to prevent stroke, non–central nervous system systemic embolus, MI, or vascular death. Unless there is an indication for antiplatelet therapy, such as coronary artery disease (CAD) or vascular disease, patients with AF should not be prescribed antiplatelet therapy to reduce stroke risk.

5. Aspirin therapy was compared with no treatment in patients with low stroke risk in a multicenter RCT in Japan, the Japan Atrial Fibrillation Stroke Trial.¹⁰ This trial was stopped early when an interim analysis showed that aspirin gave a marginally higher risk of major bleeding and was unlikely to prevent primary or secondary endpoints. The primary endpoint was noncardiovascular death, ICH, major bleeding, and peripheral embolization. A meta-analysis of 7 trials that studied aspirin versus placebo found that aspirin reduced stroke by 19%, which did not reach statistical significance with a CI that included 0 (–1% to 35%). Four of these were primary prevention trials and showed an absolute risk reduction with aspirin of 0.8% per year, with a number needed to treat of 125.¹¹ Although aspirin has not been studied in patients without any risk factors for stroke, patients without stroke risk may derive no benefit from aspirin therapy. In fact, 1 study that analyzed stroke mechanisms in the SPAF (Stroke Prevention in Atrial Fibrillation) I-III trials demonstrated that aspirin did not decrease the rate of cardioembolic stroke and that AF patients at highest risk for stroke are those with the highest rates of cardioembolic stroke and have the greatest reduction in stroke with anticoagulants.¹⁹ Thus, modest reduction in stroke observed in some trials may have been related to noncardioembolic stroke in patients with additional risk factors.

6.3.1.1. Considerations in Managing Anticoagulants

Synopsis

Multiple factors need to be considered to select an optimal OAC in patients with AF. Efficacy, safety, insurance coverage, renal/hepatic function, drug interaction screening, medication adherence, and patient preferences are the major factors for consideration. Dosing of DOACs in patients with AF should be selected based on age, renal function, weight, and concomitant medications. The hepatic function also needs to be evaluated for the appropriateness of DOACs. Characteristics and dosing of OACs are summarized in Table 13. Although DOAC drug interactions occur in practice less frequently than with warfarin, clinically significant drug interactions through CYP3A4 and/or p-glycoprotein should be carefully evaluated. Strong CYP3A4 and/or p-glycoprotein inhibitors such as ketoconazole, itraconazole, and ritonavir may significantly increase DOAC plasma levels and the risk of bleeding, while strong CYP3A4 and/or p-glycoprotein inducers such as rifampin, phenytoin, phenobarbital, primidone, carbamazepine, or St. John’s wort may lower DOAC plasma levels and increase the risks of stroke or systemic embolism. The routine measurement of DOAC plasma concentrations is not indicated in practice due to the lack of well-established therapeutic ranges in the literature. DOAC levels may be indicated when clinicians assess DOAC adherence for potentially noncompliant patients, quantify residual anticoagulation levels before emergency invasive procedures/surgeries, or evaluate the absorption of DOAC after bariatric surgery. Regular monitoring is recommended to optimize indications, ensure appropriate dosing, and avoid adverse effects. Suggested laboratory monitoring is summarized in Figure 11.

• Download figure
• Download PowerPoint

If warfarin is selected over DOACs, achievement of higher time in therapeutic range (eg, ≥70%), and thorough consideration of drug-drug interactions, vitamin K food intake advice, and patient education on adherence to dosing instructions are important to reduce adverse effects.

Recommendation-Specific Supportive Text

1. Rivaroxaban is contraindicated for patients receiving ketoconazole or ritonavir, because ketoconazole or ritonavir coadministration significantly increased rivaroxaban plasma levels by 158% or 153%, respectively.¹ Apixaban also requires dose adjustment to 2.5 mg twice daily for patients receiving apixaban 5 mg twice daily when ketoconazole, itraconazole, or ritonavir is started.¹³ Apixaban plasma level was significantly increased by 99% when ketoconazole was coadministered in healthy subjects.² The concomitant therapy of dronedarone or ketoconazole for patients with creatinine clearance (CrCl) 30 to 50 mL/min receiving dabigatran requires dose adjustment to dabigatran 75 mg twice daily because it produced comparable dabigatran exposure in patients with CrCl 30 to 50 mL/min receiving the concomitant therapy to patients with CrCl 15 to 29 mL/min receiving only dabigatran.¹⁴ A retrospective study revealed that the probability of patients with subreference DOAC levels was significantly higher in patients receiving antiepileptic drugs compared to those not receiving any CYP3A4 inducers.⁴ The claim-based retrospective cohort study showed that CYP-inducing antiepileptic drug use was significantly associated with an 86% increase in thromboembolic and ischemic adverse events in patients receiving DOACs.⁶ A prospective multicenter cohort study corroborated that the patients receiving concomitant therapy of DOACs and antiepileptic drugs developed a higher rate of stroke/TIA/systemic embolism (5.7% patient-year).⁷

2. Warfarin remains the preferred agent in patients with AF receiving CYP3A4/p-glycoprotein–inducing agents, or moderate-severe mitral stenosis or mechanical heart valve. Also, warfarin may be selected over DOACs in patients with AF due to higher cost or intolerances of DOACs. The optimal INR control, with a therapeutic INR goal of 2 to 3, needs to be achieved through the appropriate drug-drug interaction management, vitamin K dietary education, and routine INR monitoring. A systematic review of 47 studies found that time in therapeutic INR range showed negative correlation with risks of thromboembolism (R, –0.59; P=0.01) and major bleeding (R, –0.59; P=0.002).⁸ An RCT revealed that a weekly vitamin K dietary modification significantly achieved a therapeutic INR more frequently than a conventional group (74 versus 58%; P=0.04) at 90 days after randomization.⁹ A meta-analysis of 11 RCTs and 61 observational studies for warfarin drug interactions found that the concomitant use of APT (odds ratio [OR], 1.74 [95% CI, 1.56-1.94]), nonsteroidal anti-inflammatory drugs (OR, 1.83 [95% CI, 1.29-2.59]), selective serotonin reuptake inhibitors (OR, 1.62 [95% CI, 1.42-1.85]), or antimicrobial agents (OR, 1.63 [95% CI, 1.45-1.83]) was significantly associated with higher risks of clinically relevant bleeding.¹⁰

3. A significant number of patients with nonvalvular AF are receiving DOAC off-label doses not compliant with labeling.^15–18 A meta-analysis of cohort studies showed that inappropriately lower DOAC doses are significantly associated with higher stroke or systemic embolism risks compared with the standard labeled doses (OR, 1.21 [95% CI, 1.02-1.43]; P=0.03), but no significant differences in bleeding risks were observed (OR, 1.03 [95% CI, 0.92-1.15]; P=0.62).¹¹ Of note, 9 of 16 included studies were performed in Asia, while 4 studies in the United States, 2 studies in Europe, and 1 worldwide study were included. Another meta-analysis of observational studies found that inappropriately higher DOAC doses were significantly associated with higher risks of ischemic stroke/systemic embolism (hazard ratio [HR], 1.26 [95% CI, 1.11-1.43]; P=0.003) or major bleeding (HR, 1.30 [95% CI, 1.04-1.62]; P=0.025) compared with the standard doses.¹² In this study, underdosing was also associated with a higher risk of net clinical outcome (HR, 1.19 [95% CI, 1.04-1.40]; P=0.04) and all-cause death (HR, 1.24 [95% CI, 1.04-1.48]; P=0.02). Off-label dosing should be avoided to optimize the efficacy and safety of DOACs.

6.4. Silent AF and Stroke of Undetermined Cause

Synopsis

Approximately 25% to 30% of ischemic strokes remain cryptogenic after standard stroke evaluation.^2,3 A large proportion of these events are presumed to be related to occult AF, particularly those of embolic appearance.^4,5 APT is the recommended treatment of choice for patients with cryptogenic stroke, including embolic stroke of undetermined source.^6,7 For patients with AF in general, however, anticoagulation has been shown to be superior to APT for stroke risk reduction.⁸ Thus, detection of occult AF after stroke may have significant therapeutic implications. Randomized trials have shown longer durations of cardiac monitoring result in higher rates of AF detection after stroke.^1,9,10 However, data are limited regarding the effects of extended monitoring on risk reduction of recurrent stroke or poststroke mortality. For recommendations on anticoagulation and risk factors, refer to Section 6.4.1 (“Oral Anticoagulants”).

Recommendation-Specific Supportive Text

1. Growing evidence supports the use of extended cardiac monitoring for the identification of occult AF in patients with cryptogenic stroke.^11,12 In the EMBRACE (30 Day Event Monitoring Belt for Recording Atrial Fibrillation After a Cerebral Ischemic Event) trial, which randomized 572 patients ≥55 years of age with recent cryptogenic stroke or TIA to either 30-day external loop recorder or conventional 24-hour Holter monitoring, extended monitoring was associated with higher rates of AF detection at 90 days.¹

   The CRYSTAL-AF (Cryptogenic Stroke and Underlying AF) trial randomized 441 patients ≥40 years of age with recent cryptogenic stroke or TIA to cardiac monitoring with either insertable loop recorder or conventional follow-up ECGs. Implantable recorder was superior in detecting AF at 6 months (8.9% versus 1.4%), 12 months (12.4% versus 2.0%), and 3 years (30% versus 3%).⁹ In FIND-AF (Future Innovations in Novel Detection of Atrial Fibrillation), which compared repeated sets of 10-day Holter monitoring (at baseline, 3-month, and 6-month timepoints) to conventional 24-Holter in patients ≥60 years of age with recent stroke, higher rates of detection were associated with repeated monitoring (14% versus 5%; absolute difference, 9.0% [95% CI, 3.4-14.5]; P=0.002).¹⁰ Finally, the PER DIEM (Post-Embolic Rhythm Detection with Implantable vs External Monitoring) RCT also showed a significantly greater proportion of patients with AF detected at 1 year with prolonged monitoring.¹³ Additional studies are needed, however, to determine whether extended cardiac monitoring improves long-term poststroke outcomes.

6.4.1. Oral Anticoagulation for Device-Detected Atrial High-Rate Episodes Among Patients Without a Previous Diagnosis of AF

Synopsis

AHREs detected by a cardiovascular implantable electronic device are associated with a stroke risk lower than that of clinical AF⁶ that varies according to the episode duration¹ and CHA₂DS₂-VASc score.² Clinician confirmation of the duration and nature of the longest atrial high-rate episode is recommended.¹ Episodes lasting ≥24 hours are associated with a significant risk of stroke or systemic embolism^1,7 that may be reduced with oral anticoagulation.³ By contrast, short episodes, commonly defined as <5 minutes, are not associated with clinical events.^4,5 Patients with episodes of intermediate duration and with an elevated stroke risk may benefit from oral anticoagulation.² There is an interaction between AHRE duration and CHA₂DS₂-VASc score,² suggesting both may be used to guide oral anticoagulation candidacy (Figure 12). Of note, the threshold for anticoagulation is higher as in device-detected AF the risk of stroke may be lower than in clinical AF.

• Download figure
• Download PowerPoint

Recommendation-Specific Supportive Text

1. AHREs lasting ≥24 hours are associated with an elevated risk of stroke that varies according to CHA₂DS₂-VASc score. AHRE burden is most commonly defined according to the duration of the longest-detected episode, though cumulative duration has also been examined. In a secondary analysis of the ASSERT (Atrial Fibrillation Reduction Atrial Pacing Trial), adjudicated AHREs >24 hours was associated with an increased risk of subsequent stroke or systemic embolism (adjusted HR, 3.24 [95% CI, 1.51-6.95]; P=0.003).¹ Similar findings are seen in observational studies in which most AHREs were not adjudicated,⁷ though clinician confirmation is preferred. In an observational study linking administrative claims data² to a device database in which continuous rhythm monitoring data were available but not adjudicated, the risk of stroke among patents with subclinical AF >23.5 hours and a sex-modified CHA₂DS₂-VASc score ≥2 not taking anticoagulation exceeded 1% per year, a threshold that may be appropriate for use of non-VKAs, with a still higher stroke risk with increasing CHA₂DS₂-VASc score ≥3.⁸ In a retrospective cohort study of patients with subclinical AF >24 hours with an average CHA₂DS₂-VASc score of 4.2±1.4, treatment of subclinical AF >24 hours with OAC was associated with a reduced stroke risk (HR, 0.28 [95% CI, 0.10-0.81]; P=0.02).^2,3

2. Although the risk of stroke among patients with an AHRE of intermediate duration is lower than that among patients with an AHRE ≥24 hours, the precise stroke risk is uncertain and may vary according to the CHA₂DS₂-VASc score. In ASSERT, AHREs lasting >6 minutes to 24 hours were not associated with stroke or systemic embolism compared with patients without AHREs.¹ The CHA₂DS₂-VASc score of this subpopulation was 2.2±1.0 (mean±SD). In the aforementioned study linking administrative claims data to a device database in which continuous rhythm monitoring data were available, the risk of stroke and systemic embolism off anticoagulation exceeded 1% among patients with a sex-modified CHA₂DS₂-VASc score of 3 to 4 and exceeded 2% among patients with a sex-modified CHA₂DS₂-VASc score ≥5.²

3. Brief episodes of subclinical AF are at low risk of clinical events. In RATE (Registry of Atrial Tachycardia and Atrial Fibrillation), patients with clinical events such as hospitalization, death, or stroke were more likely than those without to have AHREs with either onset or offset of an adjudicated, single ECG (31.9% versus 22.1% among patients with pacemakers and 28.7% and 20.2% among patients with implantable cardioverter-defibrillators [ICDs]).⁴ AHREs >5 minutes are highly correlated with AF and AFL, whereas shorter episodes often represent other tachyarrhythmias.⁹ To remove most episodes of oversensing, AHREs lasting ≤5 minutes were excluded from analysis of MOST (Mode Selection Trial).¹⁰ A similar threshold was used in several subsequent key studies to exclude spurious events.^1,3,5,11,12

6.5. Nonpharmacological Stroke Prevention

Although oral anticoagulation is the standard of care to reduce the risk of ischemic stroke in patients with AF, it is contraindicated in some patients due to an excess risk of major bleeding. Surgical and percutaneous techniques to occlude the LAA have been developed to reduce the risk of ischemic stroke. These techniques have the potential to obviate the need for or supplement long-term oral anticoagulation in select patients. Surgical LAAO is a particularly important consideration in patients with AF who undergo cardiac surgery.

6.5.1. Percutaneous Approaches to Occlude the LAA

Synopsis

A large body of evidence supports the use of OACs to reduce the risk of ischemic stroke in patients with AF; however, OACs may be contraindicated in some patients (Table 14). pLAAO devices are designed to prevent embolization of LAA thrombi and potentially obviate the need for OAC for stroke risk reduction. RCTs have demonstrated pLAAO to be noninferior to warfarin and DOACs for stroke and systemic embolism with a reduced risk of major bleeding.^1–4 A prospective nonrandomized study has shown that patients deemed unsuitable for OAC who undergo pLAAO implant have a lower-than-expected risk of ischemic stroke.⁵ Prospective registries of pLAAO device implant in patients who are not on long-term OAC have shown a high rate of procedural success, a low risk of major procedure-related complications, and a low rate of ischemic stroke in follow-up.^6,7 Based on these findings, pLAAO is reasonable in patients with nonvalvular AF who have a contraindication to long-term OAC due to a nonreversible cause. Additionally, pLAAO may be reasonable as an alternative to OAC based on patient preference after careful consideration of procedural risk and with the understanding that there is a much greater body of evidence supporting OAC in this population in general.

Recommendation-Specific Supportive Text

1. pLAAO has been evaluated in patients with AF and a contraindication to OAC. The ASAP (ASA Plavix Feasibility Study With Watchman Left Atrial Appendage Closure Technology) study was a nonrandomized trial evaluating the performance of the Watchman device in 150 patients with AF who were deemed ineligible for OAC with warfarin.⁵ The rate of ischemic stroke was lower in patients who received the Watchman device than would be expected based on the CHADS₂ scores of the cohort (2.3% versus 7.3%). The EWOLUTION (Evaluating Real-Life Clinical Outcomes in Atrial Fibrillation Patients Receiving the WATCHMAN Left Atrial Appendage Closure Technology) prospective registry showed a high rate of Watchman device procedural success (98.5%) with a low ischemic stroke risk (1.1%). This low ischemic stroke was demonstrated despite most of the patients (73%) not using oral anticoagulation periprocedurally.⁶ The PINNACLE FLX (Protection Against Embolism for Nonvalvular AF Patients: Investigational Device Evaluation of the Watchman FLX LAA Closure Technology) prospective registry of the next-generation Watchman FLX device showed a high rate of procedural success (98.8%) and a low rate (0.5%) of the primary safety endpoint (death, ischemic stroke, systemic embolism, procedure-related events requiring open cardiac surgery or major endovascular intervention).⁷ Registry data have shown substantial variation in postprocedure antithrombotic regimens in real-world practice compared with protocols in the pivotal RCTs; the effect this has on long-term outcomes has not been clearly established.⁸ ASAP-TOO (Assessment of the Watchman Device in Patients Unsuitable for Oral Anticoagulation) is an ongoing RCT evaluating the safety and efficacy of LAAO in patients with AF who are deemed ineligible for oral anticoagulation.⁹

2. PROTECT AF (Watchman Left Atrial Appendage System for Embolic Protection in Patients With Atrial Fibrillation) was a noninferiority RCT comparing pLAAO with the Watchman device with warfarin in patients with AF at an increased risk for stroke. pLAAO was noninferior to warfarin for both primary composite efficacy (stroke, systemic embolism, and cardiovascular/unexplained death) and safety (procedure-related events and major bleeding) endpoints.¹ pLAAO was superior to warfarin in terms of cardiovascular mortality (HR, 0.40, 95% CI, 0.21-0.75; P=0.005) and all-cause mortality (HR, 0.66, 95% CI, 0.45-0.98; P=0.004). The PREVAIL (Watchman LAA Closure Device in Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy) study was a follow-up RCT that compared the Watchman device with warfarin in a similar patient population.² pLAAO did not achieve noninferiority for the first primary composite efficacy endpoint (stroke, systemic embolism, and cardiovascular/unexplained death) but was limited in statistical power because of unexpectedly low event rates. pLAAO was statistically noninferior to warfarin for nonprocedure-related strokes. Higher implant success and lower procedure-related safety event rates were observed in PREVAIL than PROTECT AF. A 5-year patient-level meta-analysis of these trials demonstrated similar efficacy of pLAAO for stroke prevention and reduced rates of major bleeding compared with warfarin.³ PRAGUE-17 (Left Atrial Appendage Closure vs Novel Anticoagulation Agents in Atrial Fibrillation) was an RCT that compared pLAAO with DOACs in patients with nonvalvular AF.⁴ pLAAO was noninferior to DOACs for the combined safety and efficacy composite endpoint (stroke, TIA, systemic embolism, cardiovascular death, major or nonmajor clinically relevant bleeding, or procedure-/device-related complications).

6.5.2. Cardiac Surgery—LAA Exclusion/Excision

Synopsis

Surgical left atrial appendage occlusion (S-LAAO) for reduction in the risk of recurrent arterial thromboemboli was first reported in 1949.⁶ After decades of relative clinical equipoise,^7,8 accumulating evidence,^2,8–10 culminating in the large RCT LAAOS III (Left Atrial Appendage Occlusion Study),¹ supports a benefit of S-LAAO in patients with AF who undergo^2,9 coronary artery bypass graft surgery (CABG) or valve surgeries. A 2019 meta-analysis before the publication of LAAOS III reviewed 22 studies, including 280 585 patients of whom 36 686 underwent S-LAAO during cardiac surgery. S-LAAO showed an association with a 29% lower risk of stroke or thromboembolism. The all-cause mortality rate at 2 years was lower in patients who underwent S-LAAO. No benefit was found in the subgroup of studies that included patients without preoperative AF.⁹ These favorable findings were validated by the 4770-subject LAAOS III, which confirmed S-LAAO in addition to OAC provides a 33% reduction in risk of stroke and systemic embolism.¹

LAAOS III enrolled only patients with preexisting AF; evidence of the effectiveness of S-LAAO in patients without AF is inconsistent and suggests the benefit of adjunctive S-LAAO is unclear in patients without preoperative AF.^7,9 The results of ongoing RCTs in this population are awaited.

Recommendation-Specific Supportive Text

1. LAAOS III was performed with 4770 patients with AF and a CHA₂DS₂-VASc score ≥2 undergoing cardiac surgery randomized to S-LAAO or no S-LAAO. Intraoperative transesophageal echocardiogram (TEE) was recommended to assess successful closure using an endorsed technique. Oral anticoagulation was continued postoperatively. After a mean follow-up of 3.8 years, the primary endpoint of stroke or systemic embolism occurred in 4.8% of the S-LAAO group and 7.0% of the no S-LAAO group (HR, 0.67 [95% CI, 0.53-0.85]; P=0.001). Mortality rate, cross-clamp and bypass time, chest tube output, and bleeding were not different between groups. More than 75% of subjects in both groups received OAC. These data indicate S-LAAO provides additional benefit to oral anticoagulation without increasing risk of adverse events.

   These findings cannot be extrapolated to patients without AF or patients who are not candidates for OAC. Most subjects underwent CABG or valve surgery, and the mean CHA₂DS₂-VASc score was 4.2. Surgical AF ablation was performed in one-third of both groups. Patients undergoing mechanical valve surgery, transplant, off-pump bypass, left ventricular assistive device surgery, and congenital heart disease (CHD) cases were excluded.

   A subsequent meta-analysis of 5 RCTs and 22 observational studies including 540 111 patients showed a significant decrease in stroke and thromboembolism with S-LAAO.¹¹ The postoperative mortality rate did not differ but, in follow-up, was reduced in the S-LAAO group after 2 years. No difference in major bleeding, all-cause rehospitalizations, or cross-clamp time was observed.¹¹

2. Substantial heterogeneity in S-LAAO methods exists in the literature, and not all techniques effectively exclude the LAA.⁴ An increased risk of stroke and thromboembolism is associated with incomplete S-LAAO^5,12; however, many studies do not report assessment of occlusion success.⁹ This inconsistency has likely contributed to the variability in outcome.^7,9 LAAOS III permitted 4 techniques: amputation and closure (which was promoted as the preferred technique), used in 56%; stapler closure, used in 11%; double-layer linear closure (if confirmed by TEE), used in 14%; or an approved LAAO device, used in 15%. Other techniques were approved on a case-by-case basis (4%). Intraoperative TEE was recommended to assess successful closure, defined as absence of flow across the suture line and a stump of <1 cm. If the initial closure was unsuccessful, additional suturing was performed for repair. Contemporary data show closure success of >95% using a surgical clip device.¹³ In the LAAOS series, the mean cross-clamp and bypass times were similar, and perioperative bleeding was not increased in the S-LAAO group.^1,4 These data support the importance of complete S-LAAO in achieving beneficial outcomes.

3. In LAAOS III, S-LAAO reduced stroke and thromboembolism by 33% in addition to OAC. LAAOS III was not designed to resolve whether oral anticoagulation can be safely stopped after S-LAAO; at hospital discharge, >80% of subjects in both groups were receiving OAC, with >75% on OAC at the 3-year visit.¹ A systematic review and meta-analysis focusing on the effects of anticoagulation found no statistically significant difference in stroke rates of S-LAAO/no S-LAAO subjects on OAC versus off OAC, with high heterogeneity noted for both ischemic events and mortality,¹⁴ and findings are subject to confounding by indication and the potentially beneficial effects of surgical AF ablation. The LAACS (Left Atrial Appendage Closure with Surgery) study of 187 subjects undergoing CABG, valve surgery, or both who were randomized to S-LAAO versus no S-LAAO showed a reduction in the risk of stroke in the S-LAAO group, independent of anticoagulation status.¹⁵ These results must be interpreted with caution, however, because the study was not powered to demonstrate a reduction in stroke, and substantial crossover occurred.

6.6. Active Bleeding on Anticoagulant Therapy and Reversal Drugs

Synopsis

About 2% to 4% of patients who receive OACs experience major bleeding and require intervention.¹³ Activated charcoal may be administered up to 6 to 8 hours after the last dose of an OAC.^14,15 Hemodialysis may be also considered to eliminate dabigatran but may be challenging and impractical due to coagulopathy and hemodynamic instability.¹⁶ The proportion of emergency department visits for bleeding from OACs due to DOACs increased from 2.3% in 2011 to 37.9% in 2017 because of the increased use of DOACs over the past decade.¹⁷ Some patients may require reversal agents to achieve rapid hemostasis. Failure to achieve effective hemostasis is associated with a >3 times higher risk of death.¹⁸ Idarucizumab for dabigatran-induced major bleeding and andexanet alfa for apixaban or rivaroxaban associated major bleeding are US Food and Drug Administration (FDA) approved. Four-factor PCC for DOAC-induced major bleeding was compared with andexanet alfa in retrospective cohort studies, but the comparison has never been investigated in clinical trials.⁷ For VKA, if life-threatening bleeding cannot be managed with supportive measures, the rapid reversal treatment with 4-factor PCC is preferred over fresh frozen plasma.^8–10Table 15 summarizes reversal agents for OACs. Table 16 describes bleeding events attributable to DOACs in pivotal clinical trials.

Recommendation-Supportive Text

1. The RE-VERSE AD (Reversal of Dabigatran Anticoagulant Effect With Idarucizumab) study was a multicenter, prospective cohort study evaluating the use of idarucizumab in patients receiving dabigatran for the management of uncontrolled or life-threatening bleeding or in patients requiring urgent surgery or invasive procedure.¹ In patients with uncontrolled or life-threatening bleeding, idarucizumab reversed the dabigatran anticoagulation effect in 100% of patients, and 67.7% had bleeding cessation. Thrombosis events after the reversal agent use was 4.7% within 30 days. However, this study did not have a control group. The post hoc analysis of the RE-VERSE AD study in patients with gastrointestinal bleeding found that 97.5% of patients achieved complete reversal of dabigatran, and bleeding cessation occurred in 76.2% of patients.² Another post hoc analysis of the same study showed that idarucizumab was effective at reversing dabigatran’s anticoagulation effect, regardless of baseline renal function.³

2. Compared with idarucizumab, the data on PCC for dabigatran-associated bleeding are very limited. A prospective multicenter cohort study compared the use of activated PCC in 14 patients with dabigatran-associated major bleeding compared with historical matched cases.⁴ The study was prematurely stopped due to the availability of idarucizumab in the market. The effectiveness of activated PCC was assessed based on the assessment guide rating from good, moderate, and poor. Nine patients (64%) were rated as good, 5 patients (36%) as moderate, and none as poor. No thromboembolic events were found. A randomized, placebo-controlled crossover trial in 12 healthy subjects evaluated the reversal effect of 4-factor PCC on rivaroxaban or dabigatran’s effect.⁵ It showed 4-factor PCC completely reversed rivaroxaban’s effect based the prothrombin time, but it did not reverse dabigatran’s effect based on the coagulation tests (aPTT, ecarin clotting time, and thrombin time). Four-factor PCC may not be appropriate for dabigatran-associated major bleeding management until further investigation is performed in prospective studies.

3. The ANNEXA 4 (Andexanet Alfa, a Novel Antidote to the Anticoagulation Effects of Factor Xa Inhibitors) trial evaluated the use of andexanet alfa in adult patients with acute major bleeding (safety population, n=254; efficacy population, n=254) receiving factor Xa inhibitors (apixaban 55%, rivaroxaban 36%, edoxaban 3%, and enoxaparin 6%).⁶ It found that after the bolus administration, the median anti-factor Xa activity decreased by 89%. Twelve hours after the andexanet infusion, clinical hemostasis was adjudicated as excellent or good in 37 of 47 patients in the efficacy analysis (79% [95% CI, 64%-89%]); however, thrombotic events occurred in 12 of 67 patients (18%) during the 30-day follow-up. Currently, andexanet alfa is FDA-approved for only apixaban- or rivaroxaban-associated life-threatening or uncontrolled bleeding, because only limited data are available for other factor Xa inhibitors. For 4-factor PCC, a meta-analysis of 22 observational studies compared andexanet alfa (n=438) with PCC (n=1278) in patients with acute bleeding associated with oral factor Xa inhibitors.⁷ It showed that mean hemostatic effectiveness for andexanet alfa was 82% at 12 hours and 71% at 24 hours versus 88% at 12 hours and 76% at 24 hours for PCC. Mean 30-day venous thromboembolism rate was 5.0% in the andexanet alfa group versus 1.9% in the PCC group. However, the meta-analysis had more significant methodological issues than the ANNEXA 4 trial. Thus, the level of evidence for the 4-factor PCC use was downgraded.

4. A meta-analysis of 13 studies (5 randomized trials and 8 observational studies) compared 4-factor PCC with fresh frozen plasma for warfarin-associated bleeding or urgent surgery/procedure.⁸ Four-factor PCC use was significantly associated with lower risk of all-cause mortality (25.1% versus 28.8%; OR, 0.56 [95% CI, 0.37-0.84]), higher probability to achieve INR correction (60.6% versus 12.8%; OR, 10.8 [95% CI, 6.12-19.07]), shortened time to INR correction (–6.5 h [95% CI, –9.75 h to –3.24 h]), and lower risk of volume overload (2.1% versus 8.8%; OR, 0.27 [95% CI, 0.13-0.58]) compared with fresh frozen plasma. No significant difference in thromboembolic events (4.2% versus 4.8%; OR, 0.91 [95% CI, 0.44-1.89]; P=0.81) was found.

5. A meta-analysis of 10 cohort studies compared anticoagulation resumption (n=2080) versus discontinuation (n=2296) after gastrointestinal bleeding in patients receiving OACs for AF, venous thromboembolism, or prosthetic valve.¹¹ Restarting anticoagulation was significantly associated with higher risks of recurrent gastrointestinal bleeding (10.1% versus 5.3%) but also with lower risks of thromboembolic events (6.3% versus 10.6% ) and a reduction in all-cause mortality (21.3% versus 31%). Another meta-analysis of 7 observational studies in patients with AF found no significant difference in stroke rates between resumption and discontinuation groups (9.1% versus 8.3%; OR, 0.75 [95% CI, 0.37-1.51]).¹² Resumption of anticoagulation therapy was associated with significant reduction in any thromboembolic events (8.0% versus 12.2%; OR, 0.54 [95% CI, 0.43-0.68]) and all-cause mortality (10.8% absolute risk reduction; OR, 0.38 [95% CI, 0.24-0.60]), at the expense of recurrent major bleeding (10.2% versus 5.0%). One study found the optimal timing of resuming a VKA after upper gastrointestinal bleeding was between 3 and 6 weeks, while other studies revealed that resuming a VKA 1 week after gastrointestinal bleeding is the optimal timing.^19–21 There is a paucity of data evaluating the timing of DOACs after gastrointestinal bleeding.

6.6.1. Management of Patients With AF and ICH

Synopsis

ICH is a term that refers to bleeding within any of the compartments of the cranial vault, including intraparenchymal (or intracerebral), subarachnoid, subdural, and epidural hemorrhage. In clinical practice, the different forms of ICH are often classified by mechanism and most broadly as traumatic and nontraumatic or spontaneous ICHs (Figure 14). In patients with AF and conditions associated with high-risk for thromboembolic complications (eg, mechanical valve, rheumatic valvular disease), anticoagulation is often resumed early after ICH, regardless of mechanism. For other patients with AF, decisions regarding if (and when) to resume anticoagulation after ICH requires more careful risk-benefit assessment of patient-specific clinical factors. Patients with traumatic etiologies generally have lower long-term risk of recurrent hemorrhage, and anticoagulation is generally considered safe to resume. Nontraumatic/spontaneous etiologies (eg, hypertensive parenchymal hemorrhage, hemorrhagic transformation after ischemic stroke, cerebral amyloid angiopathy-associated hemorrhage, chronic subdural hemorrhage) are generally associated with higher risk of recurrence. Clinical decision-making regarding the appropriateness of resumption in these patients is based largely on individual risk-benefit calculations given the lack of randomized data assessing safety and long-term outcomes. For patients with ICH deemed at high risk of recurrence, LAA closure may be a viable alternative to anticoagulation, although data on efficacy and safety are lacking in the ICH population.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Recommendation-Specific Supportive Text

1. Rheumatic valvular disease is associated with a >5% per year risk of embolic events.⁸ In patients with mechanical heart valves, the rate can range from 4% to 23% per year depending on the type and position of the prosthetic.^9–11 These rates are even higher in patients with concurrent AF but are significantly reduced with the use of VKAs. Data are limited that explore the safety and timing of anticoagulation resumption in patients with ICH and high-risk valvular disease. In 1 observational analysis of 137 patients with ICH and mechanical heart valves, resuming anticoagulation within 2 weeks from index ICH was associated with increased hemorrhagic complications (HR, 7.06 [95% CI, 2.33-21.37]).¹ However, in balancing the risk of hemorrhagic with thrombotic complications, the optimal window of resumption was found to be between 1 and 2 weeks post-ICH after weighing individual patient factors (Table 17).

2. Anticoagulation is associated with up to 25% of ICHs and is associated with worse functional outcomes and higher mortality.^12–15 After ICH, anticoagulation is generally held in the acute setting. Decisions regarding the appropriateness and timing of resumption are complex given the limited amount of randomized data addressing risks and benefits in this population. Previous observational reports evaluating risk of recurrent ICH, thromboembolism, and mortality in patients with nonvalvular AF have suggested possible reduction in thromboembolic risk and mortality with resumption of anticoagulation but are limited by discordant patient populations and study designs, predominant use of VKAs, and potential selection biases.^16–22 In a meta-analysis of 2452 patients with AF and ICH, resumption of OACs was associated with lower risk of ischemic stroke (RR, 0.46 [95% CI, 0.29-0.72]) and no difference in risk of recurrent ICH compared with patients in whom anticoagulation was not resumed² (Table 17). However, data are limited on the optimal timing for resumption. In an analysis of 2619 ICH survivors with AF, a composite net benefit minimizing risk of bleeding and thromboembolic complications occurred when anticoagulation was resumed at 7 to 8 weeks after ICH.³ In a retrospective study that included 1752 patients with OAC-related ICH across 3 Danish registries, anticoagulation resumed a median of 34 days was associated with an adjusted HR of ischemic stroke, systemic embolism, and all-cause mortality of 0.55 (95% CI, 0.39-0.78) compared with no anticoagulation.⁴ In another meta-analysis of 8 observational studies, resuming OAC was associated with reduced risk of ischemic stroke without increased risk of recurrent ICH at a median time to resumption of 10 to 39 days⁵ (Table 17). Of note, DOACs have been associated with lower rates of major bleeding complications compared with VKAs.²³

3. Anticoagulation is used for stroke risk reduction in patients with AF at risk of thromboembolism. For some patients, however, risk of recurrent ICH may be prohibitively high with anticoagulation. Cerebral amyloid angiopathy is the most common cause of lobar hemorrhage in older adults and is associated with annual ICH recurrence risk that ranges from 5% in patients with isolated microhemorrhages to 26.9% in patients with associated cortical superficial siderosis.²⁴ Thus, in patients with AF and cerebral amyloid angiopathy-related ICH, alternative strategies for secondary stroke prevention are often pursued. The LAA is considered the most common site for thrombus formation in patients with AF.^25,26 Two RCTs—PROTECT AF study and PREVAIL study—compared rates of stroke and mortality after LAA closure versus VKA in patients with AF.^6,7 In a meta-analysis examining 5-year outcomes in the 2 RCTs, LAA closure was associated with significant reductions in hemorrhagic stroke, disabling and fatal stroke, and all-cause mortality, with comparable rates of stroke and systemic embolism compared with warfarin.²⁷ Of note, both studies excluded patients who were not candidates for anticoagulation. Additionally, newer treatment options such as epicardial LAA clipping are emerging for stroke prevention without need for postprocedure APT.^28,29

6.7. Periprocedural Management

Synopsis

Periprocedural anticoagulation management is driven by 2 competing aims: to hold the anticoagulant agent for the shortest possible time to minimize the risk of thromboembolism and to ensure coagulation parameters are as close to normal as possible at the time of the procedure to facilitate hemostasis and avoid intra- and postoperative bleeding. Considerations include the bleeding risk of the procedure, consequences of bleeding should it occur, patient-specific risk factors, and thrombotic risk while off anticoagulation. Invasive procedures and surgeries have been divided into high and low bleeding risks in several different classifications.¹⁶ Generally, endoscopic, dental extraction, many ophthalmologic procedures, and percutaneous vascular access, such as cardiac catheterization, are considered to be of low bleeding risk. Higher bleeding risk surgeries include intra-abdominal, pelvic, orthopedic, neurosurgical, cardiac, and transvenous lead extraction procedures. Very high bleeding risk procedures include neuraxial anesthesia and spinal surgery. Available data do not suggest high rates of bleeding from these procedures, although bleeding could cause significant consequences.¹⁷ Patient-specific factors include abnormal liver and kidney function, bleeding or clotting disorders, concomitant antiplatelet or nonsteroidal anti-inflammatory use, alcohol, and anemia. Recent bleeding or thromboembolism may increase the risk for recurrent events. To put thrombotic risk in perspective, the incidence of thromboembolism while anticoagulation was temporarily held has generally been <1%.^3,18

Recommendation-Specific Supportive Text

1. To mitigate thromboembolic risk during interruption of warfarin therapy for a procedure or surgery, “bridging” anticoagulation can be administered. However, observational studies consistently showed an increased risk of bleeding without a difference in thromboembolic risk with this strategy.³ The BRIDGE (Bridging Anticoagulation in Patients who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery) trial included patients with a mean CHADS₂ score of 2.3 (range, 1-6); 61.7% of patients had a CHADS₂ score <3.¹ In this study, discussed in more depth below, interruption of warfarin without bridging anticoagulation was superior to bridging.¹ In a meta-analysis of 6 observational studies reporting on patients not at high thromboembolic risk receiving prophylactic dose or no bridging had an overall thromboembolic event rate of 0.6% (11 of 1702). ³ A similar RCT has not been done in the DOAC era, given the ease of brief interruption of DOAC therapy; however, observational studies have shown similar results.² The PAUSE (Perioperative Anticoagulation Use for Surgery Evaluation) cohort study, discussed below, enrolling patients with a mean CHADS₂ score of 2.1 (range, 1-6) and CHA₂DS₂-VASc score of 3.5 (range, 1-9), found a low incidence of thromboembolic events when DOAC was briefly interrupted without bridging.⁴ This suggests that in patients at low-to-moderate thromboembolic risk, interruption of anticoagulation, without bridging, may provide a low risk of bleeding with an acceptable risk of stroke/TIA. Patients with mechanical valves, and recent stroke/TIA, or other high-risk markers, were not included in most studies and may still benefit from bridging anticoagulation: in these scenarios, management should be individualized.

2. Evidence has accrued that periprocedural unfractionated or low-molecular-weight heparin bridging has been associated with an increased risk of pocket hematoma after pacemaker or ICD procedures.¹⁹ BRUISE-CONTROL (Bridge or Continue Coumadin for Device Surgery Randomized Controlled Trial) randomized patients with AF (88%) or a mechanical valve with an estimated annual stroke risk of ≥5%, and undergoing elective device surgery, to continued warfarin versus warfarin interruption with bridging anticoagulation. The incidence of clinically significant pocket hematoma was >4 times higher in the bridging group.⁵ Several other procedures have been performed on uninterrupted anticoagulation.²⁰ Further study is needed to define which procedures, in which patients, can safely be performed with uninterrupted anticoagulation. Lead extraction procedures are generally considered high bleeding risk and therefore should be treated as such (high bleeding risk procedures in Table 18) and warfarin interrupted, with occasional exceptions.

3. Two RCTs and several observational studies have shown little or no difference in the risk of pocket hematoma and thromboembolic events in patients with AF anticoagulated with DOACs and undergoing device surgery.^8–10 Therefore, brief interruption of DOAC therapy (low bleeding risk procedures in Table 18 may be preferable for many, while uninterrupted DOAC may be preferred for those at particularly high thromboembolic risk. Lead extraction procedures are generally considered high bleeding risk and therefore should be treated as such (high bleeding risk procedures in Table 18) and DOAC interrupted, with occasional exceptions.

4. To achieve the goals of reducing or eliminating anticoagulant effect at the time of surgery, while also minimizing the time the patient spends without anticoagulation to prevent thromboembolism, the duration of cessation of anticoagulation should take into account both the pharmacokinetics and pharmacodynamics of the agent, the patient’s renal function, and the bleeding risk of the procedure. Direct measurement of DOAC levels is unavailable in clinical practice, but levels <30 ng/mL, which correspond to the expected plasma DOAC concentration reached after 3 half-lives, when most of the drug (87.5% of the Cmax) has been eliminated, is reached in two-thirds of patients 24 to 48 hours after the last drug intake.¹¹ Similar results were observed for apixaban.¹² A drug “hold” of 48 hours in clinical practice typically means 2 full days without drug intake, in addition to the day of the procedure, leading to a longer interval between last drug intake and procedure start longer than 48 hours, usually an additional 12 hours. Meta-analysis of the reports of periprocedural events from the pivotal trials of each of the 4 available DOACs suggested similar outcomes in patients on DOAC or warfarin, with reduced major bleeding during an uninterrupted strategy with DOACs.²¹ The PAUSE protocol was studied in patients on apixaban, dabigatran, and rivaroxaban, and resulted in low rates of major bleeding (0.9%-1.85%) and thromboembolism (<1%). A high rate (>90%) of minimal anticoagulant level (<50 ng/mL) was achieved. Similar results have been observed with edoxaban in clinical practice.²² Different protocols have not been directly compared. Table 18 presents recommended DOAC interruption times based on agent, renal function, and the bleeding risk of the procedure.

5. The timing of resumption of interrupted anticoagulation after surgery is a complex decision that should integrate patient and surgical characteristics. Warfarin, which takes several days (usually 3-5) to become therapeutic, can often be restarted the evening of the procedure. Postoperative bridging anticoagulation has not been shown to be generally beneficial.^1,13 In the PAUSE study, DOACs were started on the first postoperative day for low bleeding risk procedures and between the evening of the second and third postoperative days for high bleeding risk procedures. This resulted in a low risk of major bleeding.⁴ Input from the proceduralist should always be incorporated into decisions.

6. Several studies have compared a strategy of bridging anticoagulation with a comparator of no bridging in patients on anticoagulants who were scheduled to undergo an invasive procedure or surgery. The BRIDGE trial randomized 1 884 patients with AF or AFL to bridging anticoagulation before and after the procedure with dalteparin versus placebo.¹ Almost 90% of patients were undergoing low bleeding risk procedures. No bridging was superior to bridging with respect to major bleeding and was noninferior with respect to thromboembolism, although the observed rate (0.4%) was lower than anticipated (1.0%). PERIOP2 (A Double Blind Randomized Control Trial of Post-Operative Low Molecular Weight Heparin Bridging Therapy Versus Placebo Bridging Therapy for Patients Who Are at High Risk for Arterial Thromboembolism) randomized patients, about two-thirds of whom were undergoing high bleeding risk procedures, and 79% of whom had AF, to postoperative bridging with dalteparin or placebo until the INR was >1.9. All patients received preoperative bridging. Bleeding and thromboembolism outcomes were similar.¹³ These studies also excluded patients with recent bleeding or thromboembolism, and those undergoing spinal and neurological surgery, and in these cases, management should be individualized and may include bridging anticoagulation.^1,13 A small randomized trial found that dental extractions could be performed on uninterrupted warfarin without bridging.¹⁴ A large multicenter registry also did not find a difference in thromboembolism, but an increase in bleeding, in bridged patients.¹⁵ Patients with mechanical valves should be managed as outlined in the “2020 ACC/AHA Guideline for the Management of Patients With Valvular Heart Disease.”²³

6.8. Anticoagulation in Specific Populations

6.8.1. AF Complicating ACS or Percutaneous Coronary Intervention (PCI)

Synopsis

Encountering a patient with AF at an increased risk for stroke who undergoes PCI for stable CAD or ACS is a common clinical scenario. The addition of APT to OAC after PCI introduces a risk for clinically significant bleeding. RCTs have compared the safety and efficacy of various antithrombotic regimens after PCI is performed in patients with AF. They have consistently shown that DOACs, when used in conjunction with APT, are associated with a lower risk of clinically relevant bleeding compared to VKAs.^1–5 They have also shown that dual therapy with an OAC and single APT with a P2Y12 inhibitor significantly lowers the risk of clinically relevant bleeding compared with triple therapy (oral anticoagulation, P2Y12 inhibitor, and aspirin) without substantially increasing risk of major adverse cardiovascular events.^1–4,6 Meta-analyses of these trials have shown a slightly increased risk of stent thrombosis with dual therapy compared to triple therapy.⁵ Thus, the risk of stent thrombosis must be weighed against the risk of bleeding in determining the antithrombotic regimen after PCI. This is particularly true in patients with a relatively high perceived risk of stent thrombosis (complex revascularization, multivessel PCI, previous history of stent thrombosis).

Recommendation-Specific Supportive Text

1. Post-PCI antithrombotic regimens comparing DOACs with VKAs in combination with APT in patients with AF who do not have a specific indication for VKAs (eg, mechanical valve) have consistently shown a reduction in clinically relevant bleeding.^1,2,4 A meta-analysis of 3 RCTs comparing DOACs with VKAs after PCI, PIONEER AF-PCI (Open-Label, Randomized, Controlled, Multicenter Study Exploring Two Treatment Strategies of Rivaroxaban and a Dose-Adjusted Oral Vitamin K Antagonist Treatment Strategy in Subjects with Atrial Fibrillation who Undergo Percutaneous Coronary Intervention) (rivaroxaban), RE-DUAL PCI trial (Randomized Evaluation of Dual Antithrombotic Therapy with Dabigatran versus Triple Therapy with Warfarin in Patients with Nonvalvular Atrial Fibrillation Undergoing Percutaneous Coronary Intervention) (dabigatran), and AUGUSTUS (Open-Label, 2×2 Factorial, Randomized, Controlled Clinical Trial to Evaluate the Safety of Apixaban vs Vitamin K Antagonist and Aspirin vs Aspirin Placebo in Patients With Atrial Fibrillation and Acute Coronary Syndrome and/or Percutaneous Coronary Intervention) (apixaban), showed a 42% relative risk reduction for major bleeding in DOAC-based regimens compared with VKA-based regimens (OR, 0.577 [95% CI, 0.477-0.698]; P<0.001).⁵ There was no statistically significant difference in the risk of ischemic events between groups. The ENTRUST-AF PCI (Edoxaban Treatment vs Vitamin K Antagonist in Patients With Atrial Fibrillation Undergoing Percutaneous Coronary Intervention) trial showed that edoxaban-based regimens were noninferior to VKA-based regimens in terms of clinically relevant bleeding after PCI.³ An important limitation of the aforementioned data is that most studies compared DOACs monotherapy against VKA plus aspirin, the only VKA-based dual antithrombotic therapy (DAT) regimen was available in AUGUSTUS trial, thus providing limited information in comparison to DOAC-based DAT regimens available in all trials.⁵

2. RCTs have supported DAT with OAC and a P2Y12 inhibitor over triple therapy with OAC, P2Y12 inhibitor, and aspirin to reduce the risk of clinically relevant bleeding in patients with AF after PCI, including WOEST (What Is the Optimal Antiplatelet and Anticoagulant Therapy in Patients with Oral Anticoagulation and Coronary Stenting) (VKA), PIONEER AF-PCI (rivaroxaban), RE-DUAL (dabigatran), and AUGUSTUS (apixaban).^1,2,4,6 A meta-analysis of PIONEER AF-PCI, RE-DUAL, and AUGUSTUS demonstrated a 40% relative risk reduction for International Society on Thrombosis and Haemostasis major bleeding in dual therapy regimens compared with triple therapy (OR, 0.598 [95% CI, 0.491-0.727]; P<0.001).⁵ Patients in the intervention group received varying duration of triple therapy with aspirin after PCI up to 7 days. This meta-analysis showed that DAT with DOAC and P2Y12 inhibitor was associated with an increased risk of stent thrombosis (OR, 1.672 [95% CI, 1.022-2.733]; P=0.041) compared with triple therapy.⁵ This finding was mainly driven by outcomes in the 110-mg dosing group of dabigatran in the RE-DUAL trial. The number needed to treat for International Society on Thrombosis and Haemostasis major bleeding (n=42) was substantially lower than the number needed to harm for stent thrombosis (n=223) with DAT compared with triple therapy.⁵ Based on these findings, extended triple therapy up to 30 days after PCI is a consideration when the perceived risk of stent thrombosis is high (eg, complex revascularization, multivessel PCI, previous history of stent thrombosis).

6.8.2. Chronic Coronary Disease (CCD)

Synopsis

APT is the first-line therapy to prevent recurrent cardiovascular events in patients with CAD, while anticoagulation therapy is recommended to prevent stroke or systemic embolism in patients with AF at increased risk of thromboembolism. Minimal evidence was available until recently if a combination therapy of APT in addition to OAC therapy provides additional value to prevent cardiovascular events in patients with AF and concomitant CCD (at least 12 months after revascularization or CAD not requiring coronary revascularization) compared with oral anticoagulation monotherapy. A recent landmark clinical trial showed that rivaroxaban monotherapy significantly decreased the primary composite outcome (stroke, systemic embolism, MI, unstable angina requiring revascularization, or death from any cause) and decreased the risk of major bleeding compared with the combination therapy.¹ The subgroup analysis of the previously mentioned clinical trial in patients with AF only after coronary stenting and another clinical trial corroborated the main findings in a similar population.^2,3

Recommendation-Specific Supportive Text

1. An RCT compared the rivaroxaban monotherapy (n=1107) with the combination of rivaroxaban and a single APT (n=1108) in patients with AF and CCD or CAD not requiring revascularization.¹ The primary efficacy outcome was the composite of stroke, systemic embolism, MI, unstable angina requiring revascularization, or death from any cause. Rivaroxaban monotherapy was noninferior to dual therapy for the primary outcome (4.14% versus 5.75%/patient-year; HR, 0.72 [95% CI, 0.55-0.95]) and was superior for primary safety endpoint of major bleeding (1.62% versus 2.76%/patient-year; HR, 0.59 [95% CI, 0.39-0.89]) compared with combination therapy. Another clinical trial also showed the composite outcome of death from all causes, MI, stroke, or systemic embolism (15.7% versus 13.6%; HR, 1.16 [95% CI, 0.79-1.72]; P=0.20 for noninferiority); the major bleeding risk in the oral anticoagulation monotherapy group was numerically lower than the combination therapy group (7.8% versus 10.4%; HR, 0.73 [95% CI, 0.44-1.20]).² However, this trial was terminated prematurely because of slow enrollment and did not have enough power for noninferiority; thus, it was inconclusive. Of note, both clinical trials excluded patients with history of stent thrombosis. Further investigation of oral anticoagulation therapy in patients with a history of stent thrombosis is needed.

6.8.3. Peripheral Artery Disease (PAD)

Synopsis

Antiplatelet monotherapy is the standard of care for patients with symptomatic PAD and reasonable in asymptomatic PAD with ankle-brachial index ≤0.90 to reduce the risk of MI, stroke, and vascular death.⁶ The 2016 AHA/ACC PAD guidelines recommend against anticoagulation for ischemic event risk reduction in the general PAD population (COR 3; LOE A) due to increased risk of bleeding without significant reduction in ischemic events compared with antiplatelet monotherapy,^6,7 without making specific recommendations for patients with comorbid PAD and AF. Conversely, anticoagulation is recommended for patients with AF at intermediate-to-high risk of stroke due to the inferiority of APT to prevent stroke and systemic thromboembolism in this population. Subgroup analysis of landmark DOAC clinical trials demonstrate comparable efficacy and safety of DOACs compared with warfarin among patients with AF with and without PAD.^2–4 Clinical trial evidence is lacking to guide decisions regarding DAT in the setting of concomitant AF and PAD (whether stable PAD or after lower extremity revascularization). Observational studies of patients with concomitant AF and PAD treated with OAC monotherapy compared with dual therapy (OAC plus antiplatelet) found an increased risk of bleeding on dual therapy without reduction in major adverse cardiac events.^5,8 However, an important limitation of these studies was the lack of PAD-specific outcomes such as acute or critical limb ischemia, revascularization, or amputation as primary or secondary outcomes.^5,8

Recommendation-Specific Supportive Text

1. In the WAVE (Warfarin Antiplatelet Vascular Evaluation) trial of patients with PAD but without AF, dual therapy (antiplatelet plus warfarin) did not reduce the primary outcome of MI, stroke, cardiovascular death (RR, 0.92 [95% CI, 0.73-1.16]; P=0.48) compared with monotherapy (antiplatelet alone), and dual therapy increased the risk of life-threatening bleeding (RR, 3.41 [95% CI, 1.84-6.35]; P<0.001).⁷ Meta-analysis of 3 RCT subgroup analyses (ENGAGE AF-TIMI 48, ARISTOTLE, ROCKET AF; total N=2 564) compared safety and efficacy outcomes between warfarin and DOACs in patients with concomitant AF and PAD.^1–4 The meta-analysis demonstrated similar efficacy for DOACs compared with warfarin, including stroke and systemic embolism (RR, 0.93 [95% CI, 0.61-1.42]; P=0.73), all-cause death (RR, 0.91 [95% CI, 0.70-1.19]; P=0.50), and MI (RR, 1.10 [95% CI, 0.64-1.90]; P=0.74), without a statistically significant difference in the risk of major bleeding (RR, 1.12 [95% CI, 0.70-1.81]; P=0.63) or ICH (RR, 0.54 [95% CI, 0.16-1.85]; P=0.33).¹ In ORBIT-AF II (The Outcomes Registry for Better Informed Treatment of Atrial Fibrillation II),⁵ among AF patients with concomitant vascular disease (CAD or PAD), major adverse cardiac or neurological events and bleeding events were compared between patients on OAC monotherapy versus dual therapy (OAC plus APT). Major adverse cardiac or neurological events (HR, 1.5 [95% CI, 1.00-2.25]) did not differ significantly between mono- and dual therapy, while dual therapy was associated with an increased risk of bleeding (HR, 2.27 [95% CI, 1.38-3.73]; P=0.001).⁵

6.8.4. Chronic Kidney Disease (CKD)/Kidney Failure

Synopsis

CKD is a risk factor for stroke in patients with AF, independent of other risk factors.⁸ The prevalence of AF among patients with renal failure on dialysis is high, including asymptomatic AF.⁹ Despite compelling evidence in the general population, conflicting data exist as to whether AF is a risk factor for stroke in patients on dialysis.^10,11 Potential explanations for this discrepancy include competing risks of stroke and mortality, a high prevalence of undiagnosed AF in cohorts thought not to have AF, and the confounding effect of anticoagulation thrice weekly during hemodialysis.¹² VKAs are limited by a markedly reduced time in therapeutic range in patients with severe CKD, including those on dialysis,¹³ and association with calciphylaxis.¹⁴ The pivotal DOAC trials excluded patients on dialysis, and recommendations for dosing are based on pharmacokinetic data only. Early experience with dabigatran and rivaroxaban in this population suggested an increased risk of bleeding relative to warfarin, driven by the higher dose of each.¹⁵ Only dabigatran is substantially removed by dialysis.¹⁶ A small RCT and a retrospective population-based cohort study found a lower risk of thromboembolism with rivaroxaban versus warfarin,^17,18 but these findings await confirmation in larger studies. Limited data on LAAO exist in this group of patients, and prospective comparison to anticoagulation or no therapy is awaited.¹⁹

Recommendation-Specific Supportive Text

1. Early trials demonstrating the role of warfarin anticoagulation in preventing stroke did not enroll or separately report outcomes in patients with CKD. An analysis of the SPAF III trial showed that dose-adjusted warfarin, compared with fixed low-dose warfarin plus aspirin, reduced the risk of stroke with a similar magnitude to patients with normal kidney function.³ In the AVERROES trial of apixaban versus aspirin in patients deemed unsuitable for warfarin, apixaban showed similar efficacy in patients with an estimated glomerular filtration rate (eGFR) of 25 to 50 mL/min as those with an eGFR >50 mL/min.²⁰ Patients with CKD stage 3, which includes those with an eGFR of 30 to 60 mL/min, were enrolled in the pivotal trials of the DOACs and showed similar results compared with warfarin as seen in patients with normal kidney function.²

2. Evidence on the use of anticoagulation in patients with AF and stage 4 CKD (eGFR, 15-30 mL/min/1.73 m²) is largely observational. A prospective, multicenter cohort study of survivors of MI with AF found that warfarin prescription at discharge associated with a reduced risk of death, readmission due to MI, ischemic stroke, and bleeding.⁵ An analysis of patients with CrCl 25 to 30 mL/min at the time of enrollment in the ARISTOTLE trial found numerically fewer stroke and major bleeding events with apixaban than warfarin.⁴ Because CrCl overestimates eGFR, a proportion of the patients enrolled in the pivotal clinical trials of DOACs, which used CrCl cut-offs for enrollment, had CKD stage 4 as assessed by eGFR in current clinical practice.

3. The role of anticoagulation in patients with AF and severe CKD, including those on dialysis, is controversial.^6,7,21,22 Pending the results of RCTs comparing warfarin with no treatment, and warfarin with DOACs, SDM incorporating risks of bleeding and thromboembolism is encouraged. The pivotal clinical trials of apixaban excluded patients with CrCl <25 mL/min.^23,24 Subsequently, the FDA-approved labeling was extended to include those with end-stage renal failure, including those on hemodialysis.²⁵ A subsequent pharmacokinetic study found that drug levels with apixaban 2.5 mg twice daily approximate those seen with 5 mg twice daily in patients with normal renal function, suggesting 2.5 mg is a more appropriate dose in such patients.²⁶ A large retrospective cohort study of Medicare beneficiaries found no difference in thromboembolism, but less major bleeding, in patients treated with apixaban versus warfarin.²⁷ Another analysis found an increased risk of thromboembolism and of fatal or intracranial bleeding with apixaban 5 mg twice daily versus no anticoagulation but not with apixaban 2.5 mg twice daily.²⁸ Two small trials of apixaban versus warfarin in patients with AF on hemodialysis could not show differences in safety or efficacy outcomes but significantly more bleeding than stroke events.^29,30

6.8.5. AF in VHD

Synopsis

At least one-third of patients with AF have some degree of VHD. Patients with AF and VHD have a higher prevalence of stroke and systemic thromboembolism than those patients with AF without VHD.⁹ Patients with VHD and AF should be evaluated for risk of thromboembolic events and treated with oral anticoagulation if they are at high risk or if they have significant mitral stenosis or a mechanical heart valve. VKAs are the anticoagulation drugs of choice for patients with rheumatic mitral stenosis, mechanical heart valves, and new-onset AF within 3 months after mechanical aortic or mitral valve surgery.¹⁰

Recommendation-Specific Supportive Text

1. Historical data have reported that in patients with mitral stenosis and AF, stroke risk was nearly 18-fold higher than an age-, sex-, and hypertension-matched population without AF.¹¹ More recently, stroke or systemic embolism rates ranged between 0.4 and 4 per 100 patient-years among anticoagulated patients and were highest in those with previous embolism. Warfarin is generally prescribed, but it has been questioned whether a DOAC may be an alternative. A retrospective cohort study from Korea examined 2230 individuals with mitral stenosis and AF who were prescribed either warfarin or a DOAC, using propensity matching on 10 clinical variables. Thromboembolic events in the DOAC group were 2.22% per year compared with 4.19% per year in the warfarin group (HR, 0.28 [95% CI, 0.18-0.45]).¹² However, in the INVICTUS (The Investigation of Rheumatic AF Treatment Using VKAs, Rivaroxaban or Aspirin Studies) RCT,¹ patients with AF, rheumatic heart disease, and mitral stenosis had a mean survival time to a primary outcome event of stroke, systemic embolism, MI, or death from vascular or unknown cause of 1 675 days if treated with a VKA compared with 1 599 days if with rivaroxaban (difference, –76 days [95% CI, –121 to –31]; P<0.001).

2. Although patients with moderate to severe mitral stenosis or mechanical valve were excluded from DOAC trials, other forms of VHD were allowed, such as aortic stenosis or mitral regurgitation. Bioprosthetic valves and valve repair were allowed in the edoxaban (ENGAGE AF) and apixaban (ARISTOTLE) trials, and valve repair in the rivaroxaban (ROCKET AF) trial. A systematic review¹³ of patients with VHD (other than moderate to severe mitral stenosis or mechanical valve) concluded that DOACs were safe. A meta-analysis confirmed that DOACs decreased the risk of stroke/systemic embolism compared with warfarin in patients with (HR, 0.70 [95% CI, 0.58-0.86]) and without VHD (HR, 0.84 [95% CI, 0.75-0.95]).¹⁴ However, for patients with mechanical heart valves,¹⁵ the RE-ALIGN (Randomized, Phase II Study to Evaluate the Safety and Pharmacokinetics of Oral Dabigatran Etexilate in Patients After Heart Valve Replacement)¹⁰ trial was halted during phase II after enrolling 252 subjects due to increased thromboembolism and bleeding in the dabigatran arm. PROACT Xa (A Trial to Determine if Participants with an On-X Aortic Valve Can be Maintained Safely on Apixaban) was also terminated early due to higher thromboembolic events with apixaban in patients who were >3 months after mechanical On-X aortic valves.

6.8.6. Anticoagulation of Typical AFL

Synopsis

AFL is a common atrial arrhythmia with a reported overall incidence of 88 per 100 000 person-years, and the incidence increases with age.¹⁴ AFL is 2.5 times more common in men than in women, and it is significantly more likely to occur in patients with underlying HF or COPD.¹⁴

Because of the high success rate and low recurrence rate of typical AFL after CTI ablation, catheter ablation is often used as a first-line treatment. The high effectiveness of CTI ablation may decrease thromboembolic risk so that many physicians may elect to discontinue oral anticoagulation >1 month after ablation in the absence of previously detected AF. However, the high incidence of new onset AF at some time after CTI ablation places this practice in question. Because AF may be subclinical or asymptomatic, patients could be at significant risk of thromboembolic events after ablation of typical AFL depending on thromboembolic risk factors.

Catheter ablation of non–CTI-dependent AFL is technically more difficult, and “atypical” AFL or AT often occurs after AF ablation. These arrhythmias should be anticoagulated and managed in a manner similar to AF. Recommendations in this section refer to treatment of “typical” AFL.

Recommendation-Specific Supportive Text

1. AFL is associated with increased thromboembolic risk, but the exact risk and benefit of oral anticoagulation are difficult to ascertain as AFL and AF often coexist. In a population-based retrospective cohort study of patients with typical AFL and no history of AF, stroke occurred in 4.1% of patients with AFL compared with 1.2% of a general population-matched cohort (P<0.001).¹ However, no large, randomized trials specifically address thromboembolic risks and benefits of anticoagulation in patients who have only AFL.

   In patients referred for ablation of typical AFL without adequate previous appropriate anticoagulation, and no previous history of AF, TEE showed evidence for LA thrombus or dense spontaneous echocardiographic contrast in 3%.² A systematic review identified thromboembolic event rates of 0% to 6% with cardioversion, with the highest incidence of thromboembolic events occurring 1 to 2 days postcardioversion of AFL.³ Echocardiographic studies reported LA thrombus in 0% to 38% and spontaneous echocardiographic contrast in 21% to 28% of cases.³ In 1 ablation study, thromboembolic events occurred in 13.9% of patients who did not receive anticoagulation.³ Observational studies report an elevated risk of stroke with AFL compared with a control group (RR, 1.4 [95% CI, 1.35-1.46]) and elevated mortality risk (risk ratio, 1.9 [95% CI, 1.2-3.1]) with long-term follow-up.³

2. The incidence of embolism after cardioversion of AFL is similar to that of AF (0.72% versus 0.46%; P=not significant).⁴ Atrial stunning has been noted on TEE in patients undergoing cardioversion of AF, with partial atrial recovery 15 to 30 days and full recovery 30 to 90 days postcardioversion.¹⁵ LAA stunning also occurs in patients with AFL, although it appears to occur at a lesser degree than what is seen with AF.¹⁶ Although limited data are available for AFL, 1 study demonstrated more forceful mechanical LAA contraction and absence of spontaneous echocardiographic contrast 2 weeks after ablation of persistent AFL, suggesting earlier recovery.¹⁷ Because LA and LAA stunning are thought to contribute to thrombus formation and thromboembolic events after cardioversion, continued anticoagulation is recommended early after conversion to sinus rhythm.

3. In a meta-analysis, the subsequent incidence of AF after ablation of CTI-dependent AFL was 34% over 14 months. However, the incidence of AF after flutter ablation is significantly higher in patients with a history of AF before ablation than those without AF before ablation (53% versus 23% after 16-18 months; P<0.05).⁵ The incidence of AF continues to increase over time in both patients with and without previous AF.⁵ Oral anticoagulation reduces risk of stroke or systemic thromboembolism in the setting of nonvalvular AF.⁶ Because AF is not necessarily reduced after ablation of AFL, continued anticoagulation is recommended based on thromboembolic risk assessment.

   The acute success rate for catheter ablation typical AFL is reported to be 92% with a single procedure and 97% with multiple procedures in a meta-analysis.¹⁸ However, the subsequent incidence of AF occurring after ablation of CTI-dependent AFL is reported to be 16% to 82%,^5,10,11 with a higher incidence in patients with a history of AF before ablation than those without (53% versus 23% after 16-18 months; P<0.05).⁵ In 1 study, 82% of patients who underwent typical AFL ablation experienced new-onset AF during long-term follow-up (mean, 39 months).¹¹

4. Because of the high rate of occurrence of new-onset AF after CTI ablation, close follow-up and monitoring are recommended. In patients undergoing AFL ablation, the detection of AF in patients without previous AF significantly increased with more frequent monitoring and/or longer duration of follow-up.⁷ This is especially important in patients at high risk of stroke or thromboembolism (>2%/year). Intermittent monitoring may be performed with ambulatory monitors or wearable devices. Alternatively, implantable devices can provide more prolonged and continuous monitoring. Implantable cardiac monitors have been used in multiple settings to detect AF, including after AF ablation or in patients with cryptogenic stroke. Implantable cardiac monitors have also been used for surveillance after AF ablation.⁹

5. Factors that predict AF after flutter ablation include LA enlargement,^19–21 inducible AF at the time of flutter ablation,^22,23 interatrial conduction time,²⁴ prolonged HV interval,²⁵ COPD,²⁶ obstructive sleep apnea,²⁰ and HATCH (hypertension, age ≥75 years, TIA or stroke, COPD, heart failure) score.²¹ One study that reported new-onset AF in 38% of patients after AFL ablation demonstrated a thromboembolic event rate of 10% during a mean follow-up of 5 years.¹² In a Danish registry of patients undergoing flutter ablation, 5% of patients suffered a stroke, and 10% died during a mean follow-up of 4 years.¹³

   Thus, patients who undergo successful flutter ablation are not free from thromboembolic complications during long-term follow-up, likely due to a high rate of development of new-onset AF.

7. RATE CONTROL

7.1. Broad Considerations for Rate Control

Synopsis

Rate control is a suitable strategy for many patients with AF.^1,3 A solitary randomized trial has evaluated the optimal heart rate among patients with AF. In the RACE II study (Rate Control Efficacy in Permanent Atrial Fibrillation: A Comparison Between Lenient Versus Strict Rate Control II), 614 patients with permanent AF were randomly assigned to either lenient rate control (resting heart rate <110 bpm) or strict rate control (heart rate, <80 bpm). A difference was not seen regarding either the primary composite outcome of death from cardiovascular causes, hospitalization for HF, stroke, systemic embolism, bleeding, and life-threatening arrhythmic events (absolute difference of –2.0 percentage points [90% CI, –7.6 to 3.5; P<0.001] for the noninferiority margin)² or the secondary outcome of QOL⁷ at 3 years. These data suggest heart rate control intensity is not a central determinant of clinical outcomes. However, interpretation is limited by a difference of only 10 bpm between study groups attributable to the proportion of patients assigned to strict control who did not achieve target rate control (32.6%; n=98/301) and because 78% of lenient rate control participants having heart rate <100 bpm.⁸ In addition, because patients with HF were underrepresented, whether the results can be extrapolated to those with HF is unknown. Nonrandomized data focused on target heart rate among outpatients,^4,9,10 as well as those with HF,^5,11–16 are divergent.

Recommendation-Specific Supportive Text

1. Rate- and rhythm-control strategies have comparable clinical outcomes in many patients with AF. In the AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) study, rate control was comparable to cardioversion and antiarrhythmic drugs regarding survival (HR, 1.15 [95% CI, 0.99-1.34]; P=0.08).¹ In the HOT CAFE (How to Treat Chronic Atrial Fibrillation) trial, rate control was comparable to cardioversion and use of antiarrhythmic drugs with respect to a composite endpoint of all-cause mortality, number of thromboembolic events, or major bleeding (OR, 1.98 [95% CI, 0.28-22.3]; P>0.71).⁶ In a meta-analysis of randomized trials and observational studies comparing rhythm with rate control, management strategies were comparable with respect to all-cause (OR, 1.34 [95% CI, 0.89-2.02]) and cardiac mortality (OR, 0.96 [95% CI, 0.77-1.20]) as well as stroke (OR, 0.99 [95% CI, 0.76-1.30]).³ Although selection of a rhythm-control therapy within a year of AF diagnosis may be considered to reduce the risk of adverse cardiovascular outcomes,¹⁷ early rate control may still be appropriate as a function of clinical presentation, comorbidity burden, medication profile, and patient preferences. In this context, providers should partner with their patients to select a value-concordant approach within an SDM framework.

2. In RACE II, rate-control intensity (target resting heart rate, <110 bpm versus <80 bpm) did not influence cardiovascular morbidity or mortality.^6,18 Rate control (AFFIRM: goal heart rate, ≤80 bpm at rest and ≤110 bpm during a 6-minute walk test; HOT CAFE: 70-90 bpm at rest and ≤140 bpm with moderate exercise) has been shown to be comparable to rhythm control in a meta-analysis of 5 trials, yet some limitation of the studies include that most patients had well-tolerated persistent AF or high risk of recurrence, no long-term follow-up, and should not be extrapolated to patients with paroxysmal AF, highly symptomatic, or patients with HF.¹ Observational data among outpatients and those with HF are conflicting.⁴ In the ORBIT-AF trial, increasing heart rate was associated with higher all-cause mortality⁹ and incident HF.¹⁰ Whether a lenient rate-control strategy may be safely and effectively used in patients with heart failure with preserved ejection fraction (HFpEF)^5,14 or reduced fraction^{11–13,15,16} has also shown conflicting results. Other populations that may benefit from a low heart rate goal include those with rate-related cardiac dysfunction,¹⁹ ICDs,²⁰ cardiac resynchronization therapy,²¹ and tachycardia-bradycardia form of sick sinus syndrome²² (Table 20).

7.2. Specific Pharmacological Agents for Rate Control

The overall goals for rate control in patients with both acute and chronic AF with a rapid ventricular response center on control of symptoms and the risk of developing LV systolic dysfunction. In general, nondihydropyridine calcium channel blockers (diltiazem and verapamil) and beta blockers are the standard of care in rate-controlling AF. Nondihydropyridine calcium channel blockers slow conduction through the atrioventricular node and have negative inotropic and chronotropic effects. These agents are useful in ventricular rate control in the absence of preexcitation. They provide reasonable rate control and also improve AF-related symptoms compared with beta blockers.^1,2 Beta blockers also slow conduction through the atrioventricular node by blocking beta-1 receptors. Digoxin is a time-honored medication for patients with AF particularly among patients with HFrEF due to its positive inotropic and vagotonic effects. Limited data exist directly comparing rate-control agents, especially in the setting of long-term rate control. Selection of specific agents should consider patient-specific characteristics and response. This section will outline the principles of treatment for acute and long-term rate control.

7.2.1. Acute Rate Control

Synopsis

Controlling the ventricular rate in AF with a rapid ventricular response in the acute setting can vary effectiveness for control of symptoms. It can also present challenges, given some agents may or may not be preferable in the context of underlying comorbidities given the myriad of contributing factors. For patients requiring intravenous rate-control agents, hypotension and/or the presence of decompensated HF may limit use of otherwise efficacious agents. In the acute setting, recognizing patients needing emergency cardioversion is important.

Recommendation-Specific Supportive Text

1. A randomized, double-blind, placebo-controlled trial investigated the safety and efficacy of continuous intravenous diltiazem infusion for 24-hour heart rate control during AF and AFL. Seventeen of the 23 patients (74%) receiving diltiazem infusion and none of the 21 with placebo infusion maintained a therapeutic response for 24 hours. Over 24 hours, patients receiving diltiazem infusion lost response significantly more slowly than did those receiving placebo infusion. In another RCT, patients were randomly assigned in 1:1:1 ratio to receive intravenous diltiazem, digoxin, or amiodarone for ventricular rate control. At 24 hours, rate control was achieved in 119 of 150 patients (79%). The time to rate control was significantly shorter among patients in the diltiazem group, with the percentage of patients who achieved rate control being higher in the diltiazem group (90%) than the digoxin group (74%) and the amiodarone group (74%). A randomized, parallel, open-label study aimed to investigate esmolol versus verapamil in the acute treatment of AF or AFL. The heart rate significantly declined with esmolol and verapamil. Fifty percent of esmolol-treated patients with new onset of arrhythmias converted to sinus rhythm, whereas only 12% of those who received verapamil converted. Mild hypotension was observed in both treatment groups.

2. The use of intravenous digoxin has been shown to be effective in controlling the rate of rapid AF compared with placebo in 1 multicenter RCT.⁵ However, other agents may be safer and more effective in achieving acute rate control. Intravenous diltiazem was more effective than intravenous digoxin in 2 small RCTs.^6,7 One trial performed at an emergency department at an academic center randomized 150 patients with AF without major comorbidities and a ventricular rate >120 bpm in a 1:1:1 ratio to intravenous diltiazem, digoxin, and amiodarone.⁶ The time to reduce the heart rate <90 bpm was significantly shorter in the diltiazem group compared with either digoxin or amiodarone, with fewer symptoms and shorter hospital lengths of stay. Another RCT of 30 patients randomized patients with AF and a rapid ventricular response to either intravenous diltiazem, digoxin, or a combination of both agents. Treatment with intravenous diltiazem significantly decreased heart rate within 5 minutes versus 3 hours with intravenous digoxin.⁷ Intravenous diltiazem reduced heart rate more rapidly than intravenous digoxin in a double-blinded RCT in 40 patients with rapid AF after CABG, with similar control at 12 to 24 hours. In another RCT of 84 patients in rapid AF who presented to the emergency department, intravenous amiodarone resulted in faster control of the ventricular response compared with intravenous digoxin.⁸ Finally, 52 patients with rapid AF were randomized to receive either an intravenous combination of diltiazem and digoxin or intravenous diltiazem alone, with a more rapid and durable response to the combination.⁹

3. In addition to atrioventricular nodal blockers and antiarrhythmics, intravenous magnesium has been investigated for rate controlling rapid AF. ^10,11 The mechanism likely stems from blockade of slow inward calcium channels in the sinoatrial and atrioventricular node, thereby slowing the heart rate and prolonging atrioventricular conduction velocity, respectively.¹⁶ Its low adverse effect profile and minimal toxicity make it a favorable option, often as an adjunct to conventional therapy (ie, atrioventricular nodal blockers). A meta-analysis including 6 RCTs (n=745 patients) investigated the effectiveness of intravenous magnesium in rate and rhythm control of rapid AF when administered in combination with standard rate-control methods (including beta blockers, diltiazem, verapamil, digoxin), as well as placebo. In the pooled analysis, intravenous magnesium was superior in achieving rate control (63% versus 40%; OR, 2.49 [95% CI, 1.80-3.45]) and modestly effective in rhythm conversion to sinus (21% versus 14%; OR, 1.75 [95% CI, 1.08-2.84]) compared with standard rate-control methods. Subgroup analysis showed the superiority of a lower dose (≤5 g) (24% versus 13%; OR, 2.10 [95% CI, 1.22-3.61]) compared with the higher dose (>5 g) (16% versus 13%; OR, 1.23 [95% CI, 0.65-2.32]) in rhythm control when compared with placebo.¹⁰

4. Intravenous amiodarone has been shown to be effective in controlling ventricular rates in patients who are critically ill. In a retrospective study of 38 patients admitted to the intensive care unit, intravenous amiodarone was associated with a statistically significant decrease in heart rate without decrease in BP compared with intravenous diltiazem and digoxin.¹² One study of 60 critically ill patients, predominantly with AF, with heart rate consistently >120 bpm randomized patients to 1 of 3 intravenous treatment regimens: diltiazem in a 25-mg bolus followed by a continuous infusion of 20 mg/h for 24 hours; amiodarone in a 300-mg bolus; and amiodarone in a 300-mg bolus followed by 45 mg/h for 24 hours.¹³ A >30% rate reduction within 4 hours was not statistically different in either treatment arm, with less effective control more frequently observed in those receiving a bolus of amiodarone alone; thus overall, diltiazem allowed for significantly better heart rate control over 24 hours, but it also caused a significantly higher incidence of hypotension requiring discontinuation of the drug.

5. Limited data exist regarding the use of nondihydropyridine calcium channel blockers in patients with HFrEF, due to the presumed limitation of negative inotropic effects. Two retrospective analyses highlight potential for increased morbidity when administered in this population. In a retrospective analysis of patients hospitalized with AF with rapid ventricular rate (RVR), diltiazem was associated with an increased risk of acute kidney injury within 48 hours of initiation of diltiazem in those patients with LVEF ≤50%, compared with those with normal EF (10% versus 3.6%; P=0.002).¹⁴ An additional retrospective analysis compared patients with HFrEF receiving either intravenous metoprolol or diltiazem. There was a higher incidence of worsening HF symptoms (increased oxygen requirement within 4 hours or initiation of inotropic support with 48 hours) in those patients receiving diltiazem (33% versus 15%; P=0.019). Neither analysis noted an increase in the in-hospital mortality rate, need for a higher level of care, or hypotension.

7.2.2. Long-Term Rate Control

Synopsis

Both nondihydropyridine calcium channel blockers (verapamil and diltiazem) and beta blockers are effective for long-term rate control. These agents are useful in ventricular rate control in the absence of preexcitation. Digoxin may be useful in patients with limited tolerability to other agents, or as adjunct therapy in patients with a difficult to control ventricular rate. Limited data exist comparing various rate control agents, especially in the setting of chronic rate control. Selection of specific agents should consider patient-specific characteristics (ie, HFrEF, reactive airway disease) and response (Figure 18).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Recommendation-Specific Supportive Text

1. One prospective, randomized, investigator-blinded, crossover study compared the effect of 4 rate-reducing once-daily drug regimens on the ventricular heart rate and arrhythmia-related symptoms in patients with permanent AF. The 24-hour mean heart rate was 96±12 bpm at baseline (no treatment), 75±10 bpm with diltiazem, 81±11 bpm with verapamil, 82±11 bpm with metoprolol, and 84±11 bpm with carvedilol. All drugs significantly reduced the heart rate compared with baseline.¹³ In a retrospective cohort study, investigators used a database from 2 urban emergency departments to identify consecutive patients with emergency department discharge diagnoses of AF. A total of 259 consecutive patients were enrolled, with 100 receiving calcium channel blockers and 159 receiving beta blockers. Baseline demographics and comorbidities were similar. Twenty-seven percent of patients taking beta blockers were admitted, 31.0% of patients taking calcium channel blockers were admitted, and there were no significant differences in emergency department length of stay, adverse events, or 7- or 30-day emergency department revisits. A follow-up analysis of the AFFIRM study examined the difference in all hospitalization and all-cause mortality in participants treated with a single rate-control agent at baseline. No difference was found among those participants receiving beta blockers, nondihydropyridine calcium channel blockers, or digoxin.²

2. Seminal work investigating digoxin toxicity was based on a small series of patients, where toxicity was determined by electrocardiographic abnormalities.¹⁴ Toxicity was seen when serum digoxin concentrations exceeded 2.0 ng/mL, and almost certainly at levels >3.0 ng/mL. Based on this small study, the narrow therapeutic range of digoxin is usually cited as 0.8 to 2.0 ng/mL, with substantial variation across institutional laboratories and published references.^15,16 However, 3 post hoc analyses of the DIG (Digitalis Investigation Group) trial suggested that safe use of digoxin is seen at lower serum concentrations.^4–6 Serum digoxin concentrations of 0.5 to 0.9 ng/mL were associated with significantly lower all-cause mortality rates and hospitalizations compared with concentrations ≥1.0 ng/mL.³ Another analysis suggested that serum digoxin concentrations >1.2 ng/mL may be harmful, particularly in women.⁷ A post hoc analysis comparing patients with AF taking digoxin with propensity score-matched controls suggested that patients with serum digoxin concentrations <0.9 ng/mL had no increased risk of death, concentrations of 0.9 to 1.1 ng/mL had a nonsignificant increased risk of death, and serum digoxin concentrations ≥1.2 ng/mL were associated with a significant (56%) increased risk of death.⁴ Serum digoxin concentrations of 0.5 to 0.8 ng/mL seem safest in terms of benefit without adverse effects in patients with HFrEF.¹⁷

3. Digoxin may have added efficacy as rate control in conjunction with beta blockers in patients with AF. A small RCT of 47 patients with persistent AF and concomitant HFrEF compared the effects of digoxin alone, carvedilol alone, and their combination.¹⁰ Combination therapy was significantly associated with better rate control and symptom relief than monotherapy with either agent. The RATE-AF trial randomized 160 patients with persistent AF and concomitant HFrEF to digoxin and bisoprolol.¹¹ No significant differences in the primary endpoint of QOL were noted at 6 months with either agent. Several secondary outcomes, including N-terminal pro-brain natriuretic peptide level, were significantly lower in the digoxin group, without a significant difference in resting heart rate between the 2 groups at 12 months. Adverse effects were less common in the digoxin group. A large meta-analysis of 75 studies and approximately 620 000 patients found a significantly increased risk of death in digoxin users overall but, when limited to 7 RCTs (∼8400 patients), no difference in mortality rate was observed between digoxin or placebo—of note, these 7 RCTs in the subanalysis were in patients with concomitant HF.⁹ A post hoc sensitivity analysis assessing the impact of digoxin on mortality in patients with HF and concomitant AF found no impact on overall mortality in patients taking digoxin, most with HFrEF.

4. MDPIT (Multicenter Diltiazem Postinfarction Trial) demonstrated an association of worsening HF in patients with a recent MI, with LV dysfunction randomized to diltiazem. Among those with a baseline EF of <40%, late HF appeared in significantly more patients receiving diltiazem compared with placebo (21% versus 12%; P=0.004). Life table analysis in patients with an EF of <40% confirmed significantly more frequent late HF in those taking diltiazem. In addition, the diltiazem-associated rise in the frequency of late congestive heart failure (CHF) was progressively greater with increasingly severe decrements in baseline EF.¹⁰

5. The safety and efficacy of dronedarone, originally designed as an antiarrhythmic to maintain sinus rhythm in patients with AF, was addressed in several seminal trials.^18–20 PALLAS (Palbociclib Collaborative Adjuvant Study) was designed to investigate whether dronedarone would reduce major vascular events or hospitalizations in patients with permanent AF. The benefits of dronedarone in patients with permanent AF were initially suggested from the ERATO (Efficacy and Safety of Dronedarone for the Control of ventricular rate during atrial fibrillation) trial, which showed a significant reduction in the ventricular response for patients with paroxysmal and permanent AF in the 24-hour setting, which were carried forward in a 6-month follow-up.¹² PALLAS randomized approximately 3200 patients with permanent AF at high risk for adverse events. The trial was terminated early, because of the strong signal for harm in 43 patients receiving dronedarone versus 19 receiving placebo, namely death (13/21 due to arrhythmia), stroke, and hospitalization for HF.¹²

7.3. Atrioventricular Nodal Ablation (AVNA)

Synopsis

AVNA provides ventricular rate control effectively and without medications yet creates dependence on pacing.³ Consideration of the consequences of lifelong pacemaker implantation, particularly with respect to age and comorbidities, is central to decision-making regarding benefit.

Small RCTs of AVNA and right ventricular pacing (RVP) have shown improvements in symptoms and QOL compared with medical rate control alone but found no significant differences in EF or other measures of cardiac performance.^3,14 Progression to persistent AF is described in many patients with previously paroxysmal AF after AVNA.^15,16 Antiarrhythmic drugs reduce the transition to permanent AF but do not improve QOL or echocardiographic parameters, and patients treated with antiarrhythmic drugs have more episodes of HF and hospitalizations.¹⁷

Early concerns regarding AVNA included risks for device-related complications, sudden cardiac death, and worsened HF. Deaths due to malignant arrhythmias have been nearly eradicated by programming higher lower-rate pacing in the early postprocedure period.^1,2,18 The risks of HF may be ameliorated by non-RVP strategies, but data are limited in patients who do not have HF before AVNA.¹⁹

Overall, long-term data on outcomes after AVNA are limited, and no evidence supports AVNA as first-line therapy.⁹

Recommendation-Specific Supportive Text

1. Early observational studies showed a concerning incidence of sudden death after AVNA in 3% to 7% of patients.^20,21 This immediate postprocedural risk results from ventricular fibrillation (VF) predominantly due to bradycardia, QT prolongation, and heterogeneity of repolarization. Initiation of VF is often pause dependent or occurs during a relatively slow pacing rate.¹⁸ Deaths due to malignant arrhythmias have been minimized by current protocols specifying higher lower-rate pacing in patients after AVNA in the early postprocedure period.^1,2,18 Subsequent lower rate adjustment is performed over the course of several weeks.

2. AVNA and pacemaker implantation provide safe and effective rate control in the fraction of patients with AF who cannot be effectively rate controlled with medical therapy^3,4,22 and in whom sinus rhythm has failed or is not deemed to be effective. A meta-analysis of 6 small RCTs of studies comparing AVNA and pacemaker versus medical therapy or medical therapy with pacemaker including 323 subjects with paroxysmal or persistent AF showed improvement in symptoms and QOL in the AVNA group.⁶ No differences in survival, stroke, hospitalization, EF, or exercise capacity were found. There was no effect of AVNA on all-cause mortality in a pooled analysis.⁶ Nonrandomized studies have shown improvement in a range of clinical outcomes⁵ in addition to data showing improved EF in patients with suspected tachycardia-related cardiomyopathy after AVNA.²³ These favorable findings in nonrandomized studies are likely subject to the placebo effect of device implantation. Early and late complication rates are not inconsequential, specifically in young patients in view of the risk of pacemaker-mediated cardiomyopathy, and long-term follow-up data are scant.⁹

3. Although few studies concerning AVNA and pacing report complication rates for the pacemaker implant, the overall risk of lead dislodgement or failure is approximately 2%, and many operators perform both AVNA and pacemaker implant during the same procedure.^9,13,24 Practice patterns vary, however, and a European survey indicated up to 80% of operators will choose to perform AVNA 1 to 3 months after pacemaker implantation to reduce risk of adverse outcome due to early lead dislodgement.²⁵ In the prospective FOLLOWPACE registry from 23 centers in the Netherlands, lead dislodgment was the most frequent lead-related complication, occurring in 3.3% of patients within the first 2 months of implant.⁷ Relevant to the older age population usually undergoing AVNA, a pooled patient-level analysis of 4814 subjects from CTOPP (Canadian Trial of Physiologic Pacing), UKPACE (United Kingdom pacing and cardiovascular events), and Danish pacing trials—designed to compare pacemaker implantation complication rates in age <75 years versus age ≥75 years—found a higher risk of early complications in those age ≥75 years (5% versus 3%), driven by a higher rate of lead dislodgement (2%) and pneumothorax (1.6%).⁸ If AVNA and pacing are performed concomitantly, ideally, leads should be placed first to ensure adequate lead function.

4. HF after AVNA is attributed to the deleterious effects of RVP,²⁶ and risk is correlated with baseline cardiac function.⁹ The PAVE (Post AV Nodal Ablation Evaluation) trial compared outcomes in patient with persistent AF and uncontrolled rates undergoing AVNA and RVP with AVNA and biventricular pacing (BiVP) and found an improvement in 6-minute walk test and EF, but benefit was predominantly seen in subjects with EF <45% or New York Heart Association (NYHA) class II to III HF at baseline.²⁷ Not all patients appear to be at high risk of HF after AVNA and RVP.²⁸ Because the benefit is less and the risk of complications is higher in BiVP compared with conventional pacing, RVP is advised in patients with preserved EF undergoing AVNA.¹⁹

   Conduction system pacing of the His bundle or left bundle area offers promise to reduce the risk of RVP-induced cardiomyopathy and improve outcomes. Early studies have justifiably focused on patients with HF undergoing AVNA, and limited data are available in patients with preserved EF and no HF diagnosis.²⁷ His bundle pacing with AVNA has been shown to be feasible and associated with improvements in EF in patients with EF <40%, and NYHA class, albeit with elevated pacing thresholds.^10,11 Left bundle area pacing may deliver the advantages of His bundle pacing with reduced risk of elevated pacing thresholds or lead dislodgement.¹² Early studies have shown stable EF with conduction system pacing and AVNA in follow-up, and improvement in patients with EF <50%.^12,24

8. RHYTHM CONTROL

8.1. Goals of Therapy With Rhythm Control

Synopsis

When caring for a patient with AF, deciding between rhythm and rate control is important but critical to acknowledge that the decision is nuanced, can evolve over time, and that the 2 strategies are necessarily not mutually exclusive depending on the definition used. For example, in EAST-AFNET 4 (Early Treatment of Atrial Fibrillation for Stroke Prevention Trial), which randomized patients with recently identified AF to rhythm- or rate-control strategies, in those patients randomized to the rhythm control arm, initially 87% were treated with antiarrhythmic drugs and 8% with AF ablation; but, at 2 years, 35% were not on antiarrhythmic drugs or had received AF ablation.¹² Similarly, although the decision on whether to use rhythm control is often based on symptoms, in a post hoc analysis of the same trial, benefits in the composite outcome of death, stroke, or cardiac hospitalization was observed in all patients regardless of whether symptoms were present or not.³⁰ Data from EAST-AFNET 4 and large registries have consistently demonstrated the importance of monitoring patients for increased AF burden once AF has been identified and that rhythm-control therapies are more likely to be successful when implemented early when AF burden begins to increase.^{13,14,30,46,47} Finally, for all patients with AF, continued long-term management of modifiable risk factors is essential.⁴⁸

Recommendation-Specific Supportive Text

1. AF and reduced LV function are often recognized in the same patient, with the combination being associated with worse prognosis when compared with patients with either HF or AF alone.^1,49 Several studies have demonstrated improved LV function in patients with AF after restoration of sinus rhythm.^2–6 The incidence of AF or other atrial arrhythmias as an important contributor to reduced LVEF is unknown and likely dependent on arrhythmia burden, ventricular rates, and other factors.^4–6 However, in patients who present with both AF and reduced LV function or develop reduced LV function after an initial diagnosis of AF, where no identifiable cause for the reduced LV function is observed, AF should initially be assumed as a very possible cause until proven otherwise. Although AF as a cause for reduced LV function is usually considered in the setting of rapid ventricular rates, improved LV function with rhythm control has been reported in the setting of AF with relatively well controlled heart rates.^2,3

2. Rhythm and rate control can improve symptoms in AF.^7,8,50,51 Multiple studies have shown improved QOL with rhythm control.^7–9,50,51 Continued symptoms due to AF are common with rate control. In the RACE II study, lenient rate control had similar primary outcomes (a combination of cardiovascular death, hospitalization for HF, and several other endpoints) compared to strict rate control, but at the end of the follow-up period (>2 years), 46% of patients in both groups had continued symptoms associated with AF.¹⁰ Although a different population, symptoms due to recurrent AF were identified in 56 of 303 (18%) patients with newly identified AF undergoing rhythm control with either AADs or catheter ablation.¹¹

3. EAST-AFNET 4 randomized 2733 patients with early onset AF (<1 year) and other risk factors for stroke to rhythm control or rate-control strategies.¹² Rhythm control was associated with a 25% reduction in the combined endpoint of mortality rate, stroke, and hospitalizations due to HF or ACS. In 2 observational studies, 1 from claims data and another from a large national database, when compared with rate control, early initiation of rhythm control (<1 year) in patients with AF was associated with 15% and 19% reductions in the combined endpoint of cardiovascular death, ischemic stroke, or hospitalization for ischemia or HF, respectively.^13,14 However, these recent findings must be balanced against studies that have reported increased emergency department visits and health care resource utilization associated with rhythm control.^52–54

4. In the AF-CHF (Atrial Fibrillation and Congestive Heart Failure) trial, rhythm control did not affect cardiovascular mortality but was associated with improved QOL.^15,55 Contemporary trials in patients with AF and HF, regardless of LV function, and often using catheter ablation as an important therapy, have reported trends or statistically significant improvement in clinical outcomes associated with rhythm control when compared to rate control.^{16,17,49,56–58} In an analysis of patients with HF in EAST-AFNET 4, rhythm control was associated with a significant reduction in the composite primary endpoint of cardiovascular death, stroke, or hospitalization due to HF or acute ischemic syndrome (HR, 0.74 [95% CI, 0.56-0.97]; P=0.03).¹⁶ Similarly, in the CABANA (Catheter Ablation vs Antiarrhythmic Drug Therapy for Atrial Fibrillation) trial, catheter ablation-based rhythm control was associated with a 46% reduction in the mortality rate in patients with HF compared with medical therapy.¹⁹ Although catheter ablation is more effective in maintaining sinus rhythm than antiarrhythmic drugs, maintenance of sinus rhythm was higher in the AF-CHF trial. In this study, amiodarone was the preferential strategy for rhythm control (82% of patients) compared with the ablation-based CASTLE-AF (Catheter Ablation versus Standard Conventional Therapy in Patients with Left Ventricular Dysfunction and Atrial Fibrillation) trial (sinus rhythm at study completion: AF-CHF: 75% versus CASTLE-AF: 66%), and additionally the benefit of rhythm control in patients with HF is likely greatest in those patients with recently identified AF.^15–18,49

5. AF commonly, but not always, progresses over time to higher burdens and becomes more sustained.^20,59 Large registry studies have reported that AF progression is more commonly observed with a rate-control strategy compared to a rhythm-control strategy over a 1- to 2-year period, although the absolute magnitude and difference has varied. For example, in 2 registry studies, progression to a more sustained form of AF burden was observed in 26% to 28% of patients undergoing a rate-control strategy compared with 6% to 11% of patients treated with a rhythm-control strategy while, in an analysis of the ORBIT AF trial, although rhythm control was associated with a lower likelihood of progression, AF progression was much more common regardless of strategy (66% for rate control and 56% for rhythm control) likely due to older patients in this cohort.^21–23 AF progression is a complex process, and although the choice between rhythm-control and rate-control strategy has an impact, many other factors such as age, presence or absence of HF or other comorbidities, LA size, heart rate, and modifiable risk factors are also important.^21–27

6. The reported prevalence of asymptomatic AF varies from 10% to 40% depending on study cohort.²⁸ Asymptomatic and symptomatic patients with AF have similar risks for death and stroke.^28,29 In a post hoc analysis of EAST-AFNET 4, similar magnitudes of benefit were observed in symptomatic and asymptomatic patients for reducing the combined endpoint of cardiovascular death, stroke, and hospitalization for HF.³⁰ Nevertheless, the evidence base for the benefits of a rhythm-control trial in asymptomatic patients with sinus rhythm is limited. Improved energy with return to sinus rhythm was noted in 1 observational study of 13 asymptomatic patients with persistent AF.³¹ In another study that included 18 asymptomatic patients, at 1-month follow-up, no patients had improvement in symptoms, although sinus rhythm was maintained in only 35% of patients.³² In 2 small studies that focused on the impact of catheter ablation in asymptomatic patients, maintenance of sinus rhythm was associated with improved QOL indices in all patients, but 24% to 34% of patients developed symptoms mainly due to AT after the ablation procedure.^53,54

7. AF is associated with increased cognitive impairment, progressive increase in AF burden, and structural changes in the heart.^33,34 Nonrandomized studies from registries or post hoc analyses of randomized trials have reported that rhythm control is associated with a reduction in the incidence of dementia and development of HF.^36–38 Furthermore, AF can also be an etiologic reason for mitral regurgitation or tricuspid regurgitation.^39–45 In 1 study of 53 patients referred for AF ablation with moderate mitral regurgitation or worse, maintenance of sinus rhythm at follow-up was associated with significant decreases in LA size, mitral annular dilatation, and mitral regurgitation.³⁹ In another study, 70% of patients undergoing ablation for AF who maintained sinus rhythm at follow-up had improvement in mitral regurgitation.⁴⁰ Similar results have been found in tricuspid valve function as assessed by echocardiography.^42–45

8.2. Electrical and Pharmacological Cardioversion

Cardioversion to restore sinus rhythm is a mode of acute rhythm control. Cardioversion can be achieved electrically or pharmacologically. Electrical cardioversion, given rapidity and efficacy, is the treatment of choice for patients with hemodynamic instability attributable to AF. Acute rhythm control with cardioversion should also be considered for patients with hemodynamically stable AF intolerant of atrioventricular dyssynchrony, loss of atrial kick, or unable to achieve adequate rate control. In patients with hemodynamically stable AF, both electrical cardioversion and pharmacological cardioversion are acceptable, safe, and efficacious methods for acute rhythm control.^1–3 Electrical cardioversion is more effective than pharmacological cardioversion alone^4,5 but involves the trade-off of anesthesia or sedation.⁶ Thromboembolic risks and considerations for anticoagulation apply to both pharmacological cardioversion and electrical cardioversion.^7,8

8.2.1. Prevention of Thromboembolism in the Setting of Cardioversion

Synopsis

Thromboembolic risks and considerations of anticoagulation apply to both pharmacological cardioversion and electrical cardioversion.^23,24 Thromboembolic risks in the setting of acute cardioversion relate to preexisting thrombus, change in atrial mechanical function with restoration of sinus rhythm, atrial stunning postcardioversion, and transient prothrombotic state.^2,25 DOACs are alternatives to VKAs as thromboprophylaxis in the setting of cardioversion.^26–29 Given additional favorable attributes of rapid time to therapeutic efficacy, reliability of maintenance of therapeutic efficacy, and ease of continuation, DOACs may be preferentially considered over VKAs for patients with AF planned for cardioversion barring contraindications to DOAC therapy. Precardioversion imaging is prudent for patients with AF who have previously undergone LAAO and are not on anticoagulation.^{9,10,14–16,19} Thromboembolic risks for patients with <48 hours of AF were not homogenous but, rather, varied by a patient’s risk profile for thromboembolic complications (as defined by risk scores in specific studies) and duration of AF.^21,22 More nuanced consideration of a patient’s thromboembolic risks in the setting of cardioversion for AF may integrate a patient’s substrate for atrial myopathy and risks for thromboembolism instead of strict dependence on the previously used 48-hour duration threshold.

Recommendation-Specific Supportive Text

1. Early experience with both pharmacological and then electrical cardioversion of AF led to the observation that thromboembolism was less frequent in patients who were chronically anticoagulated than those who were not.³⁰ This led to an empiric recommendation for 3 weeks of anticoagulation before cardioversion in the prospective, randomized, multicenter ACUTE (Assessment of Cardioversion Using Tranesophageal Echocardiography) study of patients with >2 days’ duration of AF.³¹ Precardioversion anticoagulation of at least 3 weeks may be accomplished via either therapeutic anticoagulation with VKAs or DOACs.^26–28 Cardiac computed tomography, particularly with delayed contrast-enhanced image acquisition protocol, has emerged as an alternate imaging modality to exclude intracardiac thrombus.^32–34

2. Rationales for establishment of therapeutic anticoagulation before cardioversion and continuation for the subsequent 4 weeks after cardioversion for AF are driven by elevated thromboembolic risks,^6,7,35 high early recurrence of AF,³⁶ and observation of atrial dysfunction^2–5 during this period. Thromboembolic risks are elevated around the time of cardioversion, especially within the 30 days after cardioversion.^6,7 The observation of spontaneous echocardiographic contrast in association with atrial dysfunction on TEE after cardioversion raised mechanistic concerns for atrial stunning in the setting of cardioversion,^2,3 with recovery of mechanical atrial systole over the ensuing 1 month.^4,5 Rapidity of recovery of mechanical atrial systole may be influenced by clinical variables,³⁷ duration of the antecedent AF episode leading to cardioversion,⁵ and the extent of comorbid conditions as risks factors for atrial myopathy.³⁸

3. In patients with AF, detection of intracardiac thrombus should prompt cancellation of planned cardioversion and institution of therapeutic anticoagulation in anticoagulant-naïve patients. Detection of intracardiac thrombus in patients already on anticoagulation should prompt assessment of compliance, dosing appropriateness, drug absorption, and drug interactions. In patients with thrombus detected despite already being anticoagulated, potential reasons should be investigated, such as nonadherence and drug interaction, which may cause subtherapeutic levels. Switching to an alternate anticoagulant can be considered, although the benefit is uncertain. In anticoagulant-naïve patients, 61.2% had resolution of thrombus on follow-up TEE performed 3 to 12 weeks after initiation of VKA in the CLOT-AF (Retrospective Registry Providing Baseline Data on the Outcome of Left Atrial [LA] or LA Appendage [LAA] Thrombus in Patients With Nonvalvular Atrial Fibrillation [AF] or Atrial Flutter After Standard of Care [SoC] Anticoagulant Therapy) study, and 41.5% had resolution of thrombus on follow-up TEE performed 6 weeks after therapy with rivaroxaban in the X-TRA (Exploring the Efficacy of Once Daily Oral Rivaroxaban for Treatment of Thrombus in Left Atrial/Left Atrial Appendage in Subjects With Nonvalvular Atrial Fibrillation or Atrial Flutter) study.⁸ Repeat imaging (TEE or computed tomography^32–34) is generally pursued after at least 3 to 6 weeks of therapeutic anticoagulation to assess for resolution of intracardiac thrombus before reconsidering cardioversion.^1,8

4. In the subset of patients who have undergone surgical or pLAAO procedures, inadequate occlusion may be seen acutely during the procedure or delayed after the procedure.^9–14 Inadequate LAAO has been associated with elevated risks for thromboembolism.^9–11 Device-related thrombosis, reported in patients with pLAAO devices,^15–19 has also been associated with elevated stroke/systemic embolism.^15–18 Imaging with TEE or cardiac computed tomography can assess adequacy of LAAO, residual leak, and device-related thrombosis.^39–41 In the setting of transesophageal echocardiographic imaging before electrical cardioversion, device-related thrombus was detected in 2.7% (4 of 148 patients) of a small multicenter retrospective analysis of patients with Watchman LAAO device.²⁰ These 4 patients were treated with oral anticoagulation and after 6 to 8 weeks, on confirmation of thrombus resolution by repeat TEE, underwent successful electrical cardioversion.²⁰

5. Data are limited to guide pericardioversion anticoagulation in patients with LAAO. In the broader (not limited to cardioversion) context, residual leak ≤5 mm by computed tomography after LAAO by surgical ligation has been associated with particularly elevated stroke/system embolism risks (especially in the absence of anticoagulation)⁹ and similarly for residual leak <5 mm by TEE post-Lariat.¹⁰ A large analysis that stratified National Cardiovascular Data Registry LAAO registry Watchman 2.5 device patients by transesophageal echocardiographic observation of residual leak (none, ≤5 mm, >5 mm) at 45 days, found ≤5 mm residual leak to be common (13 258 of 51 333 patients; 25.8%) and associated with modestly higher incidence of thromboembolic events over 1-year follow-up compared with no leak.¹¹ OAC use was more frequent in patients with large (>5 mm) residual leak.¹¹ Data specific to the setting of cardioversion are further limited. The largest to date is a retrospective multicenter study of 148 Watchman LAAO device patients who underwent electrical cardioversion after precardioversion exclusion of intracardiac thrombus or large residual leak (≥5 mm in this study) by TEE.²⁰ No thromboembolic events were observed within 6 weeks of cardioversion, in recipients or nonrecipients of pericardioversion anticoagulation.²⁰ However, detection of thromboembolic events may be limited by study’s sample size. Pericardioversion antithrombotic regimen was at the discretion of the treating physicians,²⁰ and the relationship of anticoagulation with presence/absence of small residual leak was not reported. Considerations for anticoagulation may also vary based on specific surgical or pLAAO procedures and devices.

6. The safety of cardioversion without further assessment or previous anticoagulation in patients with AF duration of <48 hours has been challenged. In addition to concerns for underestimation of actual duration of episode and burden due to potential for asymptomatic occurrence of AF, emerging data demonstrate that thromboembolic risks in patients with <48 hours of AF were not homogenously low. Time to cardioversion >12 hours has been reported as an independent predictor for thromboembolic complications.²¹ A single-center observational study of patients undergoing cardioversion for AF of <48 hours duration did not observe thromboembolic events in patients with CHA₂DS₂-VASc score of 0 or 1 or patients with postoperative AF but noted differential thromboembolic rates in CHA₂DS₂-VASc score ≥2.⁴² Larger studies similarly demonstrated that among patients with <48 hours of AF, postcardioversion thromboembolic risks increased with increasing CHA₂DS₂-VASc score,^22,43 especially in those with CHA₂DS₂-VASc score ≥2 in the absence of anticoagulation.²²

7. Retrospective observational data from the FinCV (Finnish CardioVersion) study noted low incidence of 30-day pericardioversion thromboembolic event rate of 0.4% in patients with AF with CHA₂DS₂-VASc of 0 to 1²² or 0.3% in patients with duration of <12 hours of AF.²¹ Other retrospective studies have similarly demonstrated low pericardioversion thromboembolic risks in patients with CHA₂DS₂-VASc score of 0 to 1.⁴² The combination of CHA₂DS₂-VASc of 0 to 1 in conjunction with duration of <12 hours of AF may identify a population at particularly low risk for pericardioversion thromboembolism.

8.2.2. Electrical Cardioversion

Synopsis

Electrical cardioversion is a strategy for acute rhythm control. Compared with pharmacological cardioversion alone, electrical cardioversion with synchronized direct current cardioversion is more rapid and efficacious for restoring sinus rhythm from AF.^1,11,12 Given rapidity and efficacy, electrical cardioversion is the treatment of choice for patients with hemodynamically unstable AF.

In hemodynamically stable patients with AF planned for acute rhythm control, electrical cardioversion can be performed as an initial strategy or after unsuccessful pharmacological cardioversion.^2,6,13,14 Electrical cardioversion is favored over pharmacological cardioversion when the patient is able to tolerate sedation, desires more immediate rhythm conversion, or has failed or not met candidacy for pharmacological cardioversion.

Although electrical cardioversion may achieve acute restoration of sinus rhythm, durability of sinus rhythm may be influenced by ongoing acute conditions, LA size, duration of the history of AF and/or specific episode, and comorbidities.¹⁴ In patients with a longer duration of AF, immediate recurrence of AF, a previous unsuccessful electrical cardioversion, or a desire to employ all reasonable efforts to avoid an AF recurrence, pretreatment antiarrhythmic drugs and subsequent continuation to facilitate success of acute cardioversion and promote maintenance of sinus rhythm postcardioversion^6,8,9 may be pursued in the context of SDM with the patient. Optimization of energy delivery, electrode vector, and reduction of transthoracic impedance with manual pressure augmentation may also improve success of electrical cardioversion.^5–10

Recommendation-Specific Supportive Text

1. In patients with hemodynamically unstable AF, immediate electrical cardioversion with synchronized direct current cardioversion is the treatment of choice. Direct data on emergency cardioversion of hemodynamically unstable patients is limited because of the emergency circumstances of the clinical context. Data on direct current cardioversion of AF in hemodynamically stable patients demonstrate high success rate of restoring sinus rhythm with direct current cardioversion.^1,11,15–18 Electrical cardioversion is rapid. Electrical cardioversion alone is more effective than pharmacological cardioversion alone.^1,11,12

2. Goals of rhythm control are summarized in Section 8.1 (“Goals of Therapy With Rhythm Control”). Real-world data support both electrical cardioversion and pharmacological cardioversion as acceptable, safe, and efficacious methods for acute rhythm control.^6,13,14 Electrical cardioversion is more effective than pharmacological cardioversion alone^1,11 but involves the trade-off of requiring sedation.¹⁹ In patients with hemodynamically stable AF of recent onset, the RAFF2 (Electrical versus pharmacological cardioversion for emergency department patients with acute atrial fibrillation) study demonstrated high success of restoration of sinus rhythm with either an upfront electrical cardioversion strategy or a stepwise strategy with initial pharmacological cardioversion and then direct current cardioversion if the pharmacological intervention did not restore sinus rhythm.² In patients with unsuccessful pharmacological cardioversion, instead of switching to another antiarrhythmic drug (and with concerns for potentiation of adverse effects such as QT prolongation or bradycardia), electrical cardioversion should be pursued as next step for acute rhythm control. Therefore, in patients with hemodynamically stable AF, electrical cardioversion is favored over pharmacological cardioversion when the patient is able to tolerate sedation, desires more immediate rhythm conversion, or has failed or not met candidacy for pharmacological cardioversion. Monitoring and preparedness during and after electrical cardioversion include monitoring for rare chronotropic, hemodynamic, or thromboembolic complications, and the effects of anesthesia administered during electrical cardioversion.^{6,13,14,20–22}

3. Unsynchronized or inappropriately synchronized cardioversion can induce VF if the shock energy is delivered during the vulnerable period, the inscription of the T wave.^3,23–25 Therefore, electrical cardioversion for AF should be synchronized with the QRS complex. The synchronization feature needs to be turned on or confirmed that it is on; in patients who require multiple shocks, the synchronization feature must be manually activated and/or confirmed before each shock attempt. Visual confirmation of synchronization to QRS is important to detect inappropriate synchronization and ensure appropriate synchronization.

4. An earlier small, multicenter, prospective randomized study demonstrated higher efficacy of initial electrical cardioversion with a biphasic waveform instead of monophasic waveform.⁴ A more contemporary single-centered, randomized study of 279 patients who were randomized to synchronized direct current cardioversion with biphasic waveform with either upfront maximum-fixed energy or strategy of low-escalating energy found higher first shock success with maximum-fixed energy (75% compared with 34% for low-escalating energy group).⁵ In the contemporary RAFF2 study of patients with hemodynamically stable AF of recent onset, the upfront electrical cardioversion group with initial biphasic 200 J shock (and up to 3 consecutive shocks allowed, with higher energy permitted for subsequent shocks) had conversion to sinus rhythm in 176 of 192 (92%) patients in this upfront electrical cardioversion shock-only group.²

5. A previous study of patients with persistent AF (median AF duration, 5 months) undergoing direct current cardioversion with escalation of energy found the anterior-posterior orientation of the electrode vector more successful at restoring sinus rhythm than an anterior-lateral vector,⁷ while a subsequent study of patients with recently diagnosed AF (≥3 hours and <7 days) undergoing direct current cardioversion using biphasic upfront 200 J showed similar success rates with electrode vectors in either anterior-posterior or anterior-lateral orientation.² Taken together, when energy output is optimized as biphasic and maximal output and the AF is of recent onset, either vector orientation may be reasonable, but for patients with longer duration of AF, the anterior-posterior orientation of electrode vector may be favorable.² Manual pressure augmentation²⁶ or administration of antiarrhythmic drugs as pretreatment (Section 8.2.3, “Pharmacological Cardioversion”)^6,8,9 may also facilitate electrical cardioversion.

6. Increased body weight has been associated with reduced success of electrical cardioversion.²⁷ In a small randomized clinical study, use of paddles, manual pressure augmentation (using electrically silent objects as insulator), and further escalation of electrical energy improved success of electrical cardioversion of obese patients with AF,¹⁰ and the benefits of these maneuvers may mechanistically relate to overcoming increased transthoracic impedance. Other alternatives in refractory cases include the use of 2 defibrillators simultaneously, effectively doubling the delivered energy.

8.2.3. Pharmacological Cardioversion

Synopsis

Pharmacological cardioversion is indicated for patients with new-onset or persistent AF that is hemodynamically stable, or in rare instances when electrical cardioversion is desired but contraindicated. Ibutilide works rapidly for pharmacological cardioversion of AF but is associated with QT interval prolongation and torsades de pointes, particularly in patients with HFrEF.^2,3,^20,21 Data support intravenous amiodarone for pharmacological cardioversion,^4–8 although intravenous amiodarone requires a longer time for AF conversion than ibutilide. Evidence from randomized studies support the efficacy of flecainide^9–11 and propafenone^10–14 for pharmacological cardioversion, and these drugs are reasonable for administration via the PITP approach for AF occurring outside of the hospital.^16–18 Intravenous procainamide is more effective than placebo for pharmacological conversion of AF¹⁹ but is less effective than ibutilide.^22,23 Dofetilide,^24,25 oral amiodarone,^26–28 and oral sotalol^27,28 have been shown to be effective for conversion of AF to sinus rhythm but require several days for efficacy and therefore are not practical choices for acute pharmacological cardioversion. Data do not support intravenous sotalol for pharmacological cardioversion of AF.^29,30

Recommendation-Specific Supportive Text

1. No studies compare the efficacy of electrical cardioversion with that of pharmacological cardioversion in patients with AF who are hemodynamically unstable. However, in a retrospective, propensity score-matched analysis of 374 hemodynamically stable patients with AF who presented to an emergency department, the incidence of successful cardioversion was higher in the direct current cardioversion group (78.2%) than in those who underwent pharmacological conversion (59.2%) or those for whom a “wait-and-watch” approach (37.9%) was used (P<0.001).¹ Compared with the “wait-and-watch” strategy, the ORs for conversion to sinus rhythm for direct current cardioversion and pharmacological cardioversion were 6.00 (95% CI, 3.38-10.66) and 2.47 (1.45-4.20), respectively (P<0.001). Due to the potential greater efficacy of direct current cardioversion compared with that of pharmacological cardioversion, and due to the need for rapid successful conversion to sinus rhythm, immediate electrical cardioversion is generally the favored option, yet some patients may not be candidates (eg, cannot undergo anesthesia), for whom pharmacological options are available.³¹

2. Randomized, double-blind placebo-controlled studies have established the efficacy of ibutilide^2,3 for conversion of AF to sinus rhythm, with AF conversion rates of about 30% (compared with 2% in placebo groups).^2,3 The efficacy of ibutilide is greater for AFL (conversion rates, 38%-63%) than for AF.^2,3 Ibutilide terminates AF rapidly, with most patients converting to sinus rhythm within 30 to 90 minutes (Table 22). Ibutilide is associated with a risk of QT interval prolongation and torsades de pointes. The incidence of torsades de pointes associated with ibutilide is higher in patients with reduced LVEF than in those with normal LVEF,^2,3 and the risk seems particularly high in patients with severely depressed LVEF.²⁰ HFrEF is an independent risk factor for ibutilide-associated torsades de pointes.^3,21 Therefore, ibutilide is best avoided in patients with LVEF ≤40%. Of note, studies have shown that a magnesium infusion immediately before administration of ibutilide may mitigate the risk of torsades de pointes and excessive QT prolongation.^32,33

3. Multiple randomized studies have found intravenous amiodarone to be effective for conversion of AF to sinus rhythm.^4–7 A meta-analysis of randomized, placebo-controlled studies reported relative risks of sinus rhythm associated with intravenous amiodarone at 6 to 8 hours and at 24 hours of 1.23 (P=0.022) and 1.44 (P<0.001), respectively.⁸ However, pharmacological cardioversion of AF with intravenous amiodarone requires 8 to 12 hours, compared with a much shorter response time with ibutilide (Table 22).

4. Single oral doses of flecainide^9–11 and propafenone^10–14 are effective for conversion of AF to sinus rhythm, with 3- to 4-hour conversion rates of 58% to 68% for flecainide (compared with 18%-29% for placebo) and 45% to 57% for propafenone (compared with 17%-29% for placebo) (Table 22). A beta blocker or nondihydropyridine calcium channel blocker is generally administered at least 30 minutes before a dose of flecainide or propafenone to prevent 1:1 atrioventricular conduction during AFL.¹⁵ The PITP strategy was studied in 268 patients with stable AF of recent onset who presented to the emergency department.¹⁶ Patients received single-dose oral flecainide or propafenone, and those who were successfully treated (n=210) were discharged with a plan for the PITP approach. In the 15±5-month follow-up period, single-dose flecainide or propafenone was successful in 94% of episodes. However, the incidence of adverse effects is not insignificant (6%-17%).^16–18 In view of this, and because in most studies, most patients received the first dose in the hospital before taking it as an outpatient, the PITP strategy should only be used for highly selected patients and after it first has been observed to be safe and effective in an inpatient setting.^17,18,34

5. Intravenous procainamide was shown to be more effective than placebo for conversion of AF to sinus rhythm (conversion rates at 1 hour, 69% versus 38%; P=0.012) in a randomized, double-blind study of 114 patients.¹⁹ However, intravenous procainamide is less effective than ibutilide for conversion of AF to sinus rhythm (conversion rates, 14% versus 76%; P=0.001).²² The efficacy of procainamide for conversion of AF to sinus rhythm was similarly inferior to that of ibutilide in an analysis of pooled data.²³ In addition, intravenous procainamide is associated with a relatively high incidence of clinically significant hypotension (5%-12%)^19,23 and can exacerbate HFrEF.

8.3. Antiarrhythmic Drugs for Maintenance of Sinus Rhythm

8.3.1. Specific Drug Therapy for Long-Term Maintenance of Sinus Rhythm

Synopsis

Antiarrhythmic drugs are reasonable for long-term maintenance of sinus rhythm for patients with AF who are not candidates for, or decline, catheter ablation or who prefer antiarrhythmic therapy. Flecainide^3–5 and propafenone^5–11 are options for maintenance of sinus rhythm in patients with no previous history of MI or significant structural heart disease.^26,27 Dronedarone is an option for maintenance of sinus rhythm in patients without recent decompensated HF or severe LV dysfunction.^5,13–15 Dofetilide^1,5,16 and sotalol^{5–7,10,17,18} are effective for maintenance of sinus rhythm but are associated with torsades de pointes and require QT interval monitoring. Sotalol is best avoided in patients with HFrEF, because most patients are already taking a beta blocker, and the addition of sotalol is unlikely to be well-tolerated. A Cochrane database meta-analysis of randomized studies⁵ reported sotalol to be associated with an increase in all-cause mortality (RR, 2.23 [95% CI, 1.03-4.81]). Low-dose amiodarone is more effective than sotalol and Class IC agents for maintenance of sinus rhythm; but, in view of its adverse effects and multiple drug interactions, is best reserved for patients for whom other antiarrhythmic drugsare ineffective, not preferred, or contraindicated. However, amiodarone and dofetilide are options for maintenance of sinus rhythm for patients with HFrEF, as they are effective,^1,2 and most other drugs are contraindicated.

Recommendation-Specific Supportive Text

1. In subanalyses of randomized trials, dofetilide¹ and amiodarone² have been shown to be effective for maintenance of sinus rhythm in patients with AF who have HF. Most antiarrhythmic agents, except amiodarone and dofetilide, are contraindicated in patients with HFrEF due to worsening of HF and/or increased mortality. Sotalol is best avoided in most patients with HFrEF, because most patients with HFrEFare already taking a beta blocker for mortality reduction, and the addition of a second beta blocker (sotalol) is unlikely to be well-tolerated. Therefore, although amiodarone is associated with a wide range of adverse effects and many clinically important drug interactions, amiodarone is often used as a first-line agent for maintenance of sinus rhythm in patients with AF and HFrEF. Dofetilide is also an option for this indication. Patients undergoing initiation or reloading of dofetilide^29,30 should be admitted for at least 3 days to a health care facility that can provide calculations of CrCl, continuous electrocardiographic monitoring, and availability of cardiac resuscitation.

2. Randomized, controlled studies and a comprehensive Cochrane database analysis have established the efficacy of flecainide^3–5 and propafenone^5–11 for maintenance of sinus rhythm in patients with AF. In the Flec-SL (Short-term versus long-term antiarrhythmic drug treatment after cardioversion of atrial fibrillation) trial,⁴ flecainide prevented episodes of paroxysmal AF in 31% of patients over 4 months compared with 9% in the placebo group (0.013). Sustained-release propafenone 425, 325, and 225 mg twice daily extended the median time to the occurrence of AF, AFL, or supraventricular tachycardia (SVT) compared with placebo (>300, 291, and 112 days versus 41 days, P<0.001, for all propafenone doses versus placebo).⁸ Patients with previous MI or significant structural heart disease (scar or fibrosis) should not take flecainide of propafenone,^26,27 and individuals taking flecainide or propafenone should be concomitantly taking an atrioventricular nodal blocking agent to reduce the risk of 1:1 AFL.¹²

3. Randomized, controlled studies^13–15 and a comprehensive Cochrane database analysis⁵ have established the efficacy of dronedarone for maintenance of sinus rhythm in patients with AF. The combined analysis (n=1237) of EURIDIS (European Trial in Atrial Fibrillation or Flutter Patients Receiving Dronedarone for the Maintenance of Sinus Rhythm) and ADONIS (American-Australian-African Trial with Dronedarone in Atrial Fibrillation or Flutter Patients for the Maintenance of Sinus Rhythm) reported longer median times to AF recurrence in dronedarone groups versus placebo (EURIDIS: 96 versus 41 days; P=0.001; ADONIS 158 versus 59 days; P=0.002).¹⁴

4. Randomized, controlled studies and a comprehensive Cochrane database analysis⁵ have established the efficacy of dofetilide for maintenance of sinus rhythm in patients with AF. In the SAFIRE-D (Symptomatic Atrial Fibrillation Investigative Research on Dofetilide) study,^1,16 the probability of remaining in sinus rhythm at 1 year was higher in patients receiving dofetilide 500 μg twice daily compared with placebo (0.58 versus 0.25; P<0.001). Patients undergoing initiation or reloading of dofetilide^29,30 should be admitted for at least 3 days to a health care facility that can provide calculations of CrCl, continuous electrocardiographic monitoring, and cardiac resuscitation.

5. Multiple randomized, controlled studies have established the efficacy of amiodarone for maintenance of sinus rhythm in AF and have shown amiodarone to be superior to other antiarrhythmic agents. In SAFE-T (Sotalol Amiodarone Atrial Fibrillation Efficacy Trial), amiodarone was superior to sotalol and placebo with respect to median time to recurrence of AF (487 versus 74 versus 6 days, respectively; P<0.001, for amiodarone versus placebo and amiodarone versus sotalol).¹⁷ In CTAF (Canadian Trial of Atrial Fibrillation), AF recurrence rates at 16 months were 35% in the amiodarone group and 63% in those randomized to combined propafenone or sotalol (P<0.001).¹⁹ However, the adverse effects profile of amiodarone is onerous. Amiodarone is associated with a wide range of adverse effects, including pulmonary fibrosis, hypo- or hyperthyroidism, elevated transaminases, and more rarely hepatotoxicity, photosensitivity, changes in skin pigmentation, peripheral neuropathy, sinus bradycardia, QT interval prolongation and torsades de pointes, corneal microdeposits, and rarely optic neuropathy.²⁴ In addition, amiodarone is associated with many clinically relevant drug interactions.³¹ Therefore, amiodarone is best reserved for patients who do not respond to other recommended antiarrhythmic agents or for whom other antiarrhythmic drugs are contraindicated.

6. Randomized, controlled studies and a comprehensive Cochrane database analysis⁵ have established the efficacy of sotalol^6,7,10,17,18 for maintenance of sinus rhythm in patients with AF. In SAFE-T, the median time to AF recurrence was 74 days in the sotalol group, compared with 6 days for placebo (P<0.001).¹⁶ A Cochrane database meta-analysis of 5 randomized studies totaling 1882 patients with AF⁵ reported sotalol to be associated with an increase in all-cause mortality (RR, 2.23 [95% CI, 1.03-4.81]). However, this study included patients with advanced HF, and sotalol may still have a role in patients with preserved heart function. Patients undergoing initiation or reloading of oral sotalol should be admitted for at least 3 days to a health care facility that can provide calculations of CrCl, continuous electrocardiographic monitoring, and cardiac resuscitation. A recent FDA-approved intravenous sotalol formulation can be used to evaluate the drug safety and tolerability within 6 hours and thus may reduce cost and length of stay, although the data are limited to a small sample size and very selected population.³²

7. In the randomized, double-blind, placebo-controlled CAST (Cardiac Arrhythmia Suppression Trial),^26,27 the Vaughan Williams class IC antiarrhythmic agents flecainide and encainide were associated with an increased mortality rate in patients with recent MI; most patients also had LVEF <50%.²⁶ In the continuation of the study (CAST-II), the class IC agent moricizine also increased the mortality rate within the first 14 days of treatment.³³ In CASH (Cardiac Arrest Study Hamburg),³⁴ propafenone was associated with worse outcomes and increased mortality in a population of cardiac arrest survivors, most of whom had structural heart disease. Thus, IC agents are best avoided in patients with AF who have a previous MI or significant structural heart disease, including HFrEF, due to the risk of worsening HF and increased mortality.

8. ANDROMEDA (Antiarrhythmic Trial with Dronedarone in Moderate to Severe CHF Evaluating Morbidity Decrease)²⁸ was a randomized, double-blind placebo-controlled study that tested the hypothesis that dronedarone would reduce the rate of hospitalization attributable to HF and possibly reduce mortality by reducing the incidence of arrhythmic death. The study enrolled patients who were hospitalized with new or worsening HF and who had at least 1 episode of NYHA class III or IV HF (shortness of breath on minimal exertion or at rest). More than 40% of the patients had NYHA class II HF at baseline. The investigators reported that, contrary to the study’s hypothesis, the mortality rate was higher in the dronedarone group compared with that in the placebo group (8.1% versus 3.8%; HR, 2.13 [95% CI, 1.07-4.25]; P=0.03). The higher mortality rate was principally associated with worsening HF.

8.3.2. Inpatient Initiation of Antiarrhythmic Agents

Synopsis

Pharmacological rhythm control is an alternative to catheter ablation in appropriately selected patients. However, many antiarrhythmic agents have a paradoxic risk of proarrhythmia and require close monitoring on initiation. Dofetilide and sotalol confer a relatively high risk of torsades de pointes given they prolong the QT interval, and this can occur early during initiation. As such, patients need to be admitted and closely monitored when starting or reinitiating dofetilide, as well as for increasing the dosage. Many practitioners choose to initiate sotalol in an inpatient setting given it can cause torsades de pointes; furthermore, unlike dofetilide, it can also cause bradyarrhythmia. Amiodarone, flecainide, and propafenone can be started in the outpatient setting. Because PITP with class IC (with concomitant atrioventricular nodal blockers) can cause brady- or tachyarrhythmias, the first attempt during an acute episode of AF may be performed in a monitored environment if a high dose is used.

Recommendation-Specific Supportive Text

1. Dofetilide, a class III antiarrhythmic drug, is a selective I_Kr blocker that also blocks late Na⁺ current (I_NaL) and poses a high risk for torsades de pointes. The SAFIRE-D (Symptomatic Atrial Fibrillation Investigative Research on Dofetilide) study investigated dofetilide in cardioverting and maintaining sinus rhythm in 325 patients with AF or AFL who underwent a minimum 3-day hospitalization.¹ Of 241 patients randomized to dofetilide, 10 patients had QTc prolongation within the first 3 days, with 2 episodes of torsades de pointes degenerating to VF. In a substudy of the AFFIRM trial, 1 of 12 patients on dofetilide experienced torsades de pointes.³ One retrospective cohort of 378 patients found no significant differences in the incidence of torsades de pointes during inpatient initiation of dofetilide and sotalol (1.3% versus 1.2%).⁴ One systematic review found a 1% to 10% incidence of torsades de pointes in patients taking dofetilide, highest in patients with HFrEF and taking higher than recommended (or not renally adjusted) doses.¹⁴ In 2 randomized RCTs from the DIAMOND (Management of Hyperkalemia in Subjects Receiving Renin-angiotensin-aldosterone System Inhibitor Medications for Heart Failure with Reduced Ejection Fraction) population, in which all patients were hospitalized for at least 3 days after dofetilide was initiated, torsades de pointes occurred in 25 of 762 patients (3.3%)⁵ and 7 of 749 patients (0.9%),⁷ respectively, with most occurring within the first 3 days of dosing. In the DIAMOND AF substudy of these trials,⁶ torsades de pointes occurred in 4 dofetilide-treated patients with AF (1.6%).

2. Sotalol is a class III antiarrhythmic with beta-blocking properties. Its action potential prolongation (due to the I_Kr-blocking properties of the d isomer in concert with I_NaL blockade) predisposes patients to torsades de pointes. One meta-analysis demonstrated that sotalol, compared with placebo or no treatment, was associated with a significantly higher all-cause mortality rate (RR, 2.23; 1882 patients in 5 RCTs) and proarrhythmia (RR, 3.55; 2989 patients in 12 RCTs).¹⁸ The SAFE-T trial randomized 665 patients to amiodarone, sotalol, or placebo, all initiated in the outpatient setting (sotalol at 80 mg twice daily for the first week and 160 mg twice daily subsequently).¹⁵ Fifteen deaths occurred in the sotalol group, 8 of which were sudden, with a nonsignificant mortality ratio of 1.8 compared with placebo. In 1 retrospective cohort of 120 patients undergoing sotalol initiation as an inpatient, 2 of 7 had torsades de pointes; 20 patients developed significant bradyarrhythmias necessitating dose reductions, and 3 patients required permanent pacing.⁸ Intravenous sotalol was introduced in the United States in 2015 and reaches therapeutic levels rapidly and with easier dose titration to QTc levels and clinical response. As such, it may be an option for more expedient inpatient initiation (obviating the need for a 3-day initiation period).¹⁶

3. The PITP strategy—a high dose of a class IC agent (flecainide 200-300 mg or propafenone 450-600 mg, with an atrioventricular nodal blocker)—can be considered for patients who experience infrequent episodes of symptomatic AF and do not prefer chronic antiarrhythmic therapy. Class IC agents are generally administered with concomitantly atrioventricular nodal blocking agents to mitigate risk of rapidly conducting AFL. Because of the risk of adverse effects and proarrhythmia with high-dose IC agents, inpatient initiation of PITP has been studied. In 1 trial evaluating the safety of outpatient PITP therapy after initiation as an inpatient, 58 of 268 (22%) patients experienced transient hypotension, 1:1 and 2:1 AFL, and symptomatic bradycardia.⁹ In a prospective study of 43 patients who presented to the emergency department for initiation of PITP,¹⁰ sinus rhythm was restored in 30 patients; 13 patients remained in AF, and/or had symptomatic hypotension, converted to rapid AFL requiring cardioversion, or experienced a syncopal conversion pause. In a retrospective cohort¹³ study of 273 patients initiating PITP (62% inpatient), 7 patients experienced significant adverse events, including syncope, symptomatic bradycardia/hypotension, and organization to 1:1 AFL—2 of 7 of these patients initiated PITP as an inpatient, and 2 patients required permanent pacemakers for bradycardia.

8.3.3. Antiarrhythmic Drug Follow-Up

Antiarrhythmic drugs are associated with important adverse effects, and monitoring is recommended to assess their efficacy for maintaining sinus rhythm and for prevention and/or early detection of adverse events.

Amiodarone

Recommendations for follow-up monitoring for patients taking oral amiodarone are provided in Table 24. Amiodarone contains 37% iodine by weight and is therefore associated with thyroid abnormalities in 2% to 24% of patients.¹ In 1 study, the median time to onset of amiodarone-induced hypothyroidism was 183 days, with a median onset of hyperthyroidism of 720 days.² Oral amiodarone may also provoke elevations in hepatic transaminases and, rarely, hepatotoxicity.³

Amiodarone may cause pulmonary toxicity, most commonly in the form of interstitial lung disease or hypersensitivity syndrome,⁴ in 1% to 2% of patients and is fatal in approximately 10% of cases. Consequently, a chest x-ray is recommended at baseline and when there is a clinical suspicion for pulmonary toxicity.^4,5 Pulmonary function testing, including diffusing capacity for carbon monoxide, can reveal relatively early changes in pulmonary function that may be attributable to amiodarone-associated pulmonary toxicity; however, the sensitivity of the test for routine screening of drug-induced lung toxicity in general is uncertain.^4–7

In patients who develop unexplained cough and dyspnea while taking amiodarone, a chest CT scan can be helpful for diagnosis.⁵

Corneal microdeposits (epithelial keratopathy) are common, but visual abnormalities and light sensitivity are rare.⁸ Therefore, an ophthalmologic examination is reasonable only if visual abnormalities develop. Amiodarone has also been associated with neurological toxicity, particularly peripheral neuropathy.

Although less common than class III antiarrhythmic agents, oral amiodarone may cause torsades de pointes⁹; 1 analysis estimated the incidence to be 0.7%.¹⁰

Other Antiarrhythmic Drugs

Recommendations for follow-up of other antiarrhythmic agents used for management of AF are provided in Table 25.

Dofetilide

Dofetilide is a potent inhibitor of I_Kr, augments I_Na-L, and is associated with torsades de pointes.⁹ The incidence of torsades de pointes associated with dofetilide ranges from 0.9% to 3.3%.^11,12 However, the incidence can be substantially higher if the dofetilide dose is not properly adjusted for kidney function. Patients undergoing initiation or reloading of dofetilide are generally admitted for at least 3 days to a health care facility that can provide calculations of CrCl, continuous electrocardiographic monitoring, and cardiac resuscitation. After discharge on a stable dose of these drugs, a 12-lead ECG, and levels for serum magnesium, potassium, and creatinine are recommended every 3 to 6 months for assessment of QTc interval duration, electrolyte balance, and renal function, and more frequently for patients concomitantly taking other QT interval-prolonging drugs or with changing kidney function, to minimize the risk of drug-associated torsades de pointes.

Dronedarone

Rare cases of severe liver injury have been reported in association with dronedarone. These cases have occurred 4.5 to 6 months after initiation of dronedarone therapy, although 1 case of hepatotoxicity was reported as early as 2 days after initiation of dronedarone (and was ultimately fatal a few days later),¹³ although a severe case of dronedarone-associated toxic hepatitis occurred 9 months after initiation of therapy.¹⁴

Ibutilide

Ibutilide may prolong the QT interval and cause torsades de pointes^15–22 and therefore continuous electrocardiographic monitoring is recommended during infusion and for 4 hours after completion of ibutilide infusion. Ibutilide is associated with nonsustained ventricular tachycardia (VT), QT interval prolongation, and torsades de pointes. The incidence of nonsustained VT has been reported to be as high as 8.3%.²³ Torsades de pointes occurs in approximately 2% to 7% of patients,^15–22 but the incidence increases substantially in patients with depressed LVEF.²² Most reported cases of ibutilide-associated nonsustained VT and torsades de pointes have occurred within 30 minutes after the last dose; torsades de pointes has rarely been reported as late as 2.5 hours after an ibutilide infusion.²² QTc intervals generally return to baseline within 2 to 3 hours after a dose.

The risk of drug-induced torsades de pointes is higher in patients with hypokalemia and/or hypomagnesemia. Hypokalemia has been reported to be present in 17% to 70% of patients who have developed drug-induced torsades de pointes.^24–27 Similarly, hypomagnesemia has been reported to be a contributing factor to numerous published cases of drug-induced torsades de pointes.^27–29 Therefore, maintenance of serum potassium and magnesium concentrations within the normal range is important before administration of ibutilide, and magnesium supplementation may mitigate the risk of torsades de pointes.

Procainamide

Intravenous procainamide is associated with hypotension in 5% to 12% of patients when used for management of AF,^30,31 and BP monitoring is recommended. Intravenous procainamide is also associated with widening of the QRS complex and prolongation of the QT interval and may provoke proarrhythmia in the form of monomorphic VT³² and torsades de pointes.³³

Sotalol

Sotalol is a potent inhibitor of I_Kr, augments I_Na-L, and is associated with torsades de pointes, with an incidence ranging from 0.4% to 2.3%.^34,35 Patients undergoing initiation or dose escalation of sotalol are often admitted for at least 3 days to a health care facility that can provide calculations of CrCl, continuous electrocardiographic monitoring, and cardiac resuscitation. A new intravenous form of sotalol was recently FDA approved, which may obviate the need for 3-day admission. After discharge on a stable dose of these drugs, a 12-lead ECG, and levels of serum magnesium, potassium, and creatinine are recommended every 3 to 6 months for assessment of QTc interval duration, electrolyte balance, and renal function, and more frequently for patients concomitantly taking other QT interval-prolonging drugs or with changing kidney function, to minimize the risk of drug-associated torsades de pointes.

8.3.4. Upstream Therapy

Pharmacological treatments targeting upstream pathways have included glucocorticoids, ACE inhibitors, ARBs, aldosterone antagonists, statins, omega-3 polyunsaturated fatty acids, antioxidants, and sodium-glucose cotransporter 2 inhibitors (see Section 5, “Lifestyle and Risk Factor Modification for AF Management”). Targeting inflammation, a randomized study of glucocorticoids reduced recurrent AF after first occurrence of persistent AF, but adverse effects inhibit long-term steroid use.^1–9 Statins reduced postoperative AF in small RCTs.^10–14 However, an adequately powered placebo-controlled trial of rosuvastatin did not reduce postoperative AF.¹⁵ Statins do not prevent AF in other cardiovascular settings.¹⁶ RCTs and a meta-analysis provide limited though consistent support for the antioxidant ascorbic acid for postoperative AF.¹⁷ Targeting fibrosis, small or secondary studies of RCTs reported lower new AF with ACE inhibitors or ARBs.^2–9 However, larger RCTs targeting AF failed to reduce recurrent AF.^18–20 RCTs showed that MRAs reduced new-onset atrial arrhythmias in patients with HFrEF along with improvement of other cardiovascular outcomes.^21,22 In patients with type 2 diabetes, HF, or CKD, sodium-glucose cotransporter 2 inhibitors appear to prevent new AF.^23–25 In contrast, omega-3 fatty acids do not appear to reduce AF and, in 1 large study, was associated with higher occurrence of AF.^26–31 With only limited or inconsistent data, no recommendations are made for use of these upstream therapies for prevention of AF.

8.4. AF Catheter Ablation

Synopsis

Catheter ablation has become an established therapy for AF because of multiple RCTs and evidence from large registries and continues to evolve as new technologies are developed. Previous professional society documents have provided different recommendations for catheter ablation dependent on whether AF was persistent or paroxysmal.^25,40 More recent information has shown that ablation for AF is more effective than antiarrhythmic drugs for both persistent and paroxysmal AF and that earlier implementation of rhythm control strategies is an important factor for improving AF ablation success rates.^4,35,41–44 As with all strategies for rhythm control of AF, impact on a patient’s goals of care and QOL should be the focus. For example, significantly reducing the frequency and duration of AF episodes but not eliminating all future episodes of AF may represent a clinically important improvement. Although RCTs have mainly used younger patients (<70 years of age) who also experience the largest benefits, observational studies have reported improvement in QOL with catheter ablation in older patients.^{3–6,8,45–47}

Recommendation-Specific Supportive Text

1. In patients who have not responded to an antiarrhythmic drug due to a high burden of recurrent AF or adverse effects from the medication, RCTs have consistently demonstrated lower risk for recurrent symptomatic AF after ablation when compared with using another antiarrhythmic medication.^1–10 As an example, in STOP-AF, patients who had failed ≥1 antiarrhythmic drug (approximately 70% and 30% for 1 or 2 failed drugs, respectively) were randomized to either another antiarrhythmic drug or catheter ablation. At 1 year follow-up, catheter ablation was associated with a treatment success rate of 70% compared with 7% in the drug arm.¹⁰ Similarly, in the Thermo-cool (NAVISTAR THERMOCOOL Catheter for the Radiofrequency Ablation of Symptomatic Paroxysmal Atrial Fibrillation) trial, patients with paroxysmal AF who had failed 1 antiarrhythmic medication were randomized to catheter ablation or another antiarrhythmic drug. After 9 months, 66% of patients in the catheter ablation group were free from recurrent arrhythmia compared with 16% in the antiarrhythmic drug group.⁸ Finally, most recently, in the CABANA trial, 80% of patients were on an antiarrhythmic medication and thought to be candidates for AF ablation and were randomized to catheter ablation or continued antiarrhythmic therapy. Catheter ablation was associated with a nearly 50% reduction in recurrent AF (HR, 0.52 [95% CI, 0.45-0.60]; P<0.001).^4,28

2. In selected patients with paroxysmal AF, ablation is a suitable first-line option. Several initial RCTs suggested a decrease in recurrent AF or AF burden with catheter ablation when compared with antiarrhythmic drugs.¹⁶ More recent trials have shown a significant reduction in recurrent AF with catheter ablation compared with antiarrhythmic drugs. In a follow-up report of 1 study, after 3-year follow-up, catheter ablation continued to be associated with a significant decrease in recurrent atrial tachyarrhythmias when compared with antiarrhythmic drugs (HR, 0.51 [95% CI, 0.38-0.67]).¹⁴ More importantly, episodes of persistent AF developed in only 1.9% of patients randomized to catheter ablation compared with 7.4% of patients in the antiarrhythmic drug arm (HR, 0.25 [95% CI, 0.13-0.70]).¹⁴ All of the studies that have evaluated catheter ablation as an initial strategy for rhythm control in patients with paroxysmal AF, although fairly standard exclusion criteria were used, enrolled relatively young patients (average age approximately 60 years) who had relatively few comorbidities (if present, mainly hypertension).¹⁶

3. AFL is most commonly due to the critical isthmus formed by the inferior vena cava and the tricuspid valve. More rarely, in patients who have not undergone previous ablation procedures atypical AFL or focal ATs can be observed. Catheter ablation of typical AFL is effective and relatively low risk.^17–19 In an older meta-analysis of AFL ablation studies, AFL ablation was associated with an acute success rate of 90% and a complication rate of 2.6%.¹⁸ The occurrence of AF after AFL ablation was 34%, with a recurrence rate of 23% in those patients without a history of AF compared with 53% in patients with a history of AF. By 5 years, AF developed in 60% to 70% of patients regardless of whether the patient had a history of AF before the AFL ablation. Little evidence is available in these studies for the clinical significance of AF after AFL ablation.

4. AF can be associated with other atrial arrhythmias, particularly AFL.^20–22 During ablation for AF, ablation of previously documented or inducible sustained SVT or AFL is useful to reduce the likelihood of recurrent arrhythmias.^22–25 Although ablation alone for AF in patients with both AF and AFL will reduce the likelihood of AFL, and ablation targeting an inducible SVT in a patient with AF may reduce future AF, in both cases the likelihood of recurrent arrhythmias is high, although the actual recurrence rate will depend on the specific population.^{17,18,25–27,48,49} Conversely, prophylactic catheter ablation of the CTI in patients without documented or inducible AFL likely has minimal benefit.⁵⁰ Alternatively, in 1 study, cryoballoon PVI as first-line treatment for AFL is equally effective compared with standard CTI ablation for preventing recurrence of atrial arrhythmia and better at preventing new-onset AF.⁵¹

5. Several randomized trials have compared catheter ablation to AADs as first-line therapy for AF.^11–13 In the MANTRA-AF study, ablation and antiarrhythmic drugs as first-line therapy for paroxysmal AF was evaluated. Although no differences in AF were identified during the first 18 months, at 24-month follow-up, ablation was associated with a lower AF burden (ablation, 9% versus antiarrhythmic drugs, 18%; P=0.004).¹¹ Subsequent studies have demonstrated similar results: In the EARLY-AF (Early Aggressive Invasive Intervention for Atrial Fibrillation) study, at 1-year follow-up, recurrent atrial arrhythmias were identified in 43% of the ablation group and 68% of the antiarrhythmic drug group (P<0.001), and in the STOP AF First (Cryoballoon Catheter Ablation in an Antiarrhythmic Drug Naive Paroxysmal Atrial Fibrillation) study, recurrent atrial arrhythmias were identified in 25% of the ablation group and 55% of the antiarrhythmic drug group (P<0.001).^12,13 Procedural complications were observed in 2% to 5% of patients in all 3 studies.^11–13 Although AF burden was significantly reduced with ablation compared with antiarrhythmic drug therapy in EARLY-AF, AF burden was low with either strategy (percentage of time in AF: ablation, 0.6% versus antiarrhythmic drugs, 3.9%) and, in all 3 trials, the average age was ≤60 years.^11–13 Although these trials enrolled patients with paroxysmal AF, recent studies have consistently found that catheter ablation is more commonly used and also effective in patients with persistent AF, particularly if of relatively recent onset (<1 year).²⁸

6. Catheter ablation for AF is a costly procedure yet appears to provide value by improving the symptoms of AF and patient QOL. Some of the initially higher costs of the ablation procedure may be offset by reductions in subsequent cardiovascular hospitalizations. Formal economic analyses have been performed alongside 2 randomized controlled clinical trials. Both trials showed higher costs and improved QOL during follow-up, with incremental cost-effectiveness ratios of $58 000 and €51 000 per quality adjusted life-year added, values within the intermediate value range (between $50 000 and $150 000 per quality^{29,30,52–55} adjusted life-year added) by ACC/AHA criteria. Models of the cost-effectiveness of catheter ablation have reported more favorable results, but their results generally depended on assumptions that ablation reduced mortality, or ischemic stroke, or both, which have not been proven.

7. In most patients, continued untreated AF will lead to progression and higher AF burden associated with worse clinical outcomes.^31–39 In patients with subclinical AF <24 hours, progression to overt AF or episodes >24 hours occurs at a rate of 8.8% per year, and those patients who develop AF are more likely to be hospitalized for HF.⁵⁶ Several observational studies have reported AF progression may be improved by using a catheter-based ablation rhythm-control strategy.^{33–35,37,38} As an example, in 1 randomized trial that was stopped prematurely, in 255 patients with paroxysmal AF randomized to catheter ablation or antiarrhythmic medication, freedom from persistent AF or AT was significantly higher with catheter ablation (ablation: 2.4%, versus antiarrhythmic medication, 17.5%, 1-sided P=0.0009).⁵⁷ However, serious adverse events associated with the ablation were observed in 8% of patients.⁵⁷ Despite the potential risks associated with catheter ablation, in a recent analysis of a large registry that specifically evaluated asymptomatic patients, rhythm control was associated with an improvement in the composite outcome of cardiac death, ischemic stroke, and hospitalization for HF.⁵⁸ The impact was largest in those patients who underwent catheter ablation for rhythm control, particularly in those with paroxysmal AF, LA diameters ≤50 mm,⁵⁸ or higher stroke risk (CHA₂DS₂-VASc score ≥3). In a post hoc analysis of EAST-AFNET 4, the magnitude of benefit from a rhythm-control strategy for reducing the primary composite outcome of cardiovascular death, stroke, or hospitalizations for HF or ACS was independent of symptom status, and catheter ablation for rhythm control was used in approximately 20% of patients, whether symptomatic or not.⁵⁹

8.4.1. Patient Selection

Patient selection is the first step and a critically important step in deciding candidacy for catheter ablation.^1–4 Younger patients are likely to derive greater long-term benefit, including delaying AF progression. However, clinical trials have demonstrated improved cardiovascular outcomes with rhythm control, even with median ages in the 70s.⁵ Patients with minimal atrial enlargement have the best outcomes,⁶ whereas increased myocardial fibrosis^7,8 and more persistent forms of AF are associated with higher rates of recurrence after ablation. Time from diagnosis to ablation is also associated with improved outcomes after ablation.^9,10 Although the likelihood of recurrence of AF is 1 factor that should be considered, it is difficult to predict,¹¹ and certain patients may derive even greater benefits from catheter ablation, such as in patients with HFrEF, who have been shown to have improved functional status, LV function, and cardiovascular outcomes.^12–14

In some patients, ablation should be avoided as it is unlikely to be successful due to overwhelming substrate or ongoing physiologic processes that either strongly perpetuate the risk of AF or make maintenance of sinus rhythm unlikely. These situations include but are not limited to advanced infiltrative cardiomyopathies like amyloid, severe mitral stenosis or regurgitation, and cor pulmonale.

8.4.2. Techniques and Technologies for AF Catheter Ablation

Synopsis

Catheter ablation is now established as an effective option for rhythm control in patients with AF and AFL.^1–7,16 However, in patients with AF, additional ablation targets beyond isolation of PVI as a routine strategy have not reduced AF recurrence or burden in RCTs.^8–18 Although achieving durable pulmonary vein ablation in patients with AF has been associated with less recurrent arrhythmias, other endpoints, such as termination of F and noninducibility, have not been associated with improved outcomes.^19–28 Significant complications of AF ablation include stroke and TIAs, pericardial effusion, and vascular complications.¹⁶ The most serious complication that occurs after the ablation procedure is development of an esophageal atrial fistula, although other complications, such as PV stenosis can occur.^8–18

Recommendation-Specific Supportive Text

1. Multiple randomized studies have shown that PVI is more effective than medical therapy for reducing AF burden.^1–7 A recent meta-analysis of 6 RCTs found that strategies that included PVI were associated with a 50% reduction in the development of recurrent AF when compared with strategies that did not include PVI.⁷

2. Studies have evaluated the use of different endpoints immediately after catheter ablation for predicting durable PVI and the likelihood of future arrhythmias with mixed results.^19–28 In the largest randomized study to date, techniques such as waiting for 30 minutes, using adenosine triphosphate to facilitate PV reconnection, or the combination of waiting and using adenosine were not superior to no testing for predicting durable pulmonary vein isolation or reduction in future arrhythmias. Although noninducibility remains important in determining whether ablation for paroxysmal SVT has been successful, noninducibility after AF ablation or termination of AF with ablation have had mixed results for predicting the likelihood of future recurrence.^19–29

   Multiple ablation targets beyond PVI have been evaluated in randomized controlled and observational trials, including the LAA, posterior wall, the ligament of Marshall, or atrial scar.^8–18 Results have been mixed, and no strategy has emerged that is broadly applicable in all patients, likely because the individual mechanisms for AF vary from patient to patient.^8–18 Targeting large areas of atrial tissue may have potential negative consequences, including a higher likelihood of complications such as atrioesophageal fistula, poor LA mechanics, atypical flutters, and increased risk for stroke.^30–35

8.4.3. Management of Recurrent AF After Catheter Ablation

Synopsis

Recurrences of AF are common after a first ablation procedure, occurring in 30% to 40% of patients in contemporary clinical trials.^1,2 In general US practice, 11% of patients undergoing de novo ablation have a repeat ablation by 1 year.³ Patients often experience improved symptoms and QOL even when they experience recurrence after catheter ablation.⁴ Nonetheless, recurrent atrial arrhythmia that leads to symptoms or LV dysfunction requires treatment.