OUP user menu

★ guidelines ★

ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation–executive summary
A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Pa…

Valentin Fuster, Lars E. Rydén, David S. Cannom, Harry J. Crijns, Anne B. Curtis, Kenneth A. Ellenbogen, Jonathan L. Halperin, Jean-Yves Le Heuzey, G. Neal Kay, James E. Lowe, S. Bertil Olsson, Eric N. Prystowsky, Juan Luis Tamargo, Samuel Wann, Silvia G. Priori, Jean-Jacques Blanc, Andrzej Budaj, A. John Camm, Veronica Dean, Jaap W. Deckers, Catherine Despres, Kenneth Dickstein, John Lekakis, Keith McGregor, Marco Metra, Joao Morais, Ady Osterspey, Juan Luis Tamargo, José Luis Zamorano, Sidney C. Smith, Alice K. Jacobs, Cynthia D. Adams, Jeffery L. Anderson, Elliott M. Antman, Jonathan L. Halperin, Sharon Ann Hunt, Rick Nishimura, Joseph P. Ornato, Richard L. Page, Barbara Riegel
DOI: http://dx.doi.org/10.1093/eurheartj/ehl176 1979-2030 First published online: 2 August 2006
  • ACC/AHA/ESC Guidelines
  • atrial fibrillation
  • arrhythmia
  • heart rate
  • anticoagulants
  • antiarrhythmia agents
  • electrophysiology
  • pharmacology


It is important that the medical profession play a significant role in critically evaluating the use of diagnostic procedures and therapies as they are introduced and tested in the detection, management, or prevention of disease states. Rigorous and expert analysis of the available data documenting absolute and relative benefits and risks of those procedures and therapies can produce helpful guidelines that improve the effectiveness of care, optimize patient outcomes, and favorably affect the overall cost of care by focusing resources on the most effective strategies.

The American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) have jointly engaged in the production of such guidelines in the area of cardiovascular disease since 1980. The ACC/AHA Task Force on Practice Guidelines, whose charge is to develop, update, or revise practice guidelines for important cardiovascular diseases and procedures, directs this effort. The Task Force is pleased to have this guideline developed in conjunction with the European Society of Cardiology (ESC). Writing committees are charged with the task of performing an assessment of the evidence and acting as an independent group of authors to develop or update written recommendations for clinical practice.

Experts in the subject under consideration have been selected from all 3 organizations to examine subject-specific data and to write guidelines. The process includes additional representatives from other medical practitioner and specialty groups when appropriate. Writing committees are specifically charged to perform a formal literature review, weigh the strength of evidence for or against a particular treatment or procedure, and include estimates of expected health outcomes where data exist. Patient-specific modifiers, comorbidities, and issues of patient preference that might influence the choice of particular tests or therapies are considered as well as frequency of follow-up and cost-effectiveness. When available, information from studies on cost will be considered; however, review of data on efficacy and clinical outcomes will constitute the primary basis for preparing recommendations in these guidelines.

The ACC/AHA Task Force on Practice Guidelines and the ESC Committee for Practice Guidelines make every effort to avoid any actual, potential, or perceived conflict of interest that might arise as a result of an outside relationship or personal interest of the Writing Committee. Specifically, all members of the Writing Committee and peer reviewers of the document are asked to provide disclosure statements of all such relationships that might be perceived as real or potential conflicts of interest. Writing Committee members are also strongly encouraged to declare a previous relationship with industry that might be perceived as relevant to guideline development. If a Writing Committee member develops a new relationship with industry during his or her tenure, he or she is required to notify guideline staff in writing. The continued participation of the Writing Committee member will be reviewed. These statements are reviewed by the parent Task Force, reported orally to all members of the Writing Committee at each meeting, and updated and reviewed by the Writing Committee as changes occur. Please refer to the methodology manuals for further description of the policies used in guideline development, including relationships with industry, available on the ACC, AHA, and ESC World Wide Web sites (http://www.acc.org/clinical/manual/manual_introltr.htm, http://circ.ahajournals.org/manual/ and http://www.escardio.org/knowledge/guidelines/Rules/). Please see Appendix I for author relationships with industry and Appendix II for peer reviewer relationships with industry that are pertinent to these guidelines.

These practice guidelines are intended to assist healthcare providers in clinical decision making by describing a range of generally acceptable approaches for the diagnosis, management, and prevention of specific diseases and conditions. These guidelines attempt to define practices that meet the needs of most patients in most circumstances. These guideline recommendations reflect a consensus of expert opinion after a thorough review of the available, current scientific evidence and are intended to improve patient care. If these guidelines are used as the basis for regulatory/payer decisions, the ultimate goal is quality of care and serving the patient's best interests. The ultimate judgment regarding care of a particular patient must be made by the healthcare provider and the patient in light of all of the circumstances presented by that patient. There are circumstances in which deviations from these guidelines are appropriate.

The guidelines will be reviewed annually by the ACC/AHA Task Force on Practice Guidelines and the ESC Committee for Practice Guidelines and will be considered current unless they are updated, revised, or sunsetted and withdrawn from distribution. The executive summary and recommendations are published in the August 15, 2006, issue of the Journal of the American College of Cardiology, August 15, 2006, issue of Circulation, and August 16, 2006, issue of the European Heart Journal. The full-text guidelines are e-published in the same issues of the Journal of the American College of Cardiology and Circulation, published in September 9, 2006, issue of Europace, as well as posted on the ACC (www.acc.org), AHA (www.americanheart.org), and ESC (www.escardio.org) World Wide Web sites. Copies of the full text and the executive summary are available from all 3 organizations.

Sidney C. Smith, Jr., MD, FACC, FAHA, FESC, Chair, ACC/AHA Task Force on Practice Guidelines

Silvia G. Priori, MD, PhD, FESC, Chair, ESC Committee for Practice Guidelines

I. Introduction

A. Organization of committee and evidence review

Atrial fibrillation (AF) is the most common sustained cardiac rhythm disturbance, increasing in prevalence with age. AF is often associated with structural heart disease, although a substantial proportion of patients with AF have no detectable heart disease. Hemodynamic impairment and thromboembolic events related to AF result in significant morbidity, mortality, and cost. Accordingly, the American College of Cardiology (ACC), the American Heart Association (AHA), and the European Society of Cardiology (ESC) created a committee to establish guidelines for optimum management of this frequent and complex arrhythmia.

The committee was composed of representatives of the ACC, AHA, ESC, the European Heart Rhythm Association (EHRA), and the Heart Rhythm Society (HRS). The document was reviewed by reviewers nominated by these organizations and will be reviewed annually by the Task Force and considered current unless the Task Force revises or withdraws it from distribution.

The ACC/AHA/ESC Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation conducted a comprehensive review of the relevant literature from 2001 to 2006 using the PubMed/MEDLINE and Cochrane Library databases. Searches focused on English-language sources and studies in human subjects. Articles related to animal experimentation were cited when important to understanding concepts pertinent to patient management.

Classification of recommendations

  • Class I: Conditions for which there is evidence and/or general agreement that a given procedure/therapy is beneficial, useful, and effective.

  • Class II: Conditions for which there is conflicting evidence and/or a divergence of opinion about the usefulness/efficacy of performing the procedure/therapy.

    • Class IIa: Weight of evidence/opinion is in favor of usefulness/efficacy.

    • Class IIb: Usefulness/efficacy is less well established by evidence/opinion.

  • Class III: Conditions for which there is evidence and/or general agreement that a procedure/therapy is not useful or effective and in some cases may be harmful.

Level of evidence

The weight of evidence was ranked from highest (A) to lowest (C), as follows:

  • Level of Evidence A: Data derived from multiple randomized clinical trials or meta-analyses.

  • Level of Evidence B: Data derived from a single randomized trial, or nonrandomized studies.

  • Level of Evidence C: Only consensus opinion of experts, case studies, or standard-of-care.

B. Changes since the initial publication of these guidelines in 2001

The Writing Committee considered evidence published since 2001 and drafted revised recommendations to incorporate results from major clinical trials such as those that compared rhythm control and rate control approaches to long-term management. The text has been reorganized to reflect the implications for patient care, beginning with recognition of AF and its pathogenesis and the general priorities of rate control, prevention of thromboembolism, and methods available for use in selected patients to correct the arrhythmia and maintain normal sinus rhythm. Advances in catheter-based ablation technologies are incorporated in expanded sections and recommendations, with the recognition that such vital details as patient selection, optimum catheter positioning, absolute rates of treatment success, and the frequency of complications remain incompletely defined. Sections on drug therapy have been confined to human studies with compounds approved for clinical use in North America and/or Europe. As data on the management of patients prone to AF in special circumstances are more robust, recommendations are based on a higher level of evidence than in the first edition of these guidelines. Every effort was made to maintain consistency with other ACC/AHA and ESC practice guidelines.

C. Recommendations for management of patients with atrial fibrillation

Classification of Recommendations and Level of Evidence are expressed in the ACC/AHA/ESC format as follows and described in Table 1. Recommendations are evidence based and derived primarily from published data. The reader is referred to the full-text guidelines for a complete description of the rationale and evidence supporting these recommendations.

Table 1

Applying Classification of Recommendations and Level of Evidenceb


1. Pharmacological rate control during atrial fibrillation

Class I

  1. Measurement of the heart rate at rest and control of the rate using pharmacological agents (either a beta blocker or nondihydropyridine calcium channel antagonist, in most cases) are recommended for patients with persistent or permanent AF. (Level of Evidence: B)

  2. In the absence of preexcitation, intravenous administration of beta blockers (esmolol, metoprolol, or propranolol) or nondihydropyridine calcium channel antagonists (verapamil, diltiazem) is recommended to slow the ventricular response to AF in the acute setting, exercising caution in patients with hypotension or heart failure (HF). (Level of Evidence: B)

  3. Intravenous administration of digoxin or amiodarone is recommended to control the heart rate in patients with AF and HF who do not have an accessory pathway. (Level of Evidence: B)

  4. In patients who experience symptoms related to AF during activity, the adequacy of heart rate control should be assessed during exercise, adjusting pharmacological treatment as necessary to keep the rate in the physiological range. (Level of Evidence: C)

  5. Digoxin is effective following oral administration to control the heart rate at rest in patients with AF and is indicated for patients with HF, left ventricular (LV) dysfunction, or for sedentary individuals. (Level of Evidence: C)

Class IIa

  1. A combination of digoxin and either a beta blocker or nondihydropyridine calcium channel antagonist is reasonable to control the heart rate both at rest and during exercise in patients with AF. The choice of medication should be individualized and the dose modulated to avoid bradycardia. (Level of Evidence: B)

  2. It is reasonable to use ablation of the AV node or accessory pathway to control heart rate when pharmacological therapy is insufficient or associated with side effects. (Level of Evidence: B)

  3. Intravenous amiodarone can be useful to control the heart rate in patients with AF when other measures are unsuccessful or contraindicated. (Level of Evidence: C)

  4. When electrical cardioversion is not necessary in patients with AF and an accessory pathway, intravenous procainamide or ibutilide is a reasonable alternative. (Level of Evidence: C)

Class IIb

  1. When the ventricular rate cannot be adequately controlled both at rest and during exercise in patients with AF using a beta blocker, nondihydropyridine calcium channel antagonist, or digoxin, alone or in combination, oral amiodarone may be administered to control the heart rate. (Level of Evidence: C)

  2. Intravenous procainamide, disopyramide, ibutilide, or amiodarone may be considered for hemodynamically stable patients with AF involving conduction over an accessory pathway. (Level of Evidence: B)

  3. When the rate cannot be controlled with pharmacological agents or tachycardia-mediated cardiomyopathy is suspected, catheter-directed ablation of the AV node may be considered in patients with AF to control the heart rate. (Level of Evidence: C)

Class III

  1. Digitalis should not be used as the sole agent to control the rate of ventricular response in patients with paroxysmal AF. (Level of Evidence: B)

  2. Catheter ablation of the AV node should not be attempted without a prior trial of medication to control the ventricular rate in patients with AF. (Level of Evidence: C)

  3. In patients with decompensated HF and AF, intravenous administration of a nondihydropyridine calcium channel antagonist may exacerbate hemodynamic compromise and is not recommended. (Level of Evidence: C)

  4. Intravenous administration of digitalis glycosides or nondihydropyridine calcium channel antagonists to patients with AF and a preexcitation syndrome may paradoxically accelerate the ventricular response and is not recommended. (Level of Evidence: C)

2. Preventing thromboembolism

(For recommendations regarding antithrombotic therapy in patients with AF undergoing cardioversion, see Section I.C.3.d.)

Class I

  1. Antithrombotic therapy to prevent thromboembolism is recommended for all patients with AF, except those with lone AF or contraindications. (Level of Evidence: A)

  2. The selection of the antithrombotic agent should be based upon the absolute risks of stroke and bleeding and the relative risk and benefit for a given patient. (Level of Evidence: A)

  3. For patients without mechanical heart valves at high risk of stroke, chronic oral anticoagulant therapy with a vitamin K antagonist is recommended in a dose adjusted to achieve the target intensity international normalized ratio (INR) of 2.0 to 3.0, unless contraindicated. Factors associated with highest risk for stroke in patients with AF are prior thromboembolism (stroke, transient ischemic attack [TIA], or systemic embolism) and rheumatic mitral stenosis. (Level of Evidence: A)

  4. Anticoagulation with a vitamin K antagonist is recommended for patients with more than 1 moderate risk factor. Such factors include age 75 y or greater, hypertension, HF, impaired LV systolic function (ejection fraction 35% or less or fractional shortening less than 25%), and diabetes mellitus. (Level of Evidence: A)

  5. INR should be determined at least weekly during initiation of therapy and monthly when anticoagulation is stable. (Level of Evidence: A)

  6. Aspirin, 81–325 mg daily, is recommended as an alternative to vitamin K antagonists in low-risk patients or in those with contraindications to oral anticoagulation. (Level of Evidence: A)

  7. For patients with AF who have mechanical heart valves, the target intensity of anticoagulation should be based on the type of prosthesis, maintaining an INR of at least 2.5. (Level of Evidence: B)

  8. Antithrombotic therapy is recommended for patients with atrial flutter as for those with AF. (Level of Evidence: C)

Class IIa

  1. For primary prevention of thromboembolism in patients with nonvalvular AF who have just 1 of the following validated risk factors, antithrombotic therapy with either aspirin or a vitamin K antagonist is reasonable, based upon an assessment of the risk of bleeding complications, ability to safely sustain adjusted chronic anticoagulation, and patient preferences: age greater than or equal to 75 y (especially in female patients), hypertension, HF, impaired LV function, or diabetes mellitus. (Level of Evidence: A)

  2. For patients with nonvalvular AF who have 1 or more of the following less well-validated risk factors, antithrombotic therapy with either aspirin or a vitamin K antagonist is reasonable for prevention of thromboembolism: age 65 to 74 y, female gender, or CAD. The choice of agent should be based upon the risk of bleeding complications, ability to safely sustain adjusted chronic anticoagulation, and patient preferences. (Level of Evidence: B)

  3. It is reasonable to select antithrombotic therapy using the same criteria irrespective of the pattern (i.e., paroxysmal, persistent, or permanent) of AF. (Level of Evidence: B)

  4. In patients with AF who do not have mechanical prosthetic heart valves, it is reasonable to interrupt anticoagulation for up to 1 wk without substituting heparin for surgical or diagnostic procedures that carry a risk of bleeding. (Level of Evidence: C)

  5. It is reasonable to reevaluate the need for anticoagulation at regular intervals. (Level of Evidence: C)

Class IIb

  1. In patients 75 y of age and older at increased risk of bleeding but without frank contraindications to oral anticoagulant therapy, and in other patients with moderate risk factors for thromboembolism who are unable to safely tolerate anticoagulation at the standard intensity of INR 2.0 to 3.0, a lower INR target of 2.0 (range 1.6 to 2.5) may be considered for primary prevention of ischemic stroke and systemic embolism. (Level of Evidence: C)

  2. When surgical procedures require interruption of oral anticoagulant therapy for longer than 1 wk in high-risk patients, unfractionated heparin may be administered or low-molecular-weight heparin given by subcutaneous injection, although the efficacy of these alternatives in this situation is uncertain. (Level of Evidence: C)

  3. Following percutaneous coronary intervention or revascularization surgery in patients with AF, low-dose aspirin (less than 100 mg per d) and/or clopidogrel (75 mg per d) may be given concurrently with anticoagulation to prevent myocardial ischemic events, but these strategies have not been thoroughly evaluated and are associated with an increased risk of bleeding. (Level of Evidence: C)

  4. In patients undergoing percutaneous coronary intervention, anticoagulation may be interrupted to prevent bleeding at the site of peripheral arterial puncture, but the vitamin K antagonist should be resumed as soon as possible after the procedure and the dose adjusted to achieve an INR in the therapeutic range. Aspirin may be given temporarily during the hiatus, but the maintenance regimen should then consist of the combination of clopidogrel, 75 mg daily, plus warfarin (INR 2.0 to 3.0). Clopidogrel should be given for a minimum of 1 mo after implantation of a bare metal stent, at least 3 mo for a sirolimus-eluting stent, at least 6 mo for a paclitaxel-eluting stent, and 12 mo or longer in selected patients, following which warfarin may be continued as monotherapy in the absence of a subsequent coronary event. When warfarin is given in combination with clopidogrel or low-dose aspirin, the dose intensity must be carefully regulated. (Level of Evidence: C)

  5. In patients with AF younger than 60 y without heart disease or risk factors for thromboembolism (lone AF), the risk of thromboembolism is low without treatment and the effectiveness of aspirin for primary prevention of stroke relative to the risk of bleeding has not been established. (Level of Evidence: C)

  6. In patients with AF who sustain ischemic stroke or systemic embolism during treatment with low-intensity anticoagulation (INR 2.0 to 3.0), rather than add an antiplatelet agent, it may be reasonable to raise the intensity of the anticoagulation to a maximum target INR of 3.0 to 3.5. (Level of Evidence: C)

Class III

Long-term anticoagulation with a vitamin K antagonist is not recommended for primary prevention of stroke in patients below the age of 60 y without heart disease (lone AF) or any risk factors for thromboembolism. (Level of Evidence: C)

3. Cardioversion of atrial fibrillation

a. Pharmacological cardioversion

Class I

Administration of flecainide, dofetilide, propafenone, or ibutilide is recommended for pharmacological cardioversion of AF. (Level of Evidence: A)

Class IIa

  1. Administration of amiodarone is a reasonable option for pharmacological cardioversion of AF. (Level of Evidence: A)

  2. A single oral bolus dose of propafenone or flecainide (‘pill-in-the-pocket’) can be administered to terminate persistent AF outside the hospital once treatment has proved safe in hospital for selected patients without sinus or AV node dysfunction, bundle-branch block, QT-interval prolongation, the Brugada syndrome, or structural heart disease. Before antiarrhythmic medication is initiated, a beta blocker or nondihydropyridine calcium channel antagonist should be given to prevent rapid AV conduction in the event atrial flutter occurs. (Level of Evidence: C)

  3. Administration of amiodarone can be beneficial on an outpatient basis in patients with paroxysmal or persistent AF when rapid restoration of sinus rhythm is not deemed necessary. (Level of Evidence: C)

Class IIb

Administration of quinidine or procainamide might be considered for pharmacological cardioversion of AF, but the usefulness of these agents is not well established. (Level of Evidence: C)

Class III

  1. Digoxin and sotalol may be harmful when used for pharmacological cardioversion of AF and are not recommended. (Level of Evidence: A)

  2. Quinidine, procainamide, disopyramide, and dofetilide should not be started out of hospital for conversion of AF to sinus rhythm. (Level of Evidence: B)

b. Direct-current cardioversion

Class I

  1. When a rapid ventricular response does not respond promptly to pharmacological measures for patients with AF with ongoing myocardial ischemia, symptomatic hypotension, angina, or HF, immediate R-wave synchronized direct-current cardioversion is recommended. (Level of Evidence: C)

  2. Immediate direct-current cardioversion is recommended for patients with AF involving preexcitation when very rapid tachycardia or hemodynamic instability occurs. (Level of Evidence: B)

  3. Cardioversion is recommended in patients without hemodynamic instability when symptoms of AF are unacceptable to the patient. In case of early relapse of AF after cardioversion, repeated direct-current cardioversion attempts may be made following administration of antiarrhythmic medication. (Level of Evidence: C)

Class IIa

  1. Direct-current cardioversion can be useful to restore sinus rhythm as part of a long-term management strategy for patients with AF. (Level of Evidence: B)

  2. Patient preference is a reasonable consideration in the selection of infrequently repeated cardioversions for the management of symptomatic or recurrent AF. (Level of Evidence: C)

Class III

  1. Frequent repetition of direct-current cardioversion is not recommended for patients who have relatively short periods of sinus rhythm between relapses of AF after multiple cardioversion procedures despite prophylactic antiarrhythmic drug therapy. (Level of Evidence: C)

  2. Electrical cardioversion is contraindicated in patients with digitalis toxicity or hypokalemia. (Level of Evidence: C)

c. Pharmacological enhancement of direct-current cardioversion

Class IIa

  1. Pretreatment with amiodarone, flecainide, ibutilide, propafenone, or sotalol can be useful to enhance the success of direct-current cardioversion and prevent recurrent AF. (Level of Evidence: B)

  2. In patients who relapse to AF after successful cardioversion, it can be useful to repeat the procedure following prophylactic administration of antiarrhythmic medication. (Level of Evidence: C)

Class IIb

  1. For patients with persistent AF, administration of beta blockers, disopyramide, diltiazem, dofetilide, procainamide, or verapamil may be considered, although the efficacy of these agents to enhance the success of direct-current cardioversion or to prevent early recurrence of AF is uncertain. (Level of Evidence: C)

  2. Out-of-hospital initiation of antiarrhythmic medications may be considered in patients without heart disease to enhance the success of cardioversion of AF. (Level of Evidence: C)

  3. Out-of-hospital administration of antiarrhythmic medications may be considered to enhance the success of cardioversion of AF in patients with certain forms of heart disease once the safety of the drug has been verified for the patient. (Level of Evidence: C)

d. Prevention of thromboembolism in patients with atrial fibrillation undergoing cardioversion

Class I

  1. For patients with AF of 48-h duration or longer, or when the duration of AF is unknown, anticoagulation (INR 2.0 to 3.0) is recommended for at least 3 wk prior to and 4 wk after cardioversion, regardless of the method (electrical or pharmacological) used to restore sinus rhythm. (Level of Evidence: B)

  2. For patients with AF of more than 48-h duration requiring immediate cardioversion because of hemodynamic instability, heparin should be administered concurrently (unless contraindicated) by an initial intravenous bolus injection followed by a continuous infusion in a dose adjusted to prolong the activated partial thromboplastin time to 1.5 to 2 times the reference control value. Thereafter, oral anticoagulation (INR 2.0 to 3.0) should be provided for at least 4 wk, as for patients undergoing elective cardioversion. Limited data support subcutaneous administration of low-molecular-weight heparin in this indication. (Level of Evidence: C)

  3. For patients with AF of less than 48-h duration associated with hemodynamic instability (angina pectoris, myocardial infarction [MI], shock, or pulmonary edema), cardioversion should be performed immediately without delay for prior initiation of anticoagulation. (Level of Evidence: C)

Class IIa

  • (1) During the 48 h after onset of AF, the need for anticoagulation before and after cardioversion may be based on the patient's risk of thromboembolism. (Level of Evidence: C)

  • (2) As an alternative to anticoagulation prior to cardioversion of AF, it is reasonable to perform transesophageal echocardiography (TEE) in search of thrombus in the left atrium (LA) or left atrial appendage (LAA). (Level of Evidence: B)

  • (2a) For patients with no identifiable thrombus, cardioversion is reasonable immediately after anticoagulation with unfractionated heparin (e.g., initiated by intravenous bolus injection and an infusion continued at a dose adjusted to prolong the activated partial thromboplastin time to 1.5 to 2 times the control value until oral anticoagulation has been established with an oral vitamin K antagonist (e.g., warfarin) as evidenced by an INR equal to or greater than 2.0). (Level of Evidence: B) Thereafter, continuation of oral anticoagulation (INR 2.0 to 3.0) is reasonable for a total anticoagulation period of at least 4 wk, as for patients undergoing elective cardioversion. (Level of Evidence: B) Limited data are available to support the subcutaneous administration of a low-molecular-weight heparin in this indication. (Level of Evidence: C)

  • (2b) For patients in whom thrombus is identified by TEE, oral anticoagulation (INR 2.0 to 3.0) is reasonable for at least 3 wk prior to and 4 wk after restoration of sinus rhythm, and a longer period of anticoagulation may be appropriate even after apparently successful cardioversion, because the risk of thromboembolism often remains elevated in such cases. (Level of Evidence: C)

  • (3) For patients with atrial flutter undergoing cardioversion, anticoagulation can be beneficial according to the recommendations as for patients with AF. (Level of Evidence: C)

4. Maintenance of sinus rhythm

Class I

Before initiating antiarrhythmic drug therapy, treatment of precipitating or reversible causes of AF is recommended. (Level of Evidence: C)

Class IIa

  1. Pharmacological therapy can be useful in patients with AF to maintain sinus rhythm and prevent tachycardia-induced cardiomyopathy. (Level of Evidence: C)

  2. Infrequent, well-tolerated recurrence of AF is reasonable as a successful outcome of antiarrhythmic drug therapy. (Level of Evidence: C)

  3. Outpatient initiation of antiarrhythmic drug therapy is reasonable in patients with AF who have no associated heart disease when the agent is well tolerated. (Level of Evidence: C)

  4. In patients with lone AF without structural heart disease, initiation of propafenone or flecainide can be beneficial on an outpatient basis in patients with paroxysmal AF who are in sinus rhythm at the time of drug initiation. (Level of Evidence: B)

  5. Sotalol can be beneficial in outpatients in sinus rhythm with little or no heart disease, prone to paroxysmal AF, if the baseline uncorrected QT interval is less than 460 ms, serum electrolytes are normal, and risk factors associated with class III drug-related proarrhythmia are not present. (Level of Evidence: C)

  6. Catheter ablation is a reasonable alternative to pharmacological therapy to prevent recurrent AF in symptomatic patients with little or no LA enlargement. (Level of Evidence: C)

Class III

  1. Antiarrhythmic therapy with a particular drug is not recommended for maintenance of sinus rhythm in patients with AF who have well-defined risk factors for proarrhythmia with that agent. (Level of Evidence: A)

  2. Pharmacological therapy is not recommended for maintenance of sinus rhythm in patients with advanced sinus node disease or atrioventricular (AV) node dysfunction unless they have a functioning electronic cardiac pacemaker. (Level of Evidence: C)

5. Special considerations

a. Postoperative atrial fibrillation

Class I

  1. Unless contraindicated, treatment with an oral beta blocker to prevent postoperative AF is recommended for patients undergoing cardiac surgery. (Level of Evidence: A)

  2. Administration of AV nodal blocking agents is recommended to achieve rate control in patients who develop postoperative AF. (Level of Evidence: B)

Class IIa

  1. Preoperative administration of amiodarone reduces the incidence of AF in patients undergoing cardiac surgery and represents appropriate prophylactic therapy for patients at high risk for postoperative AF. (Level of Evidence: A)

  2. It is reasonable to restore sinus rhythm by pharmacological cardioversion with ibutilide or direct-current cardioversion in patients who develop postoperative AF as advised for nonsurgical patients. (Level of Evidence: B)

  3. It is reasonable to administer antiarrhythmic medications in an attempt to maintain sinus rhythm in patients with recurrent or refractory postoperative AF, as recommended for other patients who develop AF. (Level of Evidence: B)

  4. It is reasonable to administer antithrombotic medication in patients who develop postoperative AF, as recommended for nonsurgical patients. (Level of Evidence: B)

Class IIb

Prophylactic administration of sotalol may be considered for patients at risk of developing AF following cardiac surgery. (Level of Evidence: B)

b. Acute myocardial infarction

Class I

  1. Direct-current cardioversion is recommended for patients with severe hemodynamic compromise or intractable ischemia, or when adequate rate control cannot be achieved with pharmacological agents in patients with acute MI and AF. (Level of Evidence: C)

  2. Intravenous administration of amiodarone is recommended to slow a rapid ventricular response to AF and improve LV function in patients with acute MI. (Level of Evidence: C)

  3. Intravenous beta blockers and nondihydropyridine calcium antagonists are recommended to slow a rapid ventricular response to AF in patients with acute MI who do not display clinical LV dysfunction, bronchospasm, or AV block. (Level of Evidence: C)

  4. For patients with AF and acute MI, administration of unfractionated heparin by either continuous intravenous infusion or intermittent subcutaneous injection is recommended in a dose sufficient to prolong the activated partial thromboplastin time to 1.5 to 2 times the control value, unless contraindications to anticoagulation exist. (Level of Evidence: C)

Class IIa

Intravenous administration of digitalis is reasonable to slow a rapid ventricular response and improve LV function in patients with acute MI and AF associated with severe LV dysfunction and HF. (Level of Evidence: C)

Class III

The administration of class IC antiarrhythmic drugs is not recommended in patients with AF in the setting of acute MI. (Level of Evidence: C)

c. Management of atrial fibrillation associated with the Wolff-Parkinson-White (WPW) preexcitation syndrome

Class I

  1. Catheter ablation of the accessory pathway is recommended in symptomatic patients with AF who have WPW syndrome, particularly those with syncope due to rapid heart rate or those with a short bypass tract refractory period. (Level of Evidence: B)

  2. Immediate direct-current cardioversion is recommended to prevent ventricular fibrillation in patients with a short anterograde bypass tract refractory period in whom AF occurs with a rapid ventricular response associated with hemodynamic instability. (Level of Evidence: B)

  3. Intravenous procainamide or ibutilide is recommended to restore sinus rhythm in patients with WPW in whom AF occurs without hemodynamic instability in association with a wide QRS complex on the electrocardiogram (ECG) (greater than or equal to 120-ms duration) or with a rapid preexcited ventricular response. (Level of Evidence: C)

Class IIa

Intravenous flecainide or direct-current cardioversion is reasonable when very rapid ventricular rates occur in patients with AF involving conduction over an accessory pathway. (Level of Evidence: B)

Class IIb

It may be reasonable to administer intravenous quinidine, procainamide, disopyramide, ibutilide, or amiodarone to hemodynamically stable patients with AF involving conduction over an accessory pathway. (Level of Evidence: B)

Class III

Intravenous administration of digitalis glycosides or nondihydropyridine calcium channel antagonists is not recommended in patients with WPW syndrome who have preexcited ventricular activation during AF. (Level of Evidence: B)

d. Hyperthyroidism

Class I

  1. Administration of a beta blocker is recommended to control the rate of ventricular response in patients with AF complicating thyrotoxicosis, unless contraindicated. (Level of Evidence: B)

  2. In circumstances when a beta blocker cannot be used, administration of a nondihydropyridine calcium channel antagonist (diltiazem or verapamil) is recommended to control the ventricular rate in patients with AF and thyrotoxicosis. (Level of Evidence: B)

  3. In patients with AF associated with thyrotoxicosis, oral anticoagulation (INR 2.0 to 3.0) is recommended to prevent thromboembolism, as recommended for AF patients with other risk factors for stroke. (Level of Evidence: C)

  4. Once a euthyroid state is restored, recommendations for antithrombotic prophylaxis are the same as for patients without hyperthyroidism. (Level of Evidence: C)

e. Management of atrial fibrillation during pregnancy

Class I

  1. Digoxin, a beta blocker, or a nondihydropyridine calcium channel antagonist is recommended to control the rate of ventricular response in pregnant patients with AF. (Level of Evidence: C)

  2. Direct-current cardioversion is recommended in pregnant patients who become hemodynamically unstable due to AF. (Level of Evidence: C)

  3. Protection against thromboembolism is recommended throughout pregnancy for all patients with AF (except those with lone AF and/or low thromboembolic risk). Therapy (anticoagulant or aspirin) should be chosen according to the stage of pregnancy. (Level of Evidence: C)

Class IIb

  1. Administration of heparin may be considered during the first trimester and last month of pregnancy for patients with AF and risk factors for thromboembolism. Unfractionated heparin may be administered either by continuous intravenous infusion in a dose sufficient to prolong the activated partial thromboplastin time to 1.5 to 2 times the control value or by intermittent subcutaneous injection in a dose of 10 000 to 20 000 units every 12 h, adjusted to prolong the mid-interval (6 h after injection) activated partial thromboplastin time to 1.5 times control. (Level of Evidence: B)

  2. Despite the limited data available, subcutaneous administration of low-molecular-weight heparin may be considered during the first trimester and last month of pregnancy for patients with AF and risk factors for thromboembolism. (Level of Evidence: C)

  3. Administration of an oral anticoagulant may be considered during the second trimester for pregnant patients with AF at high thromboembolic risk. (Level of Evidence: C)

  4. Administration of quinidine or procainamide may be considered to achieve pharmacological cardioversion in hemodynamically stable patients who develop AF during pregnancy. (Level of Evidence: C)

f. Management of atrial fibrillation in patients with hypertrophic cardiomyopathy (HCM)

Class I

Oral anticoagulation (INR 2.0 to 3.0) is recommended in patients with HCM who develop AF, as for other patients at high risk of thromboembolism. (Level of Evidence: B)

Class IIa

Antiarrhythmic medications can be useful to prevent recurrent AF in patients with HCM. Available data are insufficient to recommend one agent over another in this situation, but (a) disopyramide combined with a beta blocker or nondihydropyridine calcium channel antagonist or (b) amiodarone alone is generally preferred. (Level of Evidence: C)

g. Management of atrial fibrillation in patients with pulmonary disease

Class I

  1. Correction of hypoxemia and acidosis is the recommended primary therapeutic measure for patients who develop AF during an acute pulmonary illness or exacerbation of chronic pulmonary disease. (Level of Evidence: C)

  2. A nondihydropyridine calcium channel antagonist (diltiazem or verapamil) is recommended to control the ventricular rate in patients with obstructive pulmonary disease who develop AF. (Level of Evidence: C)

  3. Direct-current cardioversion should be attempted in patients with pulmonary disease who become hemodynamically unstable as a consequence of AF. (Level of Evidence: C)

Class III

  1. Theophylline and beta-adrenergic agonist agents are not recommended in patients with bronchospastic lung disease who develop AF. (Level of Evidence: C)

  2. Beta blockers, sotalol, propafenone, and adenosine are not recommended in patients with obstructive lung disease who develop AF. (Level of Evidence: C)

II. Definition

A. Atrial fibrillation

AF is a supraventricular tachyarrhythmia characterized by uncoordinated atrial activation with consequent deterioration of mechanical function. On the ECG, rapid oscillations, or fibrillatory waves that vary in amplitude, shape, and timing, replace consistent P waves, and there is an irregular ventricular response that is rapid when conduction is intact.1 The ventricular response depends on electrophysiological properties of the AV node and other conducting tissues, vagal and sympathetic tone, the presence or absence of accessory pathways, and the action of drugs.2 When AV block or ventricular or AV junctional tachycardia is present, the cardiac cycles (R-R intervals) may be regular. In patients with pacemakers, diagnosis of AF may require pacemaker inhibition to expose fibrillatory activity. An irregular, sustained, wide-QRS-complex tachycardia suggests AF with conduction over an accessory pathway or AF with bundle-branch block. Atrial flutter is usually readily distinguished from AF. Extremely rapid rates (greater than 200 beats per minute) suggest an accessory pathway or ventricular tachycardia.

B. Related arrhythmias

AF may occur in association with atrial flutter or atrial tachycardia. The typical form of atrial flutter is characterized by a saw-tooth pattern of regular atrial activation called flutter (f) waves on the ECG, particularly visible in leads II, III, aVF, and V1. If untreated, the atrial rate typically ranges from 240 to 320 beats per minute, with f waves inverted in ECG leads II, III, and aVF and upright in lead V1. The direction of activation in the right atrium (RA) may be reversed, resulting in upright f waves in leads II, III, and aVF and inversion in lead V1. Atrial flutter may degenerate into AF, and AF may convert to atrial flutter. Atrial flutter is usually readily distinguished from AF, but misdiagnosis may occur when fibrillatory atrial activity is prominent in more than 1 ECG lead.3

Focal atrial tachycardias, AV reentrant tachycardias, and AV nodal reentrant tachycardias may also trigger AF. In these tachycardias, distinct P waves are typically separated by an isoelectric baseline, and their morphology may localize the origin of the arrhythmia.

III. Classification

Various classification systems have been proposed for AF based on the ECG pattern,1 epicardial4 or endocavitary recordings, mapping of atrial electrical activity, or clinical features. Although the pattern of AF can change over time, it may be helpful to characterize the arrhythmia at a given moment. The classification scheme recommended here represents a consensus driven by a desire for simplicity and clinical relevance.

The clinician should distinguish a first-detected episode of AF, whether or not symptomatic or self-limited, recognizing the uncertainty about the actual duration of the episode and about previous undetected episodes (Figure 1). After 2 or more episodes, AF is considered recurrent. If the arrhythmia terminates spontaneously, recurrent AF is designated paroxysmal; when sustained beyond 7 d, it is termed persistent. Termination with pharmacological therapy or direct-current cardioversion does not alter the designation. First-detected AF may be either paroxysmal or persistent. The category of persistent AF also includes cases of long-standing AF (e.g., longer than 1 y), usually leading to permanent AF, in which cardioversion has failed or has been foregone.

Figure 1

Patterns of atrial fibrillation (AF). 1, Episodes that generally last 7 d or less (most less than 24 h); 2, episodes that usually last more than 7 d; 3, cardioversion failed or not attempted; and 4, both paroxysmal and persistent AF may be recurrent.

These categories are not mutually exclusive, and a particular patient may have several episodes of paroxysmal AF and occasional persistent AF, or the reverse, but it is practical to categorize a given patient by his or her most frequent presentation. The definition of permanent AF is often arbitrary, and the duration refers both to individual episodes and to how long the diagnosis has been present in a given patient. Thus, in a patient with paroxysmal AF, episodes lasting seconds to hours may occur repeatedly for years.

This terminology applies to episodes lasting longer than 30 s without a reversible cause. Secondary AF in the setting of acute MI, cardiac surgery, pericarditis, myocarditis, hyperthyroidism, or acute pulmonary disease is considered separately. In these situations, AF is not the primary problem, and concurrent treatment of the underlying disorder usually terminates the arrhythmia. Conversely, when AF occurs in the course of a concurrent disorder like well-controlled hypothyroidism, the general principles for management of the arrhythmia apply.

The term lone AF applies to individuals younger than 60 y without clinical or echocardiographic evidence of cardiopulmonary disease, including hypertension.5 These patients have a favorable prognosis with respect to thromboembolism and mortality. Over time, patients move out of the lone AF category due to aging or development of cardiac abnormalities such as enlargement of the LA, and the risks of thromboembolism and mortality rise. The term nonvalvular AF refers to cases without rheumatic mitral valve disease, prosthetic heart valve, or valve repair.

IV. Epidemiology and prognosis

AF is the most common arrhythmia in clinical practice, accounting for approximately one third of hospitalizations for cardiac rhythm disturbances. An estimated 2.3 million people in North America and 4.5 million people in the European Union have paroxysmal or persistent AF.9 During the past 20 y, hospital admissions for AF have increased by 66%7 due to the aging of the population, a rising prevalence of chronic heart disease, more frequent diagnosis through use of ambulatory monitoring devices, and other factors. AF is an extremely expensive public health problem (approximately €3000 [approximately U.S. $3600] annually per patient)8; the total cost burden approaches €13.5 billion (approximately U.S. $15.7 billion) in the European Union.

A. Prevalence

The estimated prevalence of AF is 0.4% to 1% in the general population,9 increasing with age to 8% in those older than 80 y.10 Among men, the age-adjusted prevalence has more than doubled over a generation,10 while the prevalence in women has remained constant.11 The median age of patients with AF is about 75 y. The number of men and women with AF is about equal, but approximately 60% of those over 75 y old are female. Based on limited data, the age-adjusted risk of developing AF in blacks seems less than half that in whites.

In population-based studies, patients with no history of cardiopulmonary disease account for fewer than 12% of all cases of AF.10 In case series, however, the observed proportion of lone AF was sometimes greater than 30%.12

B. Incidence

In prospective studies, the incidence of AF increases from less than 0.1% per year in people younger than 40 y to over 1.5% per year among women and 2% among men older than 80 y.13 In patients treated for HF, the 3-y incidence of AF was almost 10%.14 Angiotensin inhibition may be associated with a reduced incidence of AF in patients with HF15 and hypertension.16

C. Prognosis

AF is associated with an increased long-term risk of stroke,17 HF, and all-cause mortality, especially among women.18 The mortality rate of patients with AF is about double that of patients in normal sinus rhythm and is linked to the severity of underlying heart disease.19 In the Etude en Activité Libérale sur la Fibrillation Auriculaire Study (ALFA), about two thirds of the 5% annualized mortality was attributed to cardiovascular causes.12 In large HF trials (COMET [Carvedilol Or Metoprolol European Trial], Val-HeFT [Valsartan Heart Failure Trial]), AF was a strong independent risk factor for mortality and morbidity.20,21 HF promotes AF, AF aggravates HF, and individuals with either condition who develop the alternate condition share a poor prognosis.22 Thus, managing patients with the associated conditions is a major challenge, and randomized trials are needed to investigate the impact of AF on prognosis in HF.

The rate of ischemic stroke among patients with nonvalvular AF averages 5% per year, 2 to 7 times that of people without AF.23 One of every 6 strokes occurs in a patient with AF, and when TIAs and clinically ‘silent’ strokes detected by brain imaging are considered, the rate of brain ischemia accompanying nonvalvular AF exceeds 7% per year.24 In patients with rheumatic heart disease and AF in the Framingham Heart Study, stroke risk was increased 17-fold compared with age-matched controls,25 and attributable risk was 5 times greater than in those with nonrheumatic AF.23 The risk of stroke increased with age; the annual risk of stroke attributable to AF was 1.5% in participants aged 50 to 59 y and 23.5% in those aged 80 to 89 y.23

V. Pathophysiological mechanisms

A. Atrial factors

1. Atrial pathology as a cause of atrial fibrillation

The most frequent histopathological changes in AF are atrial fibrosis and loss of atrial muscle mass, but it is difficult to distinguish changes due to AF from those due to associated heart disease. Atrial fibrosis may precede the onset of AF,26 and juxtaposition of patchy fibrosis with normal atrial fibers may account for nonhomogeneity of conduction.27 Interstitial fibrosis may result from apoptosis leading to replacement of atrial myocytes,28 loss of myofibrils, accumulation of glycogen granules, disruption of cell coupling at gap junctions,29 and organelle aggregates30 and may be triggered by atrial dilation in any type of heart disease associated with AF.

Patients with valvular heart disease who have mild fibrosis respond more successfully to cardioversion than those with severe fibrosis, and fibrosis is thought to contribute to persistent AF.31 The concentration of membrane-bound glycoproteins that regulate cell–cell and cell–matrix interactions (disintegrin and metalloproteinases) in human atrial myocardium has been reported to double during AF, and these changes may contribute to atrial dilation in patients with longstanding AF. Dilation of the atria activates several molecular pathways, including the renin-angiotensin-aldosterone system (RAAS). Angiotensin II is upregulated in response to stretch,32 and atrial tissue from patients with persistent AF demonstrates increased expression of angiotensin-converting enzyme (ACE).33 Angiotensin inhibition may prevent AF by reducing fibrosis.34 Atrial dilation and interstitial fibrosis in HF facilitate sustained AF.35 The regional electrical silence (suggesting scar), voltage reduction, and conduction slowing described in patients with HF are similar to changes in the atria that occur as a consequence of aging.36

2. Mechanisms of atrial fibrillation

Available data support a ‘focal’ triggering mechanism involving automaticity or multiple reentrant wavelets, but these mechanisms are not mutually exclusive and may coexist (Figure 2).

Figure 2

Posterior view of principal electrophysiological mechanisms of atrial fibrillation. (A) Focal activation. The initiating focus (indicated by the star) often lies within the region of the pulmonary veins. The resulting wavelets represent fibrillatory conduction, as in multiple-wavelet reentry. (B) Multiple-wavelet reentry. Wavelets (indicated by arrows) randomly re-enter tissue previously activated by the same or another wavelet. The routes the wavelets travel vary. Reproduced with permission from Konings KT, Kirchhof CJ, Smeets JR, et al. High-density mapping of electrically induced atrial fibrillation in humans. Circulation 1994;89:1665–1680.45 LA indicates left atrium; PV, pulmonary vein; ICV, inferior vena cava; SCV, superior vena cava; and RA, right atrium.

The important observation that a focal source for AF could be identified and ablation of this source could extinguish AF37 supported a focal origin. While pulmonary veins (PVs) are the most frequent source of these rapidly atrial impulses, foci have also been found in the superior vena cava, ligament of Marshall, left posterior free wall, crista terminalis, and coronary sinus.3740 In histological studies, cardiac muscle with preserved electrical properties extends into the PVs,41 and the primacy of PVs as triggers of AF has prompted substantial research into the anatomical and electrophysiological properties of these structures. Atrial tissue in the PVs of patients with AF has shorter refractory periods than in control patients or other parts of the atria in patients with AF.42,43 This heterogeneity of conduction may promote reentry and form a substrate for sustained AF.44

The multiple-wavelet hypothesis as the mechanism of reentrant AF46 involves fractionation of wave fronts propagating through the atria and self-perpetuating ‘daughter wavelets.’ In this model, the number of wavelets at any time depends on the refractory period, mass, and conduction velocity in different parts of the atria. A large atrial mass with a short refractory period and delayed conduction increases the number of wavelets, favoring sustained AF. Simultaneous recordings from multiple electrodes supported the multiple-wavelet hypothesis in human subjects.47 Although the multiple-wavelet hypothesis was for years the dominant theory explaining the mechanism of AF, data from experimental47a and clinical47b,47c mapping studies challenge this notion. In patients with idiopathic paroxysmal AF, widespread distribution of abnormal electrograms in the RA predicts development of persistent AF,48 suggesting the importance of an abnormal substrate in the maintenance of AF. Furthermore, in patients with persistent AF undergoing conversion to sinus rhythm, intra-atrial conduction is prolonged compared with a control group, especially among those who develop recurrent AF.49 Among patients with HF, prolongation of the P wave on signal-averaged ECG analysis was more frequent in those prone to paroxysmal AF.50 Because many of these observations were made prior to onset of clinical AF, the findings cannot be ascribed to atrial remodeling that occurs as a consequence of AF, and the degree to which changes in the atrial architecture contribute to the initiation and maintenance of AF is not known.

3. Atrial electrical remodeling

Pharmacological or direct-current cardioversion of AF has a higher success rate when AF has been present for less than 24 h,51 whereas more prolonged AF makes restoring and maintaining sinus rhythm less likely. These observations gave rise to the adage ‘atrial fibrillation begets atrial fibrillation.’ The notion that AF is self-perpetuating takes experimental support from a goat model using an automatic atrial fibrillator that detected spontaneous termination of AF and reinduced the arrhythmia by electrical stimulation.52 Initially, electrically induced AF terminated spontaneously. After repeated inductions, however, the episodes became progressively more sustained until AF persisted at a more rapid atrial rate.52 The increasing propensity to AF was related to progressive shortening of effective refractory periods with increasing episode duration, a phenomenon known as electrophysiological remodeling.

In addition to remodeling and changes in electrical refractoriness, prolonged AF disturbs atrial contractile function. After a period of persistent AF, recovery of atrial contraction can be delayed for days or weeks following the restoration of sinus rhythm, and this has important implications for the duration of anticoagulation after cardioversion. (See Section VIII.B.2, Preventing thromboembolism.)

4. Other factors contributing to atrial fibrillation

Data are accumulating on the importance of the RAAS in the genesis of AF.53 Irbesartan plus amiodarone was associated with a lower incidence of recurrent AF after cardioversion than amiodarone alone,15 and treatment with angiotensin inhibitors and diuretics reduced the incidence of AF after catheter ablation of atrial flutter.54 Inhibition of the RAAS, alone or in combination with other therapies, may prevent the onset or maintenance of AF through several mechanisms,55 including lower atrial pressure and wall stress, prevention of structural remodeling (fibrosis, dilation, and hypertrophy) in both the LA and left ventricle (LV), inhibition of neurohumoral activation, reducing blood pressure, prevention or amelioration of HF, and avoidance of hypokalemia. Treatment with trandolapril reduced the incidence of AF in patients with LV dysfunction following acute MI,56 but it remains to be clarified whether this effect is related to reversal of structural or electrical remodeling in the atria or to another mechanism.

Other factors potentially involved in the induction or maintenance of AF are outlined in Table 2. Among these is inflammation, and ongoing studies are exploring the use of statin-type lipid-lowering drugs with this mechanism in mind.

View this table:
Table 2

Etiologies and factors predisposing patients to AF

B. Atrioventricular conduction

1. General aspects

In the absence of an accessory pathway or His-Purkinje dysfunction, the AV node limits conduction during AF.57 Of the multiple atrial inputs to the AV node that have been identified, 2 seem dominant: one directed posteriorly via the crista terminalis and the other aimed anteriorly via the interatrial septum. Other factors affecting AV conduction are the intrinsic refractoriness of the AV node, concealed conduction, and autonomic tone. Concealed conduction plays a prominent role in determining the ventricular response during AF58 by altering the refractoriness of the AV node and slowing or blocking atrial impulses and may explain the irregularity of ventricular response during AF.59 When the atrial rate is relatively slow during AF, the ventricular rate tends to rise and, conversely, higher atrial rate is associated with slower ventricular rate.

Increased parasympathetic and reduced sympathetic tone exert negative dromotropic effects on AV nodal conduction, while the opposite is true in states of decreased parasympathetic and increased sympathetic tone.58 Vagal tone also enhances the negative chronotropic effects of concealed conduction in the AV node.60 Fluctuations in autonomic tone can produce disparate ventricular responses to AF, exemplified by a slow ventricular rate during sleep but accelerated ventricular response during exercise. Digitalis, which slows the ventricular rate during AF predominantly by increasing vagal tone, is more effective for controlling heart rate at rest in AF but less effective during activity.

2. Atrioventricular conduction in preexcitation syndromes

Conduction across an accessory pathway during AF can result in dangerously rapid ventricular rates.2 Transition of AV reentry into AF in patients with the WPW syndrome can produce a rapid ventricular response that degenerates into lethal ventricular fibrillation.61 Drugs that lengthen refractoriness and slow conduction across the AV node (such as digitalis, verapamil, or diltiazem) do not block conduction over the accessory pathway and may accelerate the ventricular rate. Hence, these agents are contraindicated in this situation.62 Although the potential for beta blockers to potentiate conduction across the accessory pathway is controversial, caution should be exercised in the use of these agents as well in patients with AF associated with preexcitation.

C. Myocardial and hemodynamic consequences of atrial fibrillation

Among factors that affect hemodynamic function during AF are loss of synchronous atrial mechanical activity, irregular ventricular response, rapid heart rate, and impaired coronary arterial blood flow. Loss of atrial contraction may markedly decrease cardiac output, especially when diastolic ventricular filling is impaired by mitral stenosis, hypertension, HCM, or restrictive cardiomyopathy. Myocardial contractility is not constant during AF because of force–interval relationships associated with variations in cycle length.63 In patients with persistent AF, mean LA and RA volumes increase over time64 and restoration and maintenance of sinus rhythm decrease these volumes.65 Moreover, TEE has demonstrated that contractile function and blood flow velocity in the LAA recover after cardioversion, consistent with a reversible atrial cardiomyopathy in patients with AF.66 Although one might expect restoration of sinus rhythm to improve the other hemodynamic characteristics associated with AF, this is not always the case.67

Beyond its effects on atrial function, a persistently elevated ventricular rate during AF may adversely increase mitral regurgitation and produce dilated ventricular cardiomyopathy (tachycardia-induced cardiomyopathy).2,68 It is important to recognize this cause of cardiomyopathy, in which HF is a consequence rather than the cause of AF, because control of the ventricular rate may lead to reversal of the myopathic process. A variety of hypotheses have been proposed to explain tachycardia-mediated cardiomyopathy on the basis of myocardial energy depletion, ischemia, abnormal calcium regulation, and remodeling, but the actual mechanisms are still unclear.69

D. Thromboembolism

Although ischemic stroke and systemic arterial occlusion in AF are generally attributed to embolism of thrombus from the LA, the pathogenesis of thromboembolism is complex.70 Up to 25% of strokes in patients with AF may be due to intrinsic cerebrovascular diseases, other cardiac sources of embolism, or atheromatous pathology in the proximal aorta.71,72 The annual risk of stroke in patients with AF is in the range of 3% to 8% per year, depending on associated stroke risk factors.23 About half of all elderly AF patients have hypertension (a major risk factor for cerebrovascular disease), and approximately 12% have carotid artery stenosis.73 Carotid atherosclerosis is not substantially more prevalent in AF patients with stroke than in patients without AF, however, and is probably a relatively minor contributing epidemiological factor.74

1. Pathophysiology of thrombus formation

Thrombus formation as a result of stasis in the LAA is thought to represent the main source of disabling cardioembolic ischemic strokes in patients with AF. These thrombi cannot be regularly examined by precordial (transthoracic) echocardiography,75 and TEE is a more sensitive and specific method to assess LAA function76 and detect thrombus formation. Serial TEE studies of the LA77 and LAA78 during conversion of AF to sinus rhythm demonstrated reduced LAA flow velocities related to loss of organized mechanical contraction during AF. Thrombi are more often encountered in AF patients with ischemic stroke than in those without stroke.79 Although clinical management is based on the presumption that thrombus formation requires continuation of AF for approximately 48 h, thrombi have been identified by TEE within shorter intervals.80,81

After successful cardioversion, regardless of whether the method is electrical, pharmacological, or spontaneous,82 stunning of the LAA may account for an increased risk of thromboembolic events. Atrial stunning is at a maximum immediately after cardioversion; progressive improvement of atrial transport function usually occurs within a few days but sometimes takes as long as 3 to 4 wk, depending on the duration of AF.82,83 This corroborates the clinical observation that following cardioversion, more than 80% of thromboembolic events occur during the first 3 d and almost all occur within 10 d.84 TEE studies have verified resolution of thrombus in the majority of patients.85 Similar observations have defined the dynamic nature of LA/LAA dysfunction following conversion of AF, providing a mechanistic rationale for anticoagulation for several weeks before and after successful cardioversion. Although stunning may be milder with certain associated conditions or a short duration of AF, anticoagulation is recommended during cardioversion and for at least 4 wk afterward in all patients with AF lasting longer than 48 h or of unknown duration, including lone AF, except when contraindicated.

Decreased flow within the LA/LAA during AF has been associated with spontaneous echo contrast (SEC), thrombus formation, and embolic events.86,87 Specifically, SEC, or ‘smoke,’ a swirling haze of variable density, may be detected by transthoracic echocardiography or TEE imaging under low-flow conditions.88 There is evidence that SEC is a marker of stasis caused by AF,89,90 but the utility of SEC for prospective thromboembolic risk stratification beyond that achieved by clinical assessment alone has not been confirmed.

LAA flow velocities are lower in patients with atrial flutter than is usually seen during sinus rhythm but higher than in AF. Whether this accounts for any lower prevalence of LAA thrombus or thromboembolism associated with atrial flutter is uncertain. As in AF, atrial flutter is associated with low appendage emptying velocities following cardioversion with the potential for thromboembolism91 and anticoagulation is recommended similarly. (See Section in the full-text guidelines, Therapeutic implications.)

2. Clinical implications

Complex thromboembolic mechanisms are operative in AF and involve the interplay of risk factors related to atrial stasis, endothelial dysfunction, and systemic and possibly local hypercoagulability. The strong association between hypertension and stroke in AF is probably mediated primarily by embolism originating in the LAA,72 but hypertension also increases the risk of noncardioembolic strokes in patients with AF.92 Whether control of hypertension lowers the risk for cardioembolic stroke in patients with AF is a vital question.

The increasing stroke risk in patients with AF with advancing age is also multifactorial. Aging is a risk factor for atherosclerosis, and plaques in the aortic arch are associated with stroke independent of AF.93 Age is a more potent risk factor when combined with other risk factors such as hypertension or female gender, and women over age 75 y with AF are at particular risk for stroke.94

LV systolic dysfunction, as indicated by a history of HF or echocardiographic assessment, predicts ischemic stroke in patients with AF who receive no antithrombotic therapy95 but not in moderate-risk patients given aspirin.96,97 LV systolic dysfunction has been associated both with LA thrombus and with noncardioembolic strokes in patients with AF.72,98

VI. Causes, associated conditions, clinical manifestations, and quality of life

A. Causes and associated conditions

1. Reversible causes of atrial fibrillation

AF may be related to acute temporary causes, including alcohol intake (‘holiday heart syndrome’), surgery, electrocution, MI, pericarditis, myocarditis, pulmonary embolism or other pulmonary diseases, hyperthyroidism, and other metabolic disorders. In such cases, successful treatment of the underlying condition often eliminates AF. In the setting of acute MI, the development of AF portends an adverse prognosis compared with preinfarct AF or sinus rhythm.99,100 When AF is associated with atrial flutter, WPW syndrome, or AV nodal reentrant tachycardia, treatment of the primary arrhythmia reduces or eliminates the incidence of recurrent AF.101 AF is a common early postoperative complication of cardiac and thoracic surgery.

2. Atrial fibrillation without associated heart disease

Approximately 30% to 45% of cases of paroxysmal AF and 20% to 25% of cases of persistent AF occur in young patients without demonstrable underlying disease (‘lone AF’).12 AF can present as an isolated or familial arrhythmia, although a causal underlying disease may appear over time.102 Although AF may occur in the elderly without underlying heart disease, the changes in cardiac structure and function that accompany aging, such as increased myocardial stiffness, may be associated with AF, just as heart disease in older patients may be coincidental and unrelated to AF.

3. Medical conditions associated with atrial fibrillation

Obesity is an important risk factor for the development of AF.103 After adjustment for clinical risk factors, the excess risk of AF appears related to LA dilation. There is a graded increase in LA size as body mass index increases from normal to the overweight and obese categories, and weight has been linked to regression of LA enlargement.104 These findings suggest a physiological link between obesity, AF, and stroke and raise the intriguing possibility that weight reduction may decrease the risk associated with AF.

4. Atrial fibrillation with associated heart disease

Specific cardiovascular conditions associated with AF include valvular heart disease (most often mitral valve disease), HF, coronary artery disease (CAD), and hypertension, particularly when LV hypertrophy (LVH) is present. In addition, AF may be associated with HCM, dilated cardiomyopathy, or congenital heart disease, especially atrial septal defect in adults. Potential etiologies also include restrictive cardiomyopathies (e.g., amyloidosis, hemochromatosis, and endomyocardial fibrosis), cardiac tumors, and constrictive pericarditis. Other heart diseases, such as mitral valve prolapse with or without mitral regurgitation, calcification of the mitral annulus, cor pulmonale, and idiopathic dilation of the RA, have been associated with a high incidence of AF. AF is commonly encountered in patients with sleep apnea syndrome, but whether the arrhythmia is provoked by hypoxia, another biochemical abnormality, changes in pulmonary dynamics or RA factors, changes in autonomic tone, or systemic hypertension has not been determined.

5. Familial atrial fibrillation

Familial AF, defined as lone AF running in a family, is more common than previously recognized but should be distinguished from AF secondary to other genetic diseases like familial cardiomyopathies. The likelihood of developing AF is increased among the offspring of parents with AF, suggesting a familial susceptibility to the arrhythmia, but the mechanisms associated with transmission are not necessarily electrical, because the relationship has also been seen in patients with a family history of hypertension, diabetes, or HF.105 The molecular defects responsible for familial AF are largely unknown. Specific chromosomal loci linked to AF in some families106 suggest distinct genetic mutations.107

6. Autonomic influences in atrial fibrillation

Autonomic influences play an important role in the initiation of AF. Measurement of heart rate variability (HRV) reflects changes in relative autonomic modulation rather than the absolute level of sympathetic or parasympathetic tone. It appears, however, that the balance between sympathetic and vagal influences is important as a predictor of AF. Vagal predominance has been observed in the minutes preceding the onset of AF in some patients with structurally normal hearts, while in others there is a shift toward sympathetic predominance.108,109 Although certain patients can be characterized in terms of a vagal or an adrenergic form of AF, these cases likely represent the extremes of either influence.110 In general, vagally mediated AF occurs at night or after meals, while adrenergically induced AF typically occurs during the daytime.111 In patients with vagally mediated AF, the more common form, adrenergic blocking drugs or digitalis sometimes worsen symptoms. For AF of the adrenergic type, beta blockers are the initial treatment of choice.

B. Clinical manifestations

AF may cause a sensation of palpitations, have distinct hemodynamic or thromboembolic consequences, or follow an asymptomatic period of unknown duration. Ambulatory ECG recordings and device-based monitoring reveal that individuals may experience periods of both symptomatic and asymptomatic AF.112114 Over time, palpitation may disappear, such that patients in whom the arrhythmia has become permanent may become asymptomatic. This is particularly common among the elderly. Some patients experience symptoms only during paroxysmal AF, or only intermittently during sustained AF. When present, symptoms of AF vary with the irregularity and rate of ventricular response, underlying functional status, duration of AF, and individual patient factors.115

The initial presentation of AF may be an embolic complication or exacerbation of HF, but most patients complain of palpitations, chest pain, dyspnea, fatigue, lightheadedness, or syncope. Polyuria may be associated with the release of atrial natriuretic peptide, particularly as episodes of AF begin or terminate. AF associated with a sustained, rapid ventricular response can lead to tachycardia-mediated cardiomyopathy, especially in patients unaware of the arrhythmia. Syncope is an uncommon complication that can occur upon conversion in patients with sinus node dysfunction or because of rapid ventricular rates in patients with HCM, valvular aortic stenosis, or an accessory pathway.

C. Quality of life

Available data suggest that quality of life is considerably impaired in patients with AF compared to age-matched controls. Sustained sinus rhythm is associated with improved quality of life and better exercise performance than AF in some studies but not others.116 In a typical study, a majority of patients with paroxysmal AF considered the arrhythmia disruptive of lifestyle, but this perception was not associated with either the frequency or duration of symptomatic episodes.117

VII. Clinical evaluation

A. Basic evaluation of the patient with atrial fibrillation

1. Clinical history and physical examination

The diagnosis of AF requires confirmation by ECG recording, sometimes in the form of bedside telemetry or ambulatory Holter recordings. The initial evaluation of a patient with suspected or proven AF involves characterizing the pattern of the arrhythmia as paroxysmal or persistent, determining its cause, and defining associated cardiac and extracardiac factors pertinent to the etiology, tolerability, and management. The work-up and therapy can usually be accomplished in a single outpatient encounter (Table 3), unless the rhythm has not been specifically documented and additional monitoring is necessary.

View this table:
Table 3

Clinical evaluation in patients with AF

Physical examination may suggest AF on the basis of irregular pulse, irregular jugular venous pulsations, variation in the intensity of the first heart sound, or absence of a fourth sound heard previously during sinus rhythm. The findings are similar in patients with atrial flutter, except that the rhythm may be regular and rapid venous oscillations may occasionally be visible in the jugular pulse.

2. Investigations

The diagnosis of AF requires ECG documentation by at least a single-lead recording during the arrhythmia. In patients with implanted pacemakers or defibrillators, the diagnostic and memory functions may allow accurate and automatic detection.118 A chest radiograph is valuable mostly to detect intrinsic pulmonary pathology and evaluate the pulmonary vasculature. It is important that thyroid, renal, and hepatic functions, serum electrolytes, and the hemogram be measured at least once in the course of evaluation.119 All patients with AF should also have 2-dimensional, Doppler echocardiography to assess LA and LV dimensions, LV wall thickness, and function and to exclude occult valvular or pericardial disease and HCM. Thrombus in the LA or LAA is seldom detected without TEE. Among the TEE features associated with thromboembolism in patients with nonvalvular AF are thrombus, SEC, reduced LAA flow velocity, and aortic atheromatous abnormalities,120 but prospective investigations are needed to compare these TEE findings with clinical and transthoracic echocardiographic predictors of thromboembolism. Detection of LA/LAA thrombus in the setting of stroke or systemic embolism is convincing evidence of a cardiogenic mechanism.81

VIII. Management

A. Strategic objectives

Management of patients with AF involves 3 objectives—rate control, prevention of thromboembolism, and correction of the rhythm disturbance—and these are not mutually exclusive. The initial management decision involves primarily a rate control or rhythm control strategy. Under the rate control strategy, the ventricular rate is controlled with no commitment to restore or maintain sinus rhythm. The rhythm control strategy attempts restoration and/or maintenance of sinus rhythm. The latter strategy also requires attention to rate control. Depending on the patient's course, the strategy initially chosen may prove unsuccessful and the alternate strategy is then adopted. Regardless of whether the rate control or rhythm control strategy is pursued, attention must also be directed to antithrombotic therapy for prevention of thromboembolism.

B. Pharmacological and nonpharmacological treatment options

Drugs and ablation are effective for both rate and rhythm control, and in special circumstances surgery may be the preferred option. Regardless of the approach, the need for anticoagulation is based on stroke risk and not on whether sinus rhythm is maintained. For rhythm control, drugs are typically the first choice and LA ablation is a second-line choice, especially in patients with symptomatic lone AF. In some patients, especially young persons with very symptomatic AF who need sinus rhythm, radiofrequency ablation may be preferred over years of drug therapy. Patients with preoperative AF undergoing cardiac surgery face a unique opportunity. While few are candidates for a stand-alone surgical procedure to cure AF using the maze or LA ablation techniques, these approaches can be an effective adjunct to coronary bypass or valve repair surgery to prevent recurrent postoperative AF. Because the LAA is the site of greater than 95% of detected thrombi, this structure is commonly removed from the circulation during cardiac surgery in patients at risk of developing postoperative AF, although this has not been proved to prevent stroke.121

1. Heart rate control versus rhythm control

For patients with symptomatic AF lasting many weeks, initial therapy may be anticoagulation and rate control while the long-term goal is to restore sinus rhythm. When cardioversion is contemplated and the duration of AF is unknown or exceeds 48 h, patients who do not require long-term anticoagulation may benefit from short-term anticoagulation. If rate control offers inadequate symptomatic relief, restoration of sinus rhythm becomes a clear long-term goal. Early cardioversion may be necessary if AF causes hypotension or worsening HF. In contrast, amelioration of symptoms by rate control in older patients may steer the clinician away from attempts to restore sinus rhythm. In some circumstances, when the initiating pathophysiology of AF is reversible, as for instance in the setting of thyrotoxicosis or after cardiac surgery, no long-term therapy may be necessary.

Randomized trials comparing outcomes of rhythm versus rate control strategies in patients with AF are summarized in Tables 46. Among these, AFFIRM (Atrial Fibrillation Follow-up Investigation of Rhythm Management) found no difference in mortality or stroke rate between patients assigned to one strategy or the other. The RACE (Rate Control vs. Electrical cardioversion for persistent atrial fibrillation) trial found rate control not inferior to rhythm control for prevention of death and morbidity. Clinically silent recurrences of AF in asymptomatic patients treated with antiarrhythmic drugs may be responsible for thromboembolic events after withdrawal of anticoagulation. Hence, patients at high risk for stroke may require anticoagulation regardless of whether the rate control or rhythm control strategy is chosen, but the AFFIRM trial was not designed to address this question.122 While secondary analyses support this notion, the stroke rate in patients assigned to rhythm control who stopped warfarin is uncertain, and additional research is needed to address this important question.

View this table:
Table 4

Trials comparing rate control and rhythm control strategies in patients with AF

View this table:
Table 5

General characteristics of rhythm control and rate control trials in patients with AF

View this table:
Table 6

Comparison of adverse outcomes in rhythm control and rate control trials in patients with AF

Information about the effects of antiarrhythmic and chronotropic therapies on quality of life is inconsistent.116,130,131 Neither the AFFIRM,132 RACE,124 PIAF (Pharmacologic Intervention in Atrial Fibrillation),125 nor STAF (Strategies of Treatment of Atrial Fibrillation)126 studies found differences in quality of life with rhythm control compared with rate control. Rhythm control in the PIAF and HOT CAFÉ (How to Treat Chronic Atrial Fibrillation)127 studies resulted in better exercise tolerance than rate control, but this did not translate into improved quality of life. Symptomatic improvement has been reported after the maze procedure in patients with AF.133 Clinicians must exercise judgment, however, in translating shifts in quality of life in these selected populations to the sense of well-being experienced by individual patients. Patients with similar health status may experience an entirely different quality of life, and treatment must be tailored to each individual, depending on the nature, intensity, and frequency of symptoms, patient preferences, comorbid conditions, and the ongoing response to treatment.

Depending on symptoms, rate control may be reasonable initial therapy in older patients with persistent AF who have hypertension or heart disease. For younger individuals, especially those with paroxysmal lone AF, rhythm control may be a better initial approach. Often, medications that exert both antiarrhythmic and rate-controlling effects are required. Catheter ablation should be considered to maintain sinus rhythm in selected patients who failed to respond to antiarrhythmic drug therapy.134

In patients with AF, the ventricular rate may accelerate excessively during exercise even when it is well-controlled at rest. In addition to allowing adequate time for ventricular filling and avoiding rate-related ischemia, enhancement of intraventricular conduction with rate reduction may result in improved hemodynamics. It may be useful to evaluate the heart rate response to submaximal or maximal exercise or to monitor the rate over an extended period (e.g., by use of 24-h Holter recording). The definition of adequate rate control has been based primarily on short-term hemodynamic benefits and not well studied with respect to regularity or irregularity of the ventricular response to AF, quality of life, symptoms, or development of cardiomyopathy. No standard method for assessment of heart rate control has been established to guide the management of patients with AF. Criteria for rate control vary with patient age but usually involve achieving ventricular rates between 60 and 80 beats per minute at rest and between 90 and 115 beats per minute during moderate exercise.

Patients who are symptomatic with rapid ventricular rates during AF require prompt medical management, and cardioversion should be considered if symptomatic hypotension, angina, or HF is present. A sustained, uncontrolled tachycardia may lead to deterioration of ventricular function (tachycardia-related cardiomyopathy) that improves with adequate rate control. Tachycardia-induced cardiomyopathy tends to resolve within 6 mo of rate or rhythm control; when tachycardia recurs, LV ejection fraction declines and HF develops over a shorter period, and this is associated with a relatively poor prognosis.137

a. Pharmacological rate control during atrial fibrillation

The functional refractory period of the AV node correlates inversely with ventricular rate during AF, and drugs that prolong the refractory period are generally effective for rate control. There is no evidence that pharmacological rate control has any adverse influence on LV function, but bradycardia and heart block, as an unwanted effect of beta blockers, amiodarone, digitalis glycosides, or nondihydropyridine calcium channel antagonists, particularly in patients with paroxysmal AF, especially the elderly. When rapid control of the ventricular response to AF is required or oral administration of medication is not feasible, medication may be administered intravenously. Otherwise, in hemodynamically stable patients with a rapid ventricular response to AF, negative chronotropic medication may be administered orally (Table 7). Combinations may be necessary to achieve rate control in both acute and chronic situations. Some patients develop symptomatic bradycardia that requires permanent pacing. Nonpharmacological therapy should be considered when pharmacological measures fail.

View this table:
Table 7

Intravenous and orally administered pharmacological agents for heart rate control in patients with atrial fibrillation

Special considerations in patients with the Wolff-Parkinson-White syndrome

Intravenous administration of beta blockers, digitalis, adenosine, lidocaine, and nondihydropyridine calcium channel antagonists, all of which slow conduction across the AV node, is contraindicated in patients with the WPW syndrome and tachycardia associated with ventricular preexcitation because they can facilitate antegrade conduction along the accessory pathway during AF,2 resulting in acceleration of the ventricular rate, hypotension, or ventricular fibrillation.62 When the arrhythmia is associated with hemodynamic compromise, however, early direct-current cardioversion is indicated. In hemodynamically stable patients with preexcitation, type I antiarrhythmic agents or amiodarone may be administered intravenously. Beta blockers and calcium channel blockers are reasonable for oral maintenance therapy.138

Pharmacological therapy to control heart rate in patients with both atrial fibrillation and atrial flutter

A patient treated with AV nodal blocking drugs whose ventricular rate is well controlled during AF may experience a rise or fall in rate if he or she develops atrial flutter. This is also true when antiarrhythmic agents such as propafenone or flecainide are used to prevent recurrent AF. These compounds may increase the likelihood of 1:1 AV conduction during atrial flutter leading to a very rapid ventricular response. Thus, when these agents are given for prophylaxis against recurrent paroxysmal AF or atrial flutter, AV nodal blocking drugs should be routinely coadministered. An exception may be patients with paroxysmal AF who have undergone catheter ablation of the cavotricuspid isthmus to prevent atrial flutter.

b. Regulation of atrioventricular nodal conduction by pacing

Because ventricular pacing prolongs the AV nodal refractory period as a result of concealed retrograde penetration, it eliminates longer ventricular cycles and may reduce the number of short ventricular cycles related to rapid AV conduction during AF. Pacing at approximately the mean ventricular rate during spontaneous AV conduction can regulate the ventricular rhythm during AF.139 This may be useful for patients with marked variability in ventricular rates or for those who develop resting bradycardia during treatment with medication. In some patients, the hemodynamic benefit of revascularization may be offset by asynchronous ventricular activation during right ventricular pacing.

c. Atrioventricular nodal ablation

AV nodal ablation in conjunction with permanent pacemaker implantation provides highly effective control of the heart rate and improves symptoms in selected patients with AF.140143 In general, patients most likely to benefit from this strategy are those with symptoms or tachycardia-mediated cardiomyopathy related to rapid ventricular rate during AF that cannot be controlled adequately with antiarrhythmic or negative chronotropic medications. Meta-analyses of 21 studies published between 1989 and 1998 that included a total of 1181 patients concluded that AV nodal ablation and permanent pacemaker implantation significantly improved cardiac symptoms, quality of life, and healthcare utilization for patients with symptomatic AF refractory to medical treatment.143 Catheter ablation of inferior atrial inputs to the AV node slows the ventricular rate during AF and improves symptoms without pacemaker implantation.144,145 This technique has several limitations, however, including inadvertent complete AV block and a tendency of ventricular rate to rise over the 6 mo following ablation. Thus, AV nodal modification without pacemaker implantation is only rarely used.

Although the symptomatic benefits of AV nodal ablation are clear, limitations include the persistent need for anticoagulation, loss of AV synchrony, and lifelong pacemaker dependency. There is also a finite risk of sudden death due to torsades de pointes or ventricular fibrillation.146 Patients with abnormalities of diastolic ventricular compliance who depend on AV synchrony to maintain cardiac output, such as those with HCM or hypertensive heart disease, may experience persistent symptoms after AV nodal ablation and pacemaker implantation. Hence, patients should be counseled regarding each of these considerations before proceeding with this irreversible measure.

Patients with normal LV function or reversible LV dysfunction undergoing AV nodal ablation are most likely to benefit from standard AV nodal ablation and pacemaker implantation. For those with impaired LV function not due to tachycardia, a biventricular pacemaker with or without defibrillator capability should be considered. Upgrading to a biventricular device should be considered for patients with HF and a right ventricular pacing system who have undergone AV node ablation.147

2. Preventing thromboembolism

a. Risk stratification
Epidemiological data

In a small, retrospective, population-based study in Olmsted County, Minnesota, over 3 decades, the 15-y cumulative stroke rate in people with lone AF (defined as those younger than 60 y with no clinical history or echocardiographic signs of cardiopulmonary disease) was 1.3%.5 In the SPAF (Stroke Prevention in Atrial Fibrillation III) studies, the annualized rate of ischemic stroke during aspirin treatment was similar in those with paroxysmal (3.2%) and permanent (3.3%) AF.148 Those with prior stroke or TIA have a rate of subsequent stroke of 10% to 12% per year when treated with aspirin, and these patients benefit substantially from adjusted-dose oral anticoagulation.149 In addition to prior thromboembolism, HF, hypertension, increasing age, and diabetes mellitus have consistently emerged as independent risk factors for ischemic stroke associated with nonvalvular AF.96 Other factors, such as female gender, systolic blood pressure over 160 mm Hg, and LV dysfunction, have been variably linked to stroke.96 The relative risk for ischemic stroke associated with specific clinical features, derived from a collaborative analysis of participants given no antithrombotic therapy in the control groups of 5 randomized trials, is displayed in Table 8. In patients with nonvalvular AF, prior stroke or TIA is the strongest independent predictor of stroke, significantly associated with stroke in all 6 studies in which it was evaluated, with incremental relative risk between 1.9 and 3.7 (averaging approximately 3.0). All patients with prior stroke or TIA require anticoagulation unless contraindications exist in a given patient. Patient age is a consistent independent predictor of stroke (Figure 3), but older people are also at increased risk for anticoagulant-related bleeding.150 Special consideration of these older patients is therefore a critical aspect of effective stroke prophylaxis.151

Figure 3

Stroke rates in relation to age among patients in untreated control groups of randomized trials of antithrombotic therapy. Data are from the Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomized controlled trials. Arch Intern Med 1994;154:1449–1457.17

View this table:
Table 8

Risk factors for ischemic stroke and systemic embolism in patients with nonvalvular atrial fibrillation

Echocardiography and risk stratification

Echocardiography is valuable to define the origin of AF (e.g., detecting rheumatic mitral valve disease or HCM) and may add information useful in stratifying thromboembolic risk. Among high-risk AF patients, impaired LV systolic function on transthoracic echocardiography, thrombus, dense spontaneous echo contrast or reduced velocity of blood flow in the LAA, and complex atheromatous plaque in the thoracic aorta on TEE have been associated with thromboembolism, and oral anticoagulation effectively lowers the risk of stroke in AF patients with these features. LA diameter and fibrocalcific endocardial abnormalities have been less consistently associated with thromboembolism. Whether the absence of these echocardiographic abnormalities identifies a low-risk group of patients who could safely avoid anticoagulation has not been established, limiting the value of echocardiography as a prime determinant of the need for chronic anticoagulation in patients with AF.

Several clinical schemes have been proposed to stratify the risk of ischemic stroke in patients with AF, based on analyses of prospectively monitored cohorts of participants in clinical trials in which antithrombotic therapy was controlled. Other criteria have been developed by expert consensus to classify patients into low-, intermediate-, and high-risk groups. Still others have used recursive partitioning and other techniques to identify low-risk patients. The CHADS2 (Cardiac Failure, Hypertension, Age, Diabetes, Stroke [Doubled]) integrates elements from several of these schemes and is based on a point system in which 2 points are assigned for a history of stroke or TIA and 1 point each is assigned for age over 75 y and a history of hypertension, diabetes, or recent HF (Table 9).152,153 The predictive value of this scoring system was evaluated in 1733 Medicare beneficiaries with nonvalvular AF between the ages of 65 and 95 y who were not given warfarin at hospital discharge. Although high scores were associated with an increased stroke rate in this elderly cohort, few patients had a score of 5 or more or a score of 0.

View this table:
Table 9

Stroke risk in patients with nonvalvular AF not treated with anticoagulation according to the CHADS2 index

Although these schemes for stratification of stroke risk identify patients who benefit most and least from anticoagulation, the threshold for use of anticoagulation is controversial. Opinion is particularly divided about anticoagulation for those at intermediate risk (stroke rate 3% to 5% per year). Some advocate the routine use of anticoagulation in those with stroke rates in this range,154 whereas others favor selective anticoagulation of patients at intermediate risk, with weight given to individual bleeding risks and patient preferences.24 The threshold of benefit at which AF patients choose anticoagulation varies; some at intermediate risk elect anticoagulation, whereas others do not.155 Our recommendations for antithrombotic therapy in patients with AF are summarized in Table 10.

View this table:
Table 10

Antithrombotic therapy for patients with atrial fibrillation

The risk of thromboembolism is not as well established for atrial flutter as it is for AF but is generally estimated as higher than that for patients with sinus rhythm and less than that for those with persistent or permanent AF. Although the overall thromboembolic risk associated with atrial flutter may be somewhat lower than with AF,156 it seems prudent to estimate risk by the use of similar stratification criteria for both arrhythmias until more robust data become available.

b. Antithrombotic strategies for prevention of ischemic stroke and systemic embolism

Before 1990, antithrombotic therapy for prevention of ischemic stroke and systemic embolism in patients with AF was limited mainly to those with rheumatic heart disease or prosthetic heart valves.23 Anticoagulation was also accepted therapy for patients who had sustained ischemic stroke to prevent recurrence but was often delayed to avoid hemorrhagic transformation. Some advocated anticoagulation of patients with thyrotoxicosis or other conditions associated with cardiomyopathy. Since then, 24 randomized trials involving patients with nonvalvular AF have been published, including 20 012 participants with an average follow-up of 1.6 y, a total exposure of about 32 800 patient-y.

Anticoagulation with vitamin K antagonist agents

Five large randomized trials published between 1989 and 1992 evaluated oral anticoagulation mainly for primary prevention of thromboembolism in patients with nonvalvular AF157163 (Figure 4). A sixth trial focused on secondary prevention among patients who had survived nondisabling stroke or cerebral TIA.164 Meta-analysis according to the principle of intention to treat showed that adjusted-dose oral anticoagulation is highly efficacious for prevention of all stroke (both ischemic and hemorrhagic), with a risk reduction of 61% (95% CI 47% to 71%) versus placebo165 (Figure 4). The duration of follow-up was generally between 1 and 2 y; the longest was 2.2 y, whereas in clinical practice, the need for antithrombotic therapy in patients with AF typically extends over much longer periods.

Figure 4

Antithrombotic therapy for prevention of stroke (ischemic and hemorrhagic) in patients with nonvalvular atrial fibrillation. Adjusted-dose warfarin compared with placebo (six random trials). Modified with permission from Hart RG, Benavente O, McBride R, et al. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492–501.165 AFASAK indicates Copenhagen Atrial Fibrillation, Aspirin, Anticoagulation; BAATAF, Boston Area Anticoagulation Trial for Atrial Fibrillation; CAFA, Canadian Atrial Fibrillation Anticoagulation; EAFT, European Atrial Fibrillation Trial; SPAF, Stroke Prevention in Atrial Fibrillation; and SPINAF, Stroke Prevention in Nonrheumatic Atrial Fibrillation.

All reported trials excluded patients considered at high risk of bleeding. Patient age and the intensity of anticoagulation are the most powerful predictors of major bleeding.166169 Trial participants, at an average age of 69 y, were carefully selected and managed, however, and it is unclear whether the relatively low observed rates of major hemorrhage also apply to patients with AF in clinical practice, who have a mean age of about 75 y and less closely regulated anticoagulation therapy.

The target intensity of anticoagulation involves a balance between prevention of ischemic stroke and avoidance of hemorrhagic complications (Figure 5). Targeting the lowest adequate intensity of anticoagulation to minimize the risk of bleeding is particularly important for elderly AF patients. Maximum protection against ischemic stroke in AF is probably achieved at an INR range of 2.0 to 3.0.170 Despite anticoagulation of more elderly patients with AF, rates of intracerebral hemorrhage are considerably lower than in the past, typically between 0.1% and 0.6% in contemporary reports. This may reflect lower anticoagulation intensity, more careful dose regulation, or better control of hypertension.171173

Figure 5

Adjusted odds ratios for ischemic stroke and intracranial bleeding in relation to intensity of anticoagulation. Modified with permission from Hylek EM, Singer DE. Risk factors for intracranial hemorrhage in outpatients taking warfarin. Ann Intern Med 1994;120:897–902.166 Data from Odén A, Fahlén M, Hart RG. Optimal INR for prevention of stroke and death in atrial fibrillation: a critical appraisal. Thromb Res 2006;117:493–9.167

Aspirin for antithrombotic therapy in patients with atrial fibrillation

Aspirin offers only modest protection against stroke for patients with AF157,161,164,174180 (Figure 6). Meta-analysis of 5 randomized trials showed a stroke reduction of 19% (95% CI=2% to 34%).165 Aspirin may be more efficacious for AF patients with hypertension or diabetes181 and for reduction of noncardioembolic versus cardioembolic ischemic strokes in AF patients.72 Cardioembolic strokes are, on average, more disabling than noncardioembolic strokes.92 Aspirin appears to prevent nondisabling strokes more than disabling strokes.165 Thus, the greater the risk of disabling cardioembolic stroke in a population of patients with AF, the less protection is afforded by aspirin.92

Figure 6

Antithrombotic therapy for prevention of stroke (ischemic and hemorrhagic) in patients with nonvalvular atrial fibrillation: warfarin compared with aspirin and aspirin compared with placebo. Modified with permission from Hart RG, Benavente O, McBride R, Pearce LA. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492–501.165 AFASAK indicates Copenhagen Atrial Fibrillation, Aspirin, Anticoagulation; EAFT, European Atrial Fibrillation Trial; ESPS, European Stroke Prevention Study; LASAF, Low-Dose Aspirin, Stroke, Atrial Fibrillation; UK-TIA, United Kingdom Transient Ischaemic Attack Aspirin Trial; PATAF, Prevention of Arterial Thromboembolism in Atrial Fibrillation; SPAF, Stroke Prevention in Atrial Fibrillation; and SPINAF, Stroke Prevention in Nonrheumatic Atrial Fibrillation.

Combining anticoagulant and platelet-inhibitor therapy

Combinations of oral anticoagulants plus antiplatelet agents have not generally shown reduced risks of hemorrhage or augmented efficacy over adjusted-dose anticoagulation alone. Combining aspirin with an oral anticoagulant at higher intensities may accentuate intracranial hemorrhage, particularly in elderly AF patients.183 For most patients with AF who have stable CAD, warfarin anticoagulation alone (target INR 2.0 to 3.0) should provide satisfactory antithrombotic prophylaxis against both cerebral and myocardial ischemic events.

Platelet-inhibitor drugs are particularly valuable for prevention of recurrent myocardial ischemia in patients undergoing percutaneous coronary intervention, but no adequate studies have been published that specifically address this issue in patients who also require chronic anticoagulation because of AF. It is the consensus of the authors of these guidelines that the most important agent for the maintenance of coronary and stent patency is the thienopyridine derivative clopidogrel and that the addition of aspirin to the chronic anticoagulant regimen contributes more risk than benefit. Although it is usually necessary to interrupt or reduce anticoagulation to prevent bleeding at the site of peripheral arterial puncture, the vitamin K antagonist should be resumed as soon as possible after the procedure and the dose adjusted to achieve an INR in the therapeutic range. Aspirin may be given temporarily during the hiatus, but the maintenance regimen should then consist of the combination of clopidogrel, 75 mg daily, plus warfarin (INR 2.0 to 3.0) for 9 to 12 mo, after which warfarin may be continued as monotherapy in the absence of a subsequent coronary event.

Low-molecular-weight heparins

The use of low-molecular-weight heparin instead of unfractionated heparin in patients with AF is based largely on extrapolation from venous thromboembolic disease states and from limited observational studies.184 In general, low-molecular-weight heparins have several pharmacological advantages over unfractionated heparin. These include a longer half-life, more predictable bioavailability (greater than 90% after subcutaneous injection), predictable clearance (enabling once- or twice-daily subcutaneous administration), and a predictable antithrombotic response based on body weight, which permits fixed-dose treatment without laboratory monitoring except under special circumstances such as obesity, renal insufficiency, or pregnancy.185 The favorable properties of low-molecular-weight heparins may simplify the treatment of AF in acute situations and shorten or eliminate the need for hospitalization to initiate anticoagulation. Self-administration of low-molecular-weight heparins out of hospital by patients with AF undergoing elective cardioversion is a promising approach that may result in cost savings.186

Interruption of anticoagulation for diagnostic or therapeutic procedures

From time to time, it may be necessary to interrupt oral anticoagulant therapy in preparation for elective surgical procedures. In patients with mechanical prosthetic heart valves, it is generally appropriate to substitute unfractionated or low-molecular-weight heparin to prevent thrombosis.187 In patients with AF who do not have mechanical valves, however, based on extrapolation from the annual rate of thromboembolism in patients with nonvalvular AF, it is the consensus of the Writing Group that anticoagulation may be interrupted for a period of up to 1 wk for surgical or diagnostic procedures that carry a risk of bleeding without substituting heparin. In high-risk patients (particularly those with prior stroke, TIA, or systemic embolism), or when a series of procedures requires interruption of oral anticoagulant therapy for longer periods, unfractionated or low-molecular-weight heparin may be administered intravenously or subcutaneously.

c. Nonpharmacological approaches to prevention of thromboembolism

An emerging option for patients with AF who cannot safely undergo anticoagulation, not yet sufficiently investigated to allow general clinical application, is obliteration of the LAA to remove a principal nidus of thrombus formation.188 In addition to direct surgical amputation or truncation of appendage, several methods are under development to achieve this with intravascular catheters or transpericardial approaches.189 The efficacy of these techniques is presumably related to the completeness and permanence of elimination of blood flow into and out of the LAA. This has been demonstrated by TEE at the time of intervention, but the durability of the effect has not been confirmed by subsequent examinations over several years. Whether mechanical measures intended to prevent embolism from thrombotic material in the LAA will prove comparably effective and safer than anticoagulation for some patients remains to be established.190

3. Cardioversion of atrial fibrillation

Cardioversion may be performed electively to restore sinus rhythm in patients with persistent AF. The need for cardioversion may be immediate when the arrhythmia is the main factor responsible for acute HF, hypotension, or worsening of angina pectoris in a patient with CAD. Nevertheless, cardioversion carries a risk of thromboembolism unless anticoagulation prophylaxis is initiated before the procedure, and this risk is greatest when the arrhythmia has been present more than 48 h.

Cardioversion may be achieved by means of drugs or electrical shocks. Drugs were commonly used before direct-current cardioversion became a standard procedure. The development of new drugs has increased the popularity of pharmacological cardioversion, but the disadvantages include the risk of drug-induced torsades de pointes or other serious arrhythmias. Moreover, pharmacological cardioversion is less effective than direct-current cardioversion when biphasic shocks are used. The disadvantage of electrical cardioversion is that it requires conscious sedation or anesthesia, which pharmacological cardioversion does not.

There is no evidence that the risk of thromboembolism or stroke differs between pharmacological and electrical methods of cardioversion. The recommendations for anticoagulation are therefore the same for both methods, as outlined in Section I.C.2, Preventing thromboembolism. Cardioversion in patients with AF following recent heart surgery or MI is addressed later (see Section I.C.5, Special considerations).

a. Pharmacological cardioversion

The quality of evidence available to gauge the effectiveness of pharmacological cardioversion is limited by small samples, lack of standard inclusion criteria (many studies include both patients with AF and those with atrial flutter), variable intervals from drug administration to assessment of outcome, and arbitrary dose selection. Although pharmacological and direct-current cardioversion have not been compared directly, pharmacological approaches appear simpler but are less efficacious. The major risk is related to the toxicity of antiarrhythmic drugs. In developing these guidelines, placebo-controlled trials of pharmacological cardioversion in which drugs were administered over short periods of time specifically to restore sinus rhythm have been emphasized. Trials in which the control group was given another antiarrhythmic drug have, however, been considered as well.

Pharmacological cardioversion seems most effective when initiated within 7 d after the onset of an episode of AF.191,192 A majority of these patients have a first-documented episode of AF or an unknown pattern of AF at the time of treatment. (See Section III, Classification.) A large proportion of patients with recent-onset AF experience spontaneous cardioversion within 24 to 48 h.193 Spontaneous conversion is less frequent in patients with AF of longer than 7-d duration, and the efficacy of pharmacological cardioversion is markedly reduced in these patients as well. Pharmacological cardioversion may accelerate restoration of sinus rhythm in patients with recent-onset AF, but the advantage over placebo is modest after 24 to 48 h, and drug therapy is much less effective in patients with persistent AF. Some drugs have a delayed onset of action, and conversion may not occur for several days after initiation of treatment.194 Drug treatment abbreviated the interval to cardioversion compared with placebo in some studies without affecting the proportion of patients who remained in sinus rhythm after 24 h.195 A potential interaction of antiarrhythmic drugs with vitamin K antagonist oral anticoagulants, increasing or decreasing the anticoagulant effect, is an issue whenever these drugs are added or withdrawn from the treatment regimen. The problem is amplified when anticoagulation is initiated in preparation for elective cardioversion. Addition of an antiarrhythmic drug to enhance the likelihood that sinus rhythm will be restored and maintained may perturb the intensity of anticoagulation beyond the intended therapeutic range, raising the risk of bleeding or thromboembolic complications.

A summary of recommendations concerning the use of pharmacological agents and recommended doses for cardioversion of AF is presented in Tables 1113. Algorithms for pharmacological management of AF are given in Figures 710. Throughout this document, reference is made to the Vaughan Williams classification of antiarrhythmic drugs,196 modified to include drugs that became available after the original classification was developed (Table 14). The recommendations given in this document are based on published data and do not necessarily adhere to the regulations and labeling requirements of governmental agencies. These antiarrhythmic drugs have been approved by federal regulatory agencies in the United States and/or Europe for clinical use, but their use for the treatment of AF has not been approved in all cases. Furthermore, not all agents are approved for use in all countries.

Figure 7

Pharmacological management of patients with newly discovered atrial fibrillation (AF). *See Figure 9. HF indicates heart failure.

Figure 8

Pharmacological management of patients with recurrent paroxysmal atrial fibrillation (AF). *See Figure 9. AAD indicates antiarrhythmic drug.

Figure 9

Antiarrhythmic drug therapy to maintain sinus rhythm in patients with recurrent paroxysmal or persistent atrial fibrillation. Within each box, drugs are listed alphabetically and not in order of suggested use. The vertical flow indicates order of preference under each condition. The seriousness of heart disease proceeds from left to right, and selection of therapy in patients with multiple conditions depends on the most serious condition present. See Section in the full-text guidelines for details. LVH indicates left ventricular hypertrophy.

Figure 10

Pharmacological management of patients with recurrent persistent or permanent atrial fibrillation (AF). *See Figure 9. Initiate drug therapy before cardioversion to reduce the likelihood of early recurrence of AF. AAD indicates antiarrhythmic drug.

View this table:
Table 11

Recommendations for pharmacological cardioversion of atrial fibrillation of up to 7-d duration

View this table:
Table 12

Recommendations for pharmacological cardioversion of atrial fibrillation present for more than 7 d

View this table:
Table 13

Recommended doses of drugs proven effective for pharmacological cardioversion of atrial fibrillation

View this table:
Table 14

Vaughan Williams classification of antiarrhythmic drugs

4. Pharmacological agents to maintain sinus rhythm

a. Agents with proven efficacy to maintain sinus rhythm

Thirty-six controlled trials evaluating 7 antiarrhythmic drugs for the maintenance of sinus rhythm in patients with parox-ysmal or persistent AF, 14 controlled trials of drug prophylaxis involving patients with paroxysmal AF, and 22 trials of drug prophylaxis for maintenance of sinus rhythm in patients with persistent AF were identified. Comparative data are not sufficient to permit subclassification by drug or etiology. Individual drugs and doses for maintenance of sinus rhythm are given in Table 15. It should be noted that any membrane active agent may cause proarrhythmia.

View this table:
Table 15

Typical doses of drugs used to maintain sinus rhythm in patients with atrial fibrillationa

b. Out-of-hospital initiation of antiarrhythmic drugs in patients with atrial fibrillation

A frequent issue related to pharmacological cardioversion of AF is whether to initiate antiarrhythmic drug therapy in hospital or on an outpatient basis. The major concern is the potential for serious adverse effects, including torsades de pointes. With the exception of those involving low-dose oral amiodarone,234 virtually all studies of pharmacological cardioversion have involved hospitalized patients. However, 1 study251 provided a clinically useful approach with out-of-hospital patient-controlled conversion using class IC drugs.

The ‘pill-in-the-pocket’ strategy consists of the self-administration of a single oral dose of drug shortly after the onset of symptomatic AF to improve quality of life, decrease hospital admission, and reduce cost.252 Recommendations for out-of-hospital initiation or intermittent use of antiarrhythmic drugs differ for patients with paroxysmal and persistent AF. In patients with paroxysmal AF, the aims are to terminate an episode or to prevent recurrence. In patients with persistent AF, the aims are to achieve pharmacological cardioversion of AF, obviating the need for direct-current cardioversion, or to enhance the success of direct-current cardioversion by lowering the defibrillation threshold and prevent early recurrence of AF.

In patients with lone AF without structural heart disease, class IC drugs may be initiated on an outpatient basis. For other selected patients without sinus or AV node dysfunction, bundle-branch block, QT-interval prolongation, the Brugada syndrome, or structural heart disease, ‘pill-in-the-pocket’ administration of propafenone and flecainide outside the hospital becomes an option once treatment has proved safe in hospital, given the relative safety (lack of organ toxicity and low estimated incidence of proarrhythmia).253255 Before these agents are initiated, however, a beta blocker or nondihydropyridine calcium channel antagonist is generally recommended to prevent rapid AV conduction in the event of atrial flutter.256,257 Unless AV node conduction is impaired, a short-acting beta blocker or nondihydropyridine calcium channel antagonist should be given at least 30 min before administration of a type IC antiarrhythmic agent to terminate an acute episode of AF, or the AV nodal blocking agents should be prescribed as continuous background therapy. Because termination of paroxysmal AF may be associated with bradycardia due to sinus node or AV node dysfunction, an initial conversion trial should be undertaken in hospital before a patient is declared fit for outpatient ‘pill-in-the pocket’ use of flecainide or propafenone for conversion of subsequent recurrences of AF. Table 16 lists other factors associated with proarrhythmic toxicity to class IC agents. It should be noted that these include female gender.

View this table:
Table 16

Factors predisposing to drug-induced ventricular proarrhythmia

Few prospective data are available on the relative safety of initiating antiarrhythmic drug therapy in the outpatient versus inpatient setting, and the decision to initiate therapy out of hospital should be carefully individualized. The ‘pill-in-the pocket’ approach appears feasible and safe for selected patients with AF, but the safety of this approach without previous inpatient evaluation remains uncertain.

As long as the baseline uncorrected QT interval is less than 460 ms, serum electrolytes are normal, and risk factors associated with class III drug-related proarrhythmia are considered (Table 16), sotalol may be initiated in outpatients with little or no heart disease. It is safest to start sotalol when the patient is in sinus rhythm. Amiodarone can also usually be given safely on an outpatient basis, even in patients with persistent AF, because it causes minimal depression of myocardial function and has low proarrhythmic potential,258 but in-hospital loading may be necessary for earlier restoration of sinus rhythm in patients with HF or other forms of hemodynamic compromise related to AF. Loading regimens typically call for administration of 600 mg daily for 4 wk258 or 1 g daily for 1 wk,232 followed by lower maintenance doses. Amiodarone, class IA or IC agents, or sotalol can be associated with bradycardia requiring permanent pacemaker implantation259; this is more frequent with amiodarone, and amiodarone-associated bradycardia is more common in women than in men. Quinidine, procainamide, and disopyramide should not be started out of hospital and out-of-hospital initiation of dofetilide is not currently permitted. Transtelephonic monitoring or other methods of ECG surveillance may be used to monitor cardiac rhythm and conduction as pharmacological antiarrhythmic therapy is initiated in patients with AF. Specifically, the PR interval (when flecainide, propafenone, sotalol, or amiodarone are used), QRS duration (with flecainide or propafenone), and QT interval (with dofetilide, sotalol, amiodarone, or disopyramide) should be measured.

As a general rule, antiarrhythmic drugs should be started at a relatively low dose and titrated based on response, and the ECG should be reassessed after each dose change. The heart rate should be monitored at approximately weekly intervals, either by checking the pulse rate, by use of an event recorder, or by ECG tracings obtained in the office. The dose of other medication for rate control should be reduced when the rate slows after initiation of amiodarone and stopped if the rate slows excessively. Concomitant drug therapies should be monitored closely, and both the patient and the physician should be alert to possible deleterious interactions. The doses of digoxin and warfarin, in particular, should usually be reduced upon initiation of amiodarone in anticipation of the rises in serum digoxin levels and INR that typically occur.

5. Direct-current cardioversion of atrial fibrillation and atrial flutter

a. Technical and procedural aspects

Direct-current cardioversion involves delivery of an electrical shock synchronized with the intrinsic activity of the heart by sensing the R wave of the ECG to ensure that electrical stimulation does not occur during the vulnerable phase of the cardiac cycle.136 Direct-current cardioversion is used to normalize all abnormal cardiac rhythms except ventricular fibrillation. The term ‘defibrillation’ implies an asynchronous discharge, which is appropriate for correction of ventricular fibrillation because R-wave synchronization is not feasible, but not for AF.

Successful cardioversion of AF depends on the underlying heart disease and the current density delivered to the atrial myocardium. Current may be delivered through external chest wall electrodes or through an internal cardiac electrode. Although the latter technique has been considered superior to external countershocks in obese patients and in patients with obstructive lung disease, it has not been widely applied. The frequency of recurrent AF does not differ between the 2 methods.135,260

Cardioversion should be performed with the patient under adequate general anesthesia in a fasting state. Short-acting anesthetic drugs or agents that produce conscious sedation are preferred to enable rapid recovery after the procedure; overnight hospitalization is seldom required.261 The electric shock should be synchronized with the QRS complex, triggered by monitoring the R wave with an appropriately selected ECG lead that also clearly displays atrial activation to facilitate assessment of outcome. In 64 patients randomly assigned to initial monophasic waveform energies of 100, 200, or 360 J, high initial energy was significantly more effective than low levels (immediate success rates 14% with 100, 39% with 200, and 95% with 360 J, respectively), resulting in fewer shocks and less cumulative energy when 360 J was delivered initially.262 These data indicate that an initial shock of 100 J with monophasic waveform is often too low for direct-current cardioversion of AF; hence, an initial energy of 200 J or greater is recommended. A similar recommendation to start with 200 J applies to biphasic waveforms, particularly when cardioverting patients with AF of long duration.263

When appropriate precautions are taken, cardioversion of AF is safe in patients with implanted pacemaker or defibrillator devices. Pacemaker generators and defibrillators are designed with circuits protected against sudden external electrical discharges, but programmed data may be altered by current surges. Electricity conducted along an implanted electrode may cause endocardial injury and lead to a temporary or permanent increase in stimulation threshold resulting in loss of ventricular capture. To ensure appropriate function, the implanted device should be interrogated and, if necessary, reprogrammed before and after cardioversion.

b. Risks and complications of direct-current cardioversion of atrial fibrillation

The risks of direct-current cardioversion are mainly related to thromboembolism and arrhythmias. Thromboembolic events have been reported in 1% to 7% of patients not given prophylactic anticoagulation before cardioversion of AF.264,265 Prophylactic antithrombotic therapy is discussed in Section VIII.B.5.c, Pharmacological enhancement of direct-current cardioversion.

c. Pharmacological enhancement of direct-current cardioversion

Although most recurrences of AF occur within the first month after direct-current cardioversion, research with internal atrial cardioversion270 and postconversion studies using transthoracic shocks271 have established several patterns of AF recurrence (Figure 7). In some cases, direct-current countershock fails to elicit even a single isolated sinus or ectopic atrial beat, tantamount to a high atrial defibrillation threshold. In others, AF recurs within a few minutes after a period of sinus rhythm,272 and recurrence after cardioversion is sometimes delayed for days or weeks.271 Complete shock failure and immediate recurrence occur in approximately 25% of patients undergoing direct-current cardioversion of AF, and subacute recurrences occur within 2 wk in almost an equal proportion.273

Restoration and maintenance of sinus rhythm are less likely when AF has been present for more than 1 y than in patients with AF of shorter duration. The variation in immediate success rates for direct-current cardioversion from 70% to 99% in the literature262,274276 is partly explained by differences in patient characteristics and the waveform used but also depends on the definition of success, because the interval at which the result is evaluated ranges from moments to several days. In general, it appears that sinus rhythm can be restored in a substantial proportion of patients by direct-current cardioversion, but the rate of relapse is high without concomitant antiarrhythmic drug therapy.

When given in conjunction with direct-current cardioversion, the primary aims of antiarrhythmic medication therapy are to increase the likelihood of success (e.g., by lowering the cardioversion threshold) and to prevent recurrent AF. Antiarrhythmic medications may be initiated out of hospital or in hospital immediately prior to direct-current cardioversion. (See Section 8.1.7 in the full-text guidelines, Out-of-hospital initiation of antiarrhythmic drugs in patients.) The risks of pharmacological treatment include the possibility of paradoxically increasing the defibrillation threshold, as described with flecainide,277 accelerating the ventricular rate when class IA or IC drugs are given without an AV nodal blocking agent,256,257,278,279 and inducing ventricular arrhythmias (Table 17).

View this table:
Table 17

Types of proarrhythmia during treatment with various antiarrhythmic drugs for AF or atrial flutter according to the Vaughan Williams classification

Patients with lone AF of relatively short duration are less prone to early recurrence of AF than are those with heart disease and longer AF duration, who therefore stand to gain more from prophylactic administration of antiarrhythmic medication. Pretreatment with pharmacological agents is most appropriate in patients who fail to respond to direct-current cardioversion and in those who develop immediate or subacute recurrence of AF. In patients with late recurrence and those undergoing initial cardioversion of persistent AF, pretreatment is optional. Antiarrhythmic drug therapy is recommended in conjunction with a second cardioversion attempt, particularly when early relapse has occurred. Additional cardioversion, beyond a second attempt, is of limited value and should be reserved for carefully selected patients, but infrequently repeated cardioversions may be acceptable in patients who are highly symptomatic upon relapse to AF.

Available data suggest that starting pharmacological therapy and establishing therapeutic plasma drug concentrations before direct-current cardioversion enhances immediate success and suppresses early recurrences. After cardioversion to sinus rhythm, patients receiving drugs that prolong the QT interval should be monitored in the hospital for 24 to 48 h to evaluate the effects of heart rate slowing and allow for prompt intervention in the event torsades de pointes develops (Table 18).

View this table:
Table 18

Pharmacological treatment before cardioversion in patients with persistent AF: effects of various antiarrhythmic drugs on immediate recurrence, outcome of transthoracic direct-current shock, or both

In randomized studies of direct-current cardioversion, patients pretreated with ibutilide were more often converted to sinus rhythm than untreated controls and those in whom cardioversion initially failed could more often be converted when the procedure was repeated after treatment with ibutilide.280,281

d. Prevention of thromboembolism in patients with atrial fibrillation undergoing conversion

Randomized studies of antithrombotic therapy are lacking for patients undergoing cardioversion of AF or atrial flutter, but in case-control series the risk of thromboembolism was between 1% and 5%.265,282 The risk was near the low end of this spectrum when anticoagulation (INR 2.0 to 3.0) was given for 3 to 4 wk before and after conversion.62,269 It is common practice to administer anticoagulant drugs when preparing patients with AF of more than 2-d duration for cardioversion. Manning et al.283 suggested that TEE might be used to identify patients without LAA thrombus who do not require anticoagulation, but a subsequent investigation284 and meta-analysis found this approach unreliable.285

If most AF-associated strokes result from embolism of stasis-induced thrombus from the LAA, then restoration and maintenance of atrial contraction should logically reduce thromboembolic risk. There is no evidence, however, that cardioversion followed by prolonged maintenance of sinus rhythm effectively reduces thromboembolism in AF patients. Conversion of AF to sinus rhythm results in transient mechanical dysfunction of the LA and LAA286 (‘stunning’), which can occur after spontaneous, pharmacological,287,288 or electrical288290 conversion of AF or after radiofrequency catheter ablation of atrial flutter.91 Recovery of mechanical function may be delayed, depending in part on the duration of AF before conversion.291293 This could explain why some patients without demonstrable LA thrombus on TEE before cardioversion subsequently experience thromboembolic events.284 Presumably, thrombus forms during the period of stunning and is expelled after the return of mechanical function, explaining the clustering of thromboembolic events during the first 10 d after cardioversion.84

Patients with AF or atrial flutter in whom LAA thrombus is identified by TEE are at high risk of thromboembolism and should be anticoagulated for at least 3 wk prior to and 4 wk after pharmacological or direct-current cardioversion. In a multicenter study, 1222 patients with either AF persisting longer than 2 d or atrial flutter and previous AF294 were randomized to a TEE-guided or conventional strategy. In the group undergoing TEE, cardioversion was postponed when thrombus was identified, and warfarin was administered for 3 wk before TEE was repeated to confirm resolution of thrombus. Anticoagulation with heparin was used briefly before cardioversion and with warfarin for 4 wk after cardioversion. The other group received anticoagulation for 3 wk before and 4 wk after cardioversion without intercurrent TEE. Both approaches were associated with comparably low risks of stroke (0.81% with the TEE approach and 0.50% with the conventional approach) after 8 wk, there were no differences in the proportion of patients achieving successful cardioversion, and the risk of major bleeding did not differ significantly. The clinical benefit of the TEE-guided approach was limited to saving time before cardioversion.

Anticoagulation is recommended for 3 wk prior to and 4 wk after cardioversion for patients with AF of unknown duration or with AF for longer than 48 h. Although LA thrombus and systemic embolism have been documented in patients with AF of shorter duration, the need for anticoagulation is less clear. When acute AF produces hemodynamic instability in the form of angina pectoris, MI, shock, or pulmonary edema, immediate cardioversion should not be delayed to deliver therapeutic anticoagulation, but intravenous unfractionated heparin or subcutaneous injection of a low-molecular-weight heparin should be initiated before cardioversion by direct-current countershock or intravenous antiarrhythmic medication.

Protection against late embolism may require continuation of anticoagulation for a more extended period after the procedure, and the duration of anticoagulation after cardioversion depends both on the likelihood that AF will recur in an individual patient with or without symptoms and on the intrinsic risk of thromboembolism. Late events are probably due to both the development of thrombus as a consequence of atrial stunning and the delayed recovery of atrial contraction after cardioversion. Stroke or systemic embolism has been reported in patients with atrial flutter undergoing cardioversion,295297 and anticoagulation should be considered with either the conventional or TEE-guided strategy. TEE-guided cardioversion of atrial flutter has been performed with a low rate of systemic embolism, particularly when patients are stratified for other risk factors on the basis of clinical and/or TEE features.

6. Maintenance of sinus rhythm

a. Pharmacological therapy

Whether paroxysmal or persistent, AF is a chronic disorder, and recurrence at some point is likely in most patients.299301 Many patients eventually need prophylactic antiarrhythmic drug therapy to maintain sinus rhythm, suppress symptoms, improve exercise capacity and hemodynamic function, and prevent tachycardia-induced cardiomyopathy due to AF. Because factors that predispose to recurrent AF (advanced age, HF, hypertension, LA enlargement, and LV dysfunction) are risk factors for thromboembolism, the risk of stroke may not be reduced by correction of the rhythm disturbance. Trials in which rate versus rhythm control strategies were compared in patients with persistent and paroxysmal AF124,126,127,130,132 found no reduction in death, disabling stroke, hospitalizations, new arrhythmias, or thromboembolic complications in the rhythm control group.

b. Predictors of recurrent atrial fibrillation

Most patients with AF, except those with postoperative or self-limited AF secondary to transient or acute illness, eventually experience recurrence. Risk factors for frequent recurrence of paroxysmal AF (more than 1 episode per month) include female gender and underlying heart disease.302 In 1 study of patients with persistent AF, the 4-y arrhythmia-free survival rate was less than 10% after single-shock direct-current cardioversion without prophylactic drug therapy.300 Predictors of recurrences within that interval included hypertension, age over 55 y, and AF duration of longer than 3 mo. Serial cardioversions and prophylactic drug therapy resulted in freedom from recurrent AF in approximately 30% of patients,300 and with this approach predictors of recurrence included age over 70 y, AF duration beyond 3 mo, and HF.300 Other risk factors for recurrent AF include LA enlargement and rheumatic heart disease.

Various benign arrhythmias, especially ventricular and supraventricular premature beats, bradycardia, and short periods of sinus arrest, may arise after cardioversion and commonly subside spontaneously.266 More dangerous arrhythmias, such as ventricular tachycardia and fibrillation, may arise in the face of hypokalemia, digitalis intoxication, or improper synchronization.267,268 Serum potassium levels should be in the normal range for safe, effective cardioversion. Cardioversion is contraindicated in cases of digitalis toxicity because resulting ventricular tachyarrhythmia may be difficult to terminate.

In patients with long-standing AF, cardioversion commonly unmasks underlying sinus node dysfunction. A slow ventricular response to AF in the absence of drugs that slow conduction across the AV node may indicate an intrinsic conduction defect. The patient should be evaluated before cardioversion with this in mind so a transvenous or transcutaneous pacemaker can be used prophylactically.269

c. General approach to antiarrhythmic drug therapy

Before administering any antiarrhythmic agent, reversible precipitants of AF should be identified and corrected. Most are related to coronary or valvular heart disease, hypertension, or HF. Patients who develop AF in association with alcohol intake should abstain from alcohol consumption. Indefinite antiarrhythmic treatment is seldom prescribed after a first episode, although a period of several weeks may help stabilize sinus rhythm after cardioversion. Similarly, patients experiencing breakthrough arrhythmias may not require a change in antiarrhythmic drug therapy when recurrences are infrequent and mild. Beta-adrenergic antagonist medication may be effective in patients who develop AF only during exercise, but a single, specific inciting cause rarely accounts for all episodes of AF, and the majority of patients do not sustain sinus rhythm without antiarrhythmic therapy. Selection of an appropriate agent is based first on safety, tailored to whatever underlying heart disease may be present, considering the number and pattern of prior episodes of AF.303

In patients with lone AF, a beta blocker may be tried first, but flecainide, propafenone, and sotalol are particularly effective. Amiodarone and dofetilide are recommended as alternative therapies. Quinidine, procainamide, and disopyramide are not favored unless amiodarone fails or is contraindicated. For patients with vagally induced AF, however, the anticholinergic activity of long-acting disopyramide makes it a relatively attractive theoretical choice. In that situation, flecainide and amiodarone represent secondary and tertiary treatment options, respectively, whereas propafenone is not recommended because its (weak) intrinsic beta-blocking activity may aggravate vagally mediated paroxysmal AF. In patients with adrenergically mediated AF, beta blockers represent first-line treatment, followed by sotalol and amiodarone. In patients with adrenergically mediated lone AF, amiodarone represents a less appealing selection. Vagally induced AF can occur by itself, but more typically it is part of the overall patient profile. In patients with nocturnal AF, the possibility of sleep apnea should be considered.

When treatment with a single antiarrhythmic drug fails, combinations may be tried. Useful combinations include a beta blocker, sotalol, or amiodarone with a class IC agent. The combination of a calcium channel blocker, such as diltiazem, with a class IC agent, such as flecainide or propafenone, is advantageous in some patients. A drug that is initially safe may become proarrhythmic if coronary disease or HF develops or if the patient begins other medication that exerts a proarrhythmic interaction. Thus, the patient should be alerted to the potential significance of such symptoms as syncope, angina, or dyspnea and warned about the use of noncardiac drugs that might prolong the QT interval.

The optimum method for monitoring antiarrhythmic drug treatment varies with the agent involved as well as with patient factors. Prospectively acquired data on upper limits of drug-induced prolongation of QRS duration or QT interval are not available. With class IC drugs, prolongation of the QRS interval should not exceed 50%. Exercise testing may help detect QRS widening that occurs only at rapid heart rates (use-dependent conduction slowing). For class IA or class III drugs, with the possible exception of amiodarone, the corrected QT interval in sinus rhythm should be kept below 520 ms. During follow-up, plasma potassium and magnesium levels and renal function should be checked periodically because renal insufficiency leads to drug accumulation and predisposes to proarrhythmia. In individual patients, serial noninvasive assessment of LV function is indicated, especially when clinical HF develops during treatment of AF.

d. Selection of antiarrhythmic agents in patients with cardiac diseases

Pharmacological management algorithms to maintain sinus rhythm in patients with AF (see Figures 710) and applications in specific cardiac disease states are based on available evidence and extrapolated from experience with these agents in other situations.

Heart failure

Patients with HF are particularly prone to the ventricular proarrhythmic effects of antiarrhythmic drugs because of myocardial vulnerability and electrolyte imbalance. Randomized trials have demonstrated the safety of amiodarone and dofetilide (given separately) in patients with HF,200,304 and these are the recommended drugs for maintenance of sinus rhythm in patients with AF in the presence of HF (Figure 9). Patients with LV dysfunction and persistent AF should be treated with beta blockers and ACE inhibitors and/or angiotensin II receptor antagonists, because these agents help control the heart rate, improve ventricular function, and prolong survival.305308

Coronary artery disease

In stable patients with CAD, beta blockers may be considered first, although their use is supported by only 2 studies309,310 and data on efficacy for maintenance of sinus rhythm in patients with persistent AF after cardioversion are not convincing.310 Sotalol has substantial beta-blocking activity and may be the preferred initial antiarrhythmic agent in patients with AF who have ischemic heart disease because it is associated with less long-term toxicity than amiodarone. Amiodarone increases the risk of bradyarrhythmia requiring permanent pacemaker implantation in elderly patients with AF who have previously sustained MI311 but may be preferred over sotalol in patients with HF.312,313 Neither flecainide nor propafenone is recommended in these situations, but quinidine, procainamide, and disopyramide may be considered as third-line choices in patients with coronary disease. The DIAMOND-MI (Danish Investigations of Arrhythmias and Mortality on Dofetilide in Myocardial Infarction) trial314 involved selected post-MI patients in whom the antiarrhythmic benefit of dofetilide balanced the risk of proarrhythmic toxicity, making this a second-line antiarrhythmic agent. In patients with coronary disease who have not developed MI or HF, however, it is uncertain whether the benefit of dofetilide outweighs risk, and more experience is needed before this drug can be recommended for such patients (Figure 9).

Hypertensive heart disease

Hypertension is the most prevalent and potentially modifiable independent risk factor for the development of AF and its complications, including thromboembolism.315,316 Blood pressure control may become an opportune strategy for the prevention of AF. Patients with LVH may face an increased risk of torsades de pointes related to early ventricular afterdepolarizations.303,317 Thus, class IC agents and amiodarone are preferred over type IA and type III antiarrhythmic agents as first-line therapy. In the absence of ischemia or LVH, both propafenone and flecainide are reasonable choices. Proarrhythmia with 1 agent does not predict this response to another, and patients with LVH who develop torsades de pointes during treatment with a class III agent may tolerate a class IC agent. Amiodarone prolongs the QT interval but carries a very low risk of ventricular proarrhythmia. Its extracardiac toxicity relegates it to second-line therapy in these individuals, but it becomes a first-line agent in the face of substantial LVH. When amiodarone and sotalol either fail or are inappropriate, disopyramide, quinidine, or procainamide represents a reasonable alternative.

Beta blockers may be the first line of treatment to maintain sinus rhythm in patients with MI, HF, and hypertension. Compared to patients with lone AF, those with hypertension are more likely to maintain sinus rhythm after cardioversion of persistent AF when treated with a beta blocker.318 Drugs modulating the RAAS reduce structural cardiac changes,319 and ACE inhibition was associated with a lower incidence of AF compared with calcium channel blockade in patients with hypertension during 4.5 y of follow-up in a retrospective, longitudinal cohort study from a database of 8 million patients in a managed care setting.320 In patients at increased risk of cardiovascular events, therapy with either the ACE inhibitor ramipril321323 or angiotensin receptor antagonist losartan324,325 lowered the risk of stroke. A similar benefit has been reported with perindopril in a subset of patients with AF treated for prevention of recurrent stroke.326 New-onset AF and stroke were significantly reduced by losartan as compared to atenolol in hypertensive patients with ECG-documented LVH, despite a similar reduction of blood pressure.16 The benefit of losartan was greater in patients with AF than in those with sinus rhythm for the primary composite endpoint (cardiovascular mortality, stroke, and MI) and for cardiovascular mortality alone.327 Presumably, the beneficial effects of beta blockers and drugs modulating the RAAS are at least partly related to lower blood pressure.

7. Nonpharmacological therapy for atrial fibrillation

The inconsistent efficacy and potential toxicity of antiarrhythmic drug therapies have stimulated exploration of a wide spectrum of alternative nonpharmacological therapies for the prevention and control of AF.

a. Surgical ablation

A decade of research in the 1980s demonstrated the critical elements necessary to cure AF surgically, including techniques that entirely eliminate macroreentrant circuits in the atria while preserving sinus node and atrial transport functions. The surgical approach was based on the hypothesis that reentry is the predominant mechanism responsible for the development and maintenance of AF,328 leading to the concept that atrial incisions at critical locations would create barriers to conduction and prevent sustained AF. The procedure developed to accomplish these goals was based on the concept of a geographical maze, accounting for the term ‘maze’ procedure used to describe this type of cardiac operation.329

Since its introduction, the procedure has gone through 3 iterations (maze I, II, and III) using cut-and-sew techniques that ensure transmural lesions to isolate the PV, connect these dividing lines to the mitral valve annulus, and create electrical barriers in the RA that prevent macroentrant rhythms—atrial flutter or AF—from becoming sustained.330 Success rates of around 95% over 15 y of follow-up have been reported in patients undergoing mitral valve surgery.331 Other studies suggest success rates around 70%.332 Atrial transport function is maintained and, when combined with amputation or obliteration of the LAA, postoperative thromboembolic events are substantially reduced. Risks include death (less than 1% when performed as an isolated procedure), the need for permanent pacing (with right-sided lesions), recurrent bleeding requiring re-operation, impaired atrial transport function, delayed atrial arrhythmias (especially atrial flutter), and atrioesophageal fistula.

Despite its high success rate, the maze operation has not been widely adopted other than for patients undergoing cardiac surgery because of the need for cardiopulmonary bypass. A wide variety of less invasive modifications are under investigation, including thoracoscopic and catheter-based epicardial techniques.332 If the efficacy of these adaptations approaches that of the endocardial maze procedure and they can be performed safely, they may become acceptable alternatives for a larger proportion of patients with AF.

b. Catheter ablation

Early radiofrequency catheter ablation techniques emulated the surgical maze procedure by introducing linear scars in the atrial endocardium.333 While the success rate was approximately 40% to 50%, a relatively high complication rate diminished enthusiasm for this approach.38 The observation that potentials arising in or near the ostia of the PV often provoked AF and demonstration that elimination of these foci abolished AF escalated enthusiasm for catheter-based ablation.38 Initially, areas of automaticity within the PV were targeted, and in a series of 45 patients with paroxysmal AF, 62% became free of symptomatic AF over a mean follow-up of 8 mo, but 70% required multiple procedures.38 In another study, the success rate was 86% over a 6-mo follow-up.334 Subsequent research has demonstrated that potentials may arise in multiple regions of the RA and LA, including the LA posterior wall, superior vena cava, vein of Marshall, crista terminalis, interatrial septum, and coronary sinus,335 and modification of the procedures has incorporated linear LA ablation, mitral isthmus ablation, or both for selected patients.336

The technique of ablation has continued to evolve from early attempts to target individual ectopic foci within the PV to circumferential electrical isolation of the entire PV musculature. In a series of 70 patients, 73% were free from AF following PV isolation without antiarrhythmic medications during a mean follow-up of 4 mo, but 29 patients required a second procedure to reach this goal. However, post-ablation AF may occur transiently in the first 2 mo.337 Advances involving isolation of the PV at the antrum using a circular mapping catheter, guided by intracardiac echocardiography, have reportedly yielded approximately 80% freedom of recurrent AF or atrial flutter after the first 2 mo in patients with paroxysmal AF,338 but success rates were lower in patients with cardiac dysfunction.339 Still another approach340,341 uses a nonfluoroscopic guidance system and radiofrequency energy delivered circumferentially outside the ostia of the PV. In a series of 26 patients, 85% were free of recurrent AF during a mean follow-up of 9 mo, including 62% taking no antiarrhythmic medications. The accumulated experience involves nearly 4000 patients,341 with approximately 90% success in cases of paroxysmal AF and 80% in cases of persistent AF.339,342,343

Another anatomic approach to radiofrequency catheter ablation targets complex fractionated electrograms344 with 91% efficacy reported at 1 y. Restoration of sinus rhythm after catheter ablation for AF significantly improved LV function, exercise capacity, symptoms, and quality of life (usually within the first 3 to 6 mo), even in the presence of concurrent heart disease and when ventricular rate control was adequate before ablation.345 While that study lacked a control group of patients with HF, in another study catheter ablation of AF was associated with reduced mortality and morbidity due to HF and thromboembolism.346

In selected patients, radiofrequency catheter ablation of the AV node and pacemaker insertion decreased symptoms of AF and improved quality-of-life scores compared with medication therapy.140142,347349 A meta-analysis of 10 studies of patients with AF143 found improvement in both symptoms and quality-of-life scores after ablation and pacing. Although these studies involved selected patients who remained in AF, the consistent improvement suggests that quality of life was impaired before intervention.

Despite these advances, the long-term efficacy of catheter ablation to prevent recurrent AF requires further study. Available data demonstrate 1 or more y free from recurrent AF in most (albeit carefully selected) patients.350352 It is important to bear in mind, however, that AF can recur without symptoms and be unrecognized by the patient or the physician. Therefore, it remains uncertain whether apparent cures represent elimination of AF or transformation into an asymptomatic form of paroxysmal AF. The distinction has important implications for the duration of anticoagulation therapy in patients with risk factors for stroke associated with AF. In addition, little information is available about the late success of ablation in patients with HF and other advanced structural heart disease, who may be less likely to enjoy freedom from AF recurrence.

Complications of catheter ablation include the adverse events associated with any cardiac catheterization procedure in addition to those specific to ablation of AF. Major complications have been reported in about 6% of procedures and include PV stenosis, thromboembolism, atrioesophageal fistula, and LA flutter.343 The initial ablation approach targeting PV ectopy was associated with an unacceptably high rate of PV stenosis,334,353 but the incidence has dramatically decreased as a result of changes in technique. Current approaches avoid delivering radiofrequency energy within the PV and instead target areas outside the veins to isolate the ostia from the remainder of the LA conducting tissue. Use of intracardiac echocardiographically detected microbubble formation to titrate radiofrequency energy has also been reported to reduce the incidence of PV stenosis.338

Embolic stroke is among the most serious complications of catheter-based ablation procedures in patients with AF. The incidence varies from 0% to 5%. A higher intensity of anticoagulation reduces the risk of thrombus formation during ablation.354 Based on limited data from dose-comparison studies, it seems likely that more aggressive anticoagulation may reduce the incidence of thromboembolism associated with catheter-based ablation of AF.

Atrioesophageal fistula has been reported with both the circumferential Pappone approach355,356 and the Haissaguerre PV ablation techniques356 but is relatively rare. This complication may be more likely to occur when extensive ablative lesions are applied to the posterior LA wall, increasing the risk of atrial perforation. The typical manifestations include sudden neurological symptoms or endocarditis, and the outcome in most cases is, unfortunately, fatal. Depending on the ablation approach, LA flutter may develop during treatment of AF,357 and this is amenable to further ablation.358

Future directions in catheter-based ablation therapy for atrial fibrillation

Catheter-directed ablation of AF represents a substantial achievement that promises better therapy for a large number of patients presently resistant to pharmacological or electrical conversion to sinus rhythm. The limited available studies suggest that catheter-based ablation offers benefit to selected patients with AF, but these studies do not provide convincing evidence of optimum catheter positioning or absolute rates of treatment success. Identification of patients who might benefit from ablation must take into account both potential benefits and short- and long-term risks. Rates of success and complications vary, sometimes considerably, from one study to another because of patient factors, patterns of AF, criteria for definition of success, duration of follow-up, and technical aspects. Registries of consecutive case series should incorporate clear and prospectively defined outcome variables. Double-blind studies are almost impossible to perform, yet there is a need for randomized trials in which evaluation of outcomes is blinded as to treatment modality. A comprehensive evaluation of the favorable and adverse effects of various ablation techniques should include measures of quality of life and recurrence rates compared with pharmacological strategies for rhythm control and, when this is not successful, with such techniques of rate control as AV node ablation and pacing. Generation of these comparative data over relatively long periods of observation would address the array of invasive and conservative management approaches available for management of patients with AF and provide a valuable foundation for future practice guidelines.

c. Suppression of atrial fibrillation through pacing

Several studies have examined the role of atrial pacing, either in the RA alone or in more than 1 atrial location, to prevent recurrent paroxysmal AF. In patients with symptomatic bradycardia, the risk of AF is lower with atrial than with ventricular pacing.359 In patients with sinus node dysfunction and normal AV conduction, data from several randomized trials support atrial or dual-chamber rather than ventricular pacing for prevention of AF.360363 The mechanisms by which atrial pacing prevents AF in patients with sinus node dysfunction include prevention of bradycardia-induced dispersion of repolarization and suppression of atrial premature beats. Atrial or dual-chamber pacing also maintains AV synchrony, preventing retrograde ventriculoatrial conduction, which can cause valvular regurgitation and stretch-induced changes in atrial electrophysiology. When ventricular pacing with dual-chamber devices is unavoidable because of concomitant disease of the AV conduction system, the evidence is less clear that atrial-based pacing is superior. Although atrial-based pacing is associated with a lower risk of AF and stroke than ventricular-based pacing in patients requiring pacemakers for bradyarrhythmias, the value of pacing as a primary therapy for prevention of recurrent AF has not been proven.

d. Internal atrial defibrillators

There has been an interest in internal cardioversion of AF for the past 10 y.135 Intense basic and clinical research to find tolerable shock waveforms led to evaluation of an implantable device capable of both atrial sensing and cardioversion and ventricular sensing and pacing in 290 patients with mean LV ejection fraction greater than 50% who had not responded satisfactorily to therapy with 4 antiarrhythmic drugs.135 The rate of conversion to sinus rhythm was 93%. As spontaneous episodes were treated quickly, the interval between episodes of AF lengthened. Several available devices combining both atrial cardioversion and ventricular defibrillation capabilities with dual-chamber sensing and pacing have been designed to treat both atrial and ventricular arrhythmias by pacing before delivering low- or high-energy shocks.

An important limitation of atrial defibrillators, unrelated to efficacy, is that most patients find discharge energies over 1 J uncomfortable. Candidates for atrial cardioverters with infrequent episodes of poorly tolerated AF are typically also candidates for catheter ablation. As a result, implanted devices have limited utility, except for patients with LV dysfunction who are candidates for implantable ventricular defibrillators.

C. Primary prevention

Although measures aimed at the primary prevention of AF have not been widely investigated, it has been suggested that atrial or AV synchronous pacing may reduce the incidence of subsequent AF in patients with bradycardia compared with ventricular pacing.359,360 On the other hand, studies in patients with intermittent atrial tachyarrhythmias failed to illustrate a general benefit of atrial pacing.360,365,366 Another potential avenue for primary prevention has been suggested following secondary analysis of placebo-controlled trials of treatment with ACE inhibitors.56,367 In the LIFE (Losartan Intervention For End Point Reduction in Hypertension)16 and CHARM (Candesartan in Heart Failure Assessment of Reduction in Mortality and Morbidity)368 trials, the angiotensin receptor antagonists losartan and candesartan reduced the incidence of AF in hypertensive patients with LVH16 and symptomatic HF,21,368 respectively. These results, together with their favorable safety profile compared with antiarrhythmic agents, suggest a role for ACE inhibitor or angiotensin receptor antagonists for primary prevention of initial or recurrent episodes of AF associated with hypertension, MI, HF, or diabetes mellitus. An overview of 11 clinical trials involving more than 56 000 patients with different underlying cardiovascular diseases suggests that ACE inhibitors or angiotensin receptor blockers may reduce the occurrence and recurrence of AF.55

Yet inadequately explored, the use of statins has also been suggested to protect against AF,369,370 and dietary lipid components may influence the propensity of patients to develop AF.371 In 449 patients with CAD followed for 5 y, statin therapy reduced the incidence of AF—an effect not observed with other lipid-lowering drugs.369 Insufficient data are available at this time to permit recommendations for primary prevention of AF in populations at risk using dietary interventions, pharmacological interventions, or pacing or other devices.

IX. Proposed management strategies

A. Overview of algorithms for management of patients with atrial fibrillation

Management of patients with AF requires knowledge of its pattern of presentation (paroxysmal, persistent, or permanent), underlying conditions, and decisions about restoration and maintenance of sinus rhythm, control of the ventricular rate, and antithrombotic therapy. These issues are addressed in the various management algorithms for each presentation of AF (see Figures 710).

1. Newly discovered atrial fibrillation

It is not always clear whether the initial presentation of AF is actually the first episode, particularly in patients with minimal or no symptoms related to the arrhythmia. In patients who have self-limited episodes of AF, antiarrhythmic drugs are usually unnecessary to prevent recurrence unless AF is associated with severe symptoms related to hypotension, myocardial ischemia, or HF. Regarding anticoagulation, the results of the AFFIRM study132 indicate that patients with AF who are at high risk for stroke on the basis of identified risk factors generally benefit from anticoagulation even after sinus rhythm has been restored. Therefore, unless there is a clear reversible precipitating factor for AF, such as hyperthyroidism that has been corrected, the diagnosis of AF in a patient with risk factors for thromboembolism should prompt long-term anticoagulation.

When AF persists, one option is to accept progression to permanent AF, with attention to antithrombotic therapy and control of the ventricular rate. Although it may seem reasonable to make at least one attempt to restore sinus rhythm, the AFFIRM study showed no difference in survival or quality of life with rate control compared to a rhythm control strategies.132 Other trials that addressed this issue reached similar conclusions.124,126,127,130 Hence, the decision to attempt restoration of sinus rhythm should be based on the severity of arrhythmia-related symptoms and the potential risk of antiarrhythmic drugs. If the decision is made to attempt to restore and maintain sinus rhythm, then anticoagulation and rate control are important before cardioversion. Although long-term antiarrhythmic therapy may not be needed to prevent recurrent AF after cardioversion, short-term therapy may be beneficial. In patients with AF that has been present for longer than 3 mo, early recurrence is common after cardioversion. In such cases, antiarrhythmic medication may be initiated before cardioversion (after adequate anticoagulation) to reduce the likelihood of recurrence, and the duration of drug therapy would be brief (e.g., 1 mo).

2. Recurrent paroxysmal atrial fibrillation

In patients who experience brief or minimally symptomatic recurrences of paroxysmal AF, it is reasonable to avoid antiarrhythmic drugs, but troublesome symptoms generally call for suppressive antiarrhythmic therapy. Rate control and prevention of thromboembolism are appropriate in both situations. In a given patient, several antiarrhythmic drugs may be effective and the initial selection is based mainly on safety and tolerability (Figure 9). For individuals with no or minimal heart disease, flecainide, propafenone, or sotalol is recommended as initial antiarrhythmic therapy because they are generally well tolerated and carry relatively little risk of toxicity. For patients with recurrent episodes of symptomatic AF who tolerate these agents, an as-needed, pill-in-the-pocket approach may reduce the risk of toxicity compared with sustained therapy. When these drugs prove ineffective or are associated with side effects, the second- or third-line choices include amiodarone, dofetilide, disopyramide, procainamide, or quinidine, all of which carry greater potential for adverse reactions. As an alternative to treatment with amiodarone or dofetilide when first-line antiarrhythmic drugs fail or are not tolerated, PV isolation or LA substrate modification may be considered. When a consistent initiating scenario suggests vagally mediated AF, drugs such as disopyramide or flecainide are appropriate initial agents, and a beta blocker or sotalol is suggested for patients with adrenergically induced AF. In particularly symptomatic patients, nonpharmacological options such as LA ablation may be considered when antiarrhythmic drug treatment alone fails to control the arrhythmia.

Many patients with organic heart disease can be broadly categorized into those with HF, CAD, or hypertension. Other types of heart disease can be associated with AF, and the clinician must determine which category best describes the individual patient. For patients with HF, safety data support the selection of amiodarone or dofetilide to maintain sinus rhythm. Patients with CAD often require beta-blocker medication, and sotalol, a drug with both beta-blocking activity and primary antiarrhythmic efficacy, is considered first, unless the patient has HF. Amiodarone and dofetilide are considered secondary agents, and the clinician should consider disopyramide, procainamide, or quinidine on an individual basis.

The selection of antiarrhythmic drugs for patients with a history of hypertension is confounded by the dearth of prospective, controlled trials comparing the safety and efficacy of drug therapy for AF. In patients with hypertension without LVH, drugs such as flecainide and propafenone, which do not prolong repolarization or the QT interval, may offer a safety advantage and are recommended first. If these agents either prove ineffective or produce side effects, then amiodarone, dofetilide, or sotalol represents an appropriate secondary choice. Disopyramide, procainamide, and quinidine are considered third-line agents in this situation. Hypertrophied myocardium may be prone to proarrhythmic toxicity and torsades de pointes ventricular tachycardia. Amiodarone is suggested as first-line therapy in patients with LVH because of its relative safety compared with several other agents. Because neither ECG nor echocardiography reliably detects LVH as defined by measurement of myocardial mass, clinicians may face a conundrum.

The scarcity of data from randomized trials of antiarrhythmic medications for treatment of patients with AF applies generally to all patient groups. Accordingly, the drug-selection algorithms presented here have been developed by consensus and are subject to revision as additional evidence emerges.

3. Recurrent persistent atrial fibrillation

Patients with minimal or no symptoms referable to AF who have undergone at least one attempt to restore sinus rhythm may remain in AF after recurrence, with therapy for rate control and prevention of thromboembolism as needed. Alternatively, those with symptoms favoring sinus rhythm should be treated with an antiarrhythmic agent (in addition to medications for rate control and anticoagulation) before cardioversion. The selection of an antiarrhythmic drug should be based on the same algorithm used for patients with recurrent paroxysmal AF. If patients remain symptomatic with heart rate control and antiarrhythmic medication is either not tolerated or ineffective, then nonpharmacological therapies may be considered. These include LA ablation, the maze operation, or AV nodal ablation and pacing.

4. Permanent atrial fibrillation

Permanent AF is the designation given to cases in which sinus rhythm cannot be sustained after cardioversion of AF or when the patient and physician have decided to allow AF to continue without further efforts to restore sinus rhythm. It is important to maintain control of the ventricular rate and to use antithrombotic therapy, as outlined elsewhere in this document, for all patients in this category.


View this table:
Appendix I

Relationships with industry—ACC/AHA committee to update the 2001 guidelines for the management of patients with atrial fibrillation

View this table:
Appendix II

Relationships with industry—external peer review for the ACC/AHA/ESC Committee to update the 2001 guidelines for the management of patients with atrial fibrillation

View this table:
Appendix III



  • European Heart Rhythm Association Official Representative.

  • Heart Rhythm Society Official Representative.

  • Immediate Past Chair.

  • This document was approved by the American College of Cardiology Foundation Board of Trustees in June 2006; by the American Heart Association Science Advisory and Coordinating Committee in June 2006; and by the European Society of Cardiology Committee for Practice Guidelines in June 2006.

  • When this document is cited, the American College of Cardiology Foundation, the American Heart Association, and the European Society of Cardiology request that the following citation format be used: Fuster V, Rydén LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, Halperin JL, Le Heuzey J-Y, Kay GN, Lowe JE, Olsson SB, Prystowsky EN, Tamargo JL, Wann S, Smith SC, Jacobs AK, Adams CD, Anderson JL, Antman EM, Hunt SA, Nishimura R, Ornato JP, Page RL, Riegel B, Priori SG, Blanc J-J, Budaj A, Camm AJ, Dean V, Deckers JW, Despres C, Dickstein K, Lekakis J, McGregor K, Metra M, Morais J, Osterspey A, Zamorano JL. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation—executive summary: a report of the American College of Cardiology/American Heart Association Task Force and the European Society of Cardiology Committee for Practice Guidelines (Writing Committee to Revise the 2001 Guidelines for the Management of Patients With Atrial Fibrillation). Eur Heart J 2006;27:1979–2030.

  • This article has been copublished in the August 15, 2006, issues of Circulation and the Journal of the American College of Cardiology and the August 16, 2006, issue of the European Heart Journal.

  • Copies: This document is available on the World Wide Web sites of the American College of Cardiology (www.acc.org), the American Heart Association (www.americanheart.org), and the European Society of Cardiology (www.escardio.org). Single and bulk reprints of both the online full-text guidelines and the published executive summary (published in the August 15, 2006, issues of Circulation and the Journal of the American College of Cardiology and the August 16, 2006, issue of the European Heart Journal) are available from Oxford University Press by contacting Special Sales (special.sales{at}oxfordjournals.org), Journals Division, Oxford University Press, Great Clarendon Street, Oxford, OX2 6DP, UK. Phone +44 (0) 1865 353827, Fax +44 (0) 1865 353774, Work Mobile +44 07841322925. Single copies of the executive summary and the full-text guidelines are also available by calling 800-253-4636 or writing to the American College of Cardiology Foundation, Resource Center, at 9111 Old Georgetown Road, Bethesda, MD 20814-1699. To purchase bulk reprints, fax 212-633-3820 or e-mail reprints{at}elsevier.com. To purchase Circulation reprints: Up to 999 copies, call 800-611-6083 (US only) or fax 413-665-2671; 1000 or more copies, call 410-528-4121, fax 410-528-4264, or e-mail kelle.ramsay{at}wolterskluwer.com.

  • Permissions: Multiple copies, modification, alteration, enhancement, and/or distribution of this document are not permitted without the express permission of the American Heart Association or the European Society of Cardiology. Please direct requests to copyright.permissions{at}heart.org or journals.permissions{at}oxfordjournals.org.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 47a.
  49. 47b.
  50. 47c.
  51. 48.
  52. 49.
  53. 50.
  54. 51.
  55. 52.
  56. 53.
  57. 54.
  58. 55.
  59. 56.
  60. 57.
  61. 58.
  62. 59.
  63. 60.
  64. 61.
  65. 62.
  66. 63.
  67. 64.
  68. 65.
  69. 66.
  70. 67.
  71. 68.
  72. 69.
  73. 70.
  74. 71.
  75. 72.
  76. 73.
  77. 74.
  78. 75.
  79. 76.
  80. 77.
  81. 78.
  82. 79.
  83. 80.
  84. 81.
  85. 82.
  86. 83.
  87. 84.
  88. 85.
  89. 86.
  90. 87.
  91. 88.
  92. 89.
  93. 90.
  94. 91.
  95. 92.
  96. 93.
  97. 94.
  98. 95.
  99. 96.
  100. 97.
  101. 98.
  102. 99.
  103. 100.
  104. 101.
  105. 102.
  106. 103.
  107. 104.
  108. 105.
  109. 106.
  110. 107.
  111. 108.
  112. 109.
  113. 110.
  114. 111.
  115. 112.
  116. 113.
  117. 114.
  118. 115.
  119. 116.
  120. 117.
  121. 118.
  122. 119.
  123. 120.
  124. 121.
  125. 122.
  126. 123.
  127. 124.
  128. 125.
  129. 126.
  130. 127.
  131. 128.
  132. 129.
  133. 130.
  134. 131.
  135. 132.
  136. 133.
  137. 134.
  138. 135.
  139. 136.
  140. 137.
  141. 138.
  142. 139.
  143. 140.
  144. 141.
  145. 142.
  146. 143.
  147. 144.
  148. 145.
  149. 146.
  150. 147.
  151. 148.
  152. 149.
  153. 150.
  154. 151.
  155. 152.
  156. 153.
  157. 154.
  158. 155.
  159. 156.
  160. 157.
  161. 158.
  162. 159.
  163. 160.
  164. 161.
  165. 162.
  166. 163.
  167. 164.
  168. 165.
  169. 166.
  170. 167.
  171. 168.
  172. 169.
  173. 170.
  174. 171.
  175. 172.
  176. 173.
  177. 174.
  178. 175.
  179. 176.
  180. 177.
  181. 178.
  182. 179.
  183. 180.
  184. 181.
  185. 182.
  186. 183.
  187. 184.
  188. 185.
  189. 186.
  190. 187.
  191. 188.
  192. 189.
  193. 190.
  194. 191.
  195. 192.
  196. 193.
  197. 194.
  198. 195.
  199. 196.
  200. 197.
  201. 198.
  202. 199.
  203. 200.
  204. 201.
  205. 202.
  206. 203.
  207. 204.
  208. 205.
  209. 206.
  210. 207.
  211. 208.
  212. 209.
  213. 210.
  214. 211.
  215. 212.
  216. 213.
  217. 214.
  218. 215.
  219. 216.
  220. 217.
  221. 218.
  222. 219.
  223. 220.
  224. 221.
  225. 222.
  226. 223.
  227. 224.
  228. 225.
  229. 226.
  230. 227.
  231. 228.
  232. 229.
  233. 230.
  234. 231.
  235. 232.
  236. 233.
  237. 234.
  238. 235.
  239. 236.
  240. 237.
  241. 238.
  242. 239.