Antiarrhythmic Drug Therapy for Rhythm Control in Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia and affects over 33 million people worldwide. AF is associated with stroke and systemic thromboembolism, unpleasant symptoms and reduced quality of life, heart failure, and increased mortality, and treatment of AF and its complications are associated with significant cost. Antiarrhythmic drugs (AADs) can suppress AF, allowing long-term maintenance of sinus rhythm, and have the potential to relieve symptoms and reverse or prevent adverse effects associated with AF. However, large randomized controlled studies evaluating use of AADs have not demonstrated a clear benefit to maintaining sinus rhythm, and AADs often have significant limitations, including a modest rate of overall success at maintaining sinus rhythm, frequent side effects, and potentially life-threatening toxicities. Although some of the currently available AADs have been available for almost 100 years, better tolerated and more efficacious AADs have recently been developed both for long-term maintenance of sinus rhythm and for chemical cardioversion of AF to sinus rhythm. Advances in automated AF detection with cardiac implantable electronic devices have suggested that AADs might be useful for suppressing AF to allow safe discontinuation of anticoagulation in select patients who are in sinus rhythm for prolonged periods of time. AADs may also have synergistic effects with catheter ablation of AF. This review summarizes the pharmacology and clinical use of currently available AADs for treatment of AF and discusses novel AADs and future directions for rhythm control in AF.

Introduction

Atrial fibrillation (AF) currently affects over 33 million people worldwide. It is the most common sustained cardiac arrhythmia, ¹ and its incidence and prevalence are increasing in parallel with the aging of the world’s population. ² Atrial fibrillation is associated with unpleasant symptoms, poor quality of life, ^3,4 reduced left ventricular (LV) function and symptomatic heart failure, ^5,6 stroke and thromboembolism, ^7,8 and increased mortality, independent of other cardiovascular risk factors. ⁹ Treatment of AF and its associated complications are associated with significant health-care expenditures in both the United States and Europe. ^{10 -12}

Antiarrhythmic drugs (AADs) used for suppression of AF and maintenance of sinus rhythm (SR; rhythm control strategy) have been available for over a century ^13,14 and are currently a critical tool in the successful treatment of AF. However, widespread use of AADs is limited by an overall modest rate of success, side effects, proarrhythmia, and long-term drug toxicities. Although randomized trials evaluating use of AADs to maintain SR have not demonstrated a clear mortality benefit or decreased thromboembolic risk associated with a rhythm control strategy, ^{15 -20} AADs have been used effectively to reduce the severity, frequency, and duration of AF symptoms. Recent data from large anticoagulation trials have also suggested that a low burden of AF may be associated with a reduction in thromboembolism and mortality. ^21,22 This review summarizes the pharmacology and clinical use of currently available and novel AADs used in the treatment of AF.

Using AADs to Maintain Sinus Rhythm

Despite the observation that the presence of AF is associated with increased thromboembolism ^7,8 and mortality, ⁹ large randomized controlled trials comparing rate control (allowing the patient to remain in AF, but preventing very rapid ventricular rates) and use of AADs for rhythm control have not demonstrated a reduction in thromboembolic complications or mortality associated with use of AADs to maintain SR. ^{15 -20} Additionally, these studies found that a rhythm control strategy was associated with higher rates of adverse events, including drug side effects, hospitalization, and stroke. The lack of mortality benefit associated with rhythm control of AF in these studies, however, may be related to frequent discontinuation of anticoagulation in patients who appeared to maintain SR, increasing the risk of stroke and thromboembolism. ²³ As no AADs are 100% effective in suppressing AF, many of the patients on AADs who had strokes may have had subclinical AF which was not detected. Nonetheless, the American College of Cardiology (ACC), American Heart Association (AHA), Heart Rhythm Society (HRS), and European Society of Cardiology (ESC) guidelines recommend antiarrhythmic therapy primarily as a means to reduce symptoms attributable to AF and not as a way of reducing thromboembolic risk or mortality. ^24,25 It is also reasonable to consider rhythm control in patients <60 years regardless of symptoms or adequate rate control in AF; since these patients were not well represented in studies comparing rhythm and rate control, chronic AF may render future attempts at rate control more difficult, and the long-term consequences of AF are unclear.

Currently Available AADs

Multiple AADs (described in detail subsequently) are available for rhythm control in AF, with drug choice driven by patient characteristics/comorbidities and the risks of side effects and long-term drug toxicities. Historically, AADs have been classified into 1 of the 4 Vaughn-Williams classifications based on their effect on sodium channels (class I), β-receptors (class II), potassium channels (class III), or calcium channels (class IV). ²⁶ Class I drugs are further divided into class IA, class IB, and class IC based on drug affinity for sodium channels. This classification, however, oversimplifies the electrophysiological properties of AADs, as almost all available drugs have effects on multiple ion channels/currents, and some drugs, such as amiodarone, have characteristics of all 4 Vaughn-Williams classifications. As our understanding of the electrophysiological effects of AADs has increased, other, more complicated systems of AAD classification have been developed as well. ²⁷ The pharmacologic properties and electrophysiological effects of currently available AADs are summarized in Table 1. ^24,28

Table 1.

Characteristics of Antiarrhythmic Drugs Used for Rhythm Control of AF.

Drug	Route	Half-Life	Dose	Currents/Channels	Main Side Effects	Metabolism	Select Drug Interactions	Contraindications
Amiodarone	Orally/IV	∼50 d	Oral: Load 10 g over 7-10 days; maintenance dose of 200 mg daily IV: 150 mg bolus, then 1 mg/min infusion for 6 hours followed by 0.5 mg/min thereafter	I Na, I Kr, I Ks, I Kur, I CaL, I KAch, I to, α, β	Hepatic toxicity/cirrhosis, thyroid toxicity (hyper- and hypothyroidism), pulmonary toxicity, bradycardia, photosensitivity, blue–gray skin discoloration, tremor/neurologic side effects, alopecia, corneal deposits, optic neuritis, GI side effects	Hepatic via CYP3A4	Inhibits multiple CYP enzymes (increases concentration of warfarin, statins, cyclosporine, phenytoin, and carbamazepine) and P-gP (reduces clearance of digoxin)	Significant conduction disease, hepatic dysfunction
Dofetilide (United States only)	Orally	6-10 h	CrCl > 60 mL/min: 500 μg every 12 hours CrCl 40-60 mL/min: 250 μg every 12 hours CrCl 20-39 mL/min: 125 μg every 12 hours	I Kr	QT prolongation/torsades de pointes	80% renal, 20% hepatic via CYP3A4	Verapamil, thiazide diuretics, cimetidine, ketoconazole, trimethoprim, megestrol, and prochlorperazine. Does not inhibit or induce any CYP enzymes	QT prolongation, significant renal dysfunction
Disopyramide	Orally	6-8 h	Immediate release: 100-200 mg every 6 hours Extended release: 200-400 mg every 12 hours. Reduce dose in renal and hepatic dysfunction	I Kr, I Na, I KAch	Anticholinergic effects (constipation, dry eyes/mouth, visual disturbances, urinary retention), heart failure exacerbation, QT prolongation/torsades de pointes	Renal and hepatic via CYP3A4	Inhibitors/inducers of CYP3A4 (verapamil, diltiazem, macrolides, azoles, protease inhibitors, grapefruit juice, phenytoin, rifampin)	Urinary retention/prostatic hypertrophy, glaucoma, heart failure
Dronedarone	Orally	13-19 h	400 mg every 12 hours	I Na, I Kr, I Ks, I K1, I CaL, I KAch, α, β	Bradycardia, GI side effects	Gastrointestinal (first-pass metabolism), hepatic CYP3A4	Inhibits P-gP (reduces clearance of digoxin) and CYP2D6. Inhibits and metabolized by CYP3A4 (verapamil, diltiazem, macrolides, azoles, protease inhibitors, grapefruit juice, phenytoin, rifampin)	Bradycardia, heart failure, persistent AF, hepatic dysfunction
Flecainide	Orally	7-23 h	50-200 mg every 12 hours	I Na	QRS widening, atrial flutter with 1:1 AV conduction, unmasking of Brugada-type ECG, dizziness, headache, blurred vision, ventricular arrhythmias	Renal and hepatic via CYP2D6	Inhibitors of CYP2D6 (haloperidol, tricyclic antidepressants, narcotics, fluoxetine)	Heart failure, coronary artery disease, Brudaga syndrome, significant AV conduction disease
Ibutilide	IV	6 h	1 mg infusion over 10 minutes; may repeat once if needed 10 minutes after initial infusion ends	I Kr	QT prolongation, torsades de pointes, nausea	Hepatic via CYP3A4	Minimal	QT prolongation
Propofanone	Orally	9 h	Immediate release: 150-300 mg every 8 hours Extended release: 225-425 mg every 12 hours	I Na, β	QRS widening, atrial flutter with 1:1 AV conduction, unmasking of Brugada-type ECG, dizziness, headache, blurred vision, ventricular arrhythmias, metallic taste	Hepatic via CYP2D6	Inhibits P-gP (reduces clearance of digoxin), inhibits CYP2C9 (increases warfarin effect)	Heart failure, coronary artery disease, Brudaga syndrome, significant AV conduction disease
Ranolazine	Orally	7 h	500-1000 mg every 12 hours	Late I Na, I Kr, I Ks, I CaL	Increase in serum creatinine, bradycardia, QT prolongation, GI side effects, headache, dizziness	Renal and hepatic via CYP3A4	Inhibitors of CYP3A4 (verapamil, diltiazem, macrolides, azoles, protease inhibitors, grapefruit juice, phenytoin, rifampin), inhibitors of P-gP	Severe renal or hepatic dysfunction, QT prolongation
Sotalol	Orally	8 h	80-160 mg every 12 hours	I Kr, β	Bradycardia, QT prolongation/ torsades de pointes, bronchospasm	Renal	Minimal	Bradycardia, renal dysfunction
Quinidine	Orally	8 h	Quinidine gluconate: 324-648 md 3 times daily Quinidine sulfate: 200-600 mg 3 times daily	I Kr, I Ks, I to, I KATP, I CaL, I Na, I KAch	Diarrhea/GI side effects, thrombocytopenia, rash, cinchonism, QRS widening (high doses), QT prolongation/torsades de pointes (unrelated to dose)	30% renal, 70% hepatic via CYP3A4	Inhibits CYP2D6 (reduces clearance of haloperidol, tricyclic antidepressants, narcotics, fluoxetine) and P-gP (reduces clearance of digoxin)	QT prolongation
Vernakalent (Europe only)	IV	2 h	3 mg/kg infusion over 10 minutes; may repeat 2 mg/kg infusion over 10 minutes if needed 15 minutes after initial infusion ends	I Kur	Dysgeusia, cough, sneezing, hypotension, QT prolongation	Hepatic via CYP2D6	Inhibitors of CYP2D6	Heart failure

Abbreviations: AF, atrial fibrillation; AV, atrioventricular; P-gP, p-glycoprotein; CrCl, creatinine clearance; CYP, cytochrome P; ECG, electrocardiogram; GI, gastrointestinal; IV, intravenous; I _Kr, rapid component of the delayed rectifier potassium current; I _Na, sodium current; I _CaL, l-type calcium channels.

^aAdapted from January et al ²⁴ and Zimetbaum. ²⁸

Quinidine

Quinidine, an alkaloid derived from the bark of the cinchona plant (and a diastereomer of the antimalarial drug quinine ²⁹ ), was first discovered in the 19th century ³⁰ and is one of the oldest cardiac medications still in use today. Quinidine blocks sodium and potassium channels, with variable electrophysiologic properties depending on drug concentration. At low concentrations, it primarily blocks sodium current (I _Na) and the rapid component of the delayed rectifier potassium current (I _Kr). At high concentrations, however, it also blocks multiple other potassium currents (including I _Ks, I _KATP, and I _to) and l-type calcium channels (I _CaL). ²⁹ Quinidine’s sodium channel blocking effect exhibits “use dependence,” whereby it is more pronounced at fast heart rates and high concentrations, while its potassium channel blocking effect exhibits “reverse use dependence,” whereby it is most pronounced at slow heart rates and lower drug concentrations. In addition to its effect on potassium, sodium, and calcium channels, quinidine has additional vagolytic, α-blocking, and β-blocking properties. ²⁹ The majority of quinidine is metabolized in the liver via cytochrome P (CYP) 3A4, although some drug is renally eliminated. Quinidine is a potent inhibitor of CYP2D6 and P-glycoprotein (P-gP) and as a result reduces digoxin clearance. ²⁹

Quinidine is currently available in 2 formulations, quinidine gluconate and quinidine sulfate, both of which are taken 3 times per day. Both formulations are well absorbed with a half-life of approximately 8 hours. One beneficial and often overlooked feature of quinidine is that it can be used safely in patients with structural heart disease and renal dysfunction. Despite its efficacy in suppressing AF, however, use of quinidine has fallen out of favor due to concerns about side effects and safety and the availability of newer AADs. Current use of quinidine in the treatment of AF is therefore rare, and patients often experience difficulty obtaining the drug due to high costs and lack of availability. ³¹

Side effects are common, with diarrhea and other gastrointestinal symptoms being most frequent. Less common noncardiac side effects include thrombocytopenia, rash, and cinchonism (headache and tinnitus). Adverse cardiac side effects include severe QRS widening at high doses, and QT prolongation and torsades de pointes (TdP) independent of drug concentration. ²⁹

Concern about quinidine originated from a meta-analysis of 6 small trials that revealed increased mortality associated with quinidine use (mortality rate 2.9% vs 0.8% for quinidine and control, respectively). ³² A recent Cochrane review incorporating larger and more recent studies demonstrated that quinidine use was associated with a nonstatistically significant trend toward an increase in mortality (pooled odds ratio 2.26, 95% confidence interval [0.93-5.45], P = .07). ³³ Many of the older studies of quinidine were conducted before the important adverse interaction between quinidine and digoxin was appreciated, and this might account for some of the excess mortality observed early on. ³² More recent studies that evaluated the combination of quinidine and verapamil (which may suppress early after depolarizations that can trigger TdP ³⁴ ) demonstrated low rates of all-cause mortality. ^35,36 If tolerated, quinidine remains an effective option to treat AF, especially when other AADs have failed or are contraindicated due to patient comorbidities. ³⁷

Disopyramide

Disopyramide, which was first evaluated in the 1960s, ³⁸ shares many electrophysiological properties with quinidine, despite being chemically distinct. Disopyramide blocks both I _Na and I _Kr and prolongs both action potential duration and refractory period. ³⁹ As a result, similar to quinidine, it causes QT prolongation and QRS widening (at higher doses). Disopyramide is available in an immediate-release formulation, which is dosed 4 times per day, and a sustained-release formulation, which is taken twice daily. It has high oral bioavailability and a half-life of approximately 6 to 8 hours in patients with normal renal function. Drug is metabolized primarily by the kidneys although some drug is also metabolized by the liver via CYP3A4, ²⁹ and dose adjustment is necessary in patients with renal dysfunction and hepatic dysfunction. ³⁹

One of the main limitations of disopyramide is that it has strong anticholinergic effects, and due to the frequency with which it can induce urinary retention, its use is limited in men, especially those with prostatic hypertrophy. Other anticholinergic effects such as constipation, dry mouth, visual disturbances, and glaucoma can also occur. ²⁹ Additionally, due to strong negative inotropic effects, use of disopyramide is contraindicated in patients with heart failure and reduced LV systolic function. ⁴⁰ Hemodynamic effects in patients with structurally normal hearts, however, are minimal. Due to its negative inotropic properties, disopyramide may be a reasonable AAD to use in patients with AF who also have hypertrophic cardiomyopathy and outflow tract obstruction, ⁴¹ although strong data supporting its use in this situation are lacking. Like quinidine, use of disopyramide has been associated with increased mortality, ³³ although study of disopyramide for suppression of AF in a clinical trial settings is lacking in comparison to other AADs. Overall, use of disopyramide in the treatment of AF is exceedingly rare.

Sotalol

Sotalol, a nonselective β-blocker that exerts its antiarrhythmic effects through inhibition of the delayed rectifier potassium current I _Kr, was developed in the 1960s. ⁴² It is administered as an equal combination of l- and d-enantiomers, and although both enantiomers have equivalent potassium channel blocking properties, the l-enantiomer is a much stronger β-antagonist. ⁴² d-Sotalol, a pure I _Kr blocker associated with increased mortality in patients with prior myocardial infarction (MI) and LV dysfunction potentially due to lack of β-blockade and proarrhythmia, is not clinically available. ⁴³

Sotalol is easily absorbed, has a very high bioavailability, and is not protein bound. It is administered twice daily and has a half-life of approximately 8 hours. It is excreted unchanged in the urine, with dose adjustments required in patients with impaired renal function, and does not undergo any hepatic metabolism. ⁴⁴ Sotalol causes dose-dependent prolongation of action potential duration, atrial and atrioventricular (AV) nodal refractory periods, and the QT interval in a reverse use-dependent manner. In general, despite its β-blocking properties, sotalol is not a negative inotrope due to a concomitant increase in contractility due to action potential duration prolongation; however, in some cases, it can exacerbate heart failure. ⁴⁴ As sotalol has β-blocking properties, in some patients, it can exacerbate reactive airway disease. Overall, however, it has few other side effects. Drug interactions are predominantly related to potentially dangerous interactions with other QT-prolonging medications that can precipitate TdP. The risk of adverse events such as TdP is dose-/QT interval dependent and may be increased due to the bradycardic effects of the drug. When used properly and at usual doses, the risk of TdP is <1%. ⁴⁵

Drug therapy is usually initiated at a dose of 80 mg twice daily. In patients with a creatinine clearance between 40 and 60 mL/min, the dose is reduced to once daily. Use is not recommended for patients with creatinine clearance <40 mL/min. After allowing 3 days for the drug to reach steady state, an electrocardiogram (ECG) is obtained, and if the QT interval (<500 ms) and heart rate allow, the dose can be uptitrated as needed (increasing by 40 mg increments) until reaching a maximum dose which is usually 160 mg twice daily. The drug manufacturer recommends that sotalol be initiated in the hospital, ⁴⁵ but data have suggested that outpatient drug initiation can be safely performed in select patients in SR (see subsequent sections).

In general, sotalol is a modestly effective AAD. In the Canadian Trial of Atrial Fibrillation (CTAF) amiodarone was superior to sotalol or propofanone (at a mean of 16 months, 65% of patients treated with amiodarone and 37% of patients treated with sotalol or propofanone remained free of AF recurrence). ⁴⁶ Similarly, in the Sotalol Amiodarone Atrial Fibrillation Efficacy Trial (SAFE-T) comparing amiodarone, sotalol, and placebo in patients with persistent AF, sotalol was superior to placebo but inferior to amiodarone in maintaining SR (median time to recurrence 487 days, 74 days, and 6 days for amiodarone, sotalol, and placebo, respectively, P < .001). In the subgroup of patients with ischemic heart disease, however, amiodarone and sotalol were equally efficacious in preventing AF recurrence (median time to recurrence 569 days and 428 days for amiodarone and sotalol, respectively, P = .53). ⁴⁷ In meta-analyses, efficacy of sotalol is similar to efficacy of most AADs other than amiodarone. ³³

Flecainide

Flecainide was developed in the 1970s. ⁴⁸ It is administered twice daily (at doses between 50 and 200 mg), has a bioavailability of approximately 90%, and its half-life ranges from 7 to 23 hours. ⁴⁹ Metabolism takes place primarily in the liver via CYP2D6 and CYP1A2 with further renal excretion of metabolites, and dose adjustment is therefore necessary in patients with advanced renal failure. Flecainide exerts its electrophysiologic effects primarily by blocking the rapid sodium current responsible for phase 0 of the cardiac action potential (I _Na). Flecainide kinetics are strongly influenced by heart rate; because it has very slow dissociation from ion channels at fast heart rates, drug binding to sodium channels during tachycardia is enhanced and drug effect is therefore magnified. Compared to other sodium channel blocking, drugs with intermediate dissociation kinetics, such as quinidine and disopyramide, flecainide’s very slow dissociation kinetics, can produce significant QRS widening at normal heart rates. If this is observed, it should prompt drug discontinuation. ⁴⁸

The Cardiac Arrhythmia Suppression Trial (CAST) demonstrated an increase in mortality associated with flecainide use in patients with MI and premature ventricular beats. ⁵⁰ Flecainide also has negative inotropic effects likely related to both its sodium channel–blocking properties and the fact that drug clearance is significantly reduced in patients with heart failure. ⁴⁸ Flecainide is therefore not appropriate for use in patients with structural heart disease and prior MI. Whether flecainide can be safely used in patients with mild/stable coronary artery disease without MI is unclear, but such use is not currently recommended. ²⁴

The major cardiovascular toxicity of flecainide when used in the treatment of AF is conversion of AF to atrial flutter with rapid ventricular rates due to 1:1 AV conduction. This can occur due to slowing and organization of atrial arrhythmias to the point that the AV node is able to conduct 1:1 rather than at intervals of 2:1 or higher. Thus, flecainide use may result in atrial flutter with very rapid ventricular rates > 200 bpm which can cause syncope, or even induce ventricular tachycardia or ventricular fibrillation. Therefore, flecainide should always be used with another drug which slows AV nodal conduction (such as a β-blocker or calcium channel blocker), unless AV nodal conduction is already significantly impaired. ²⁴ A hybrid strategy of using flecainide after ablation of atrial flutter may also be helpful in patients who have AF and flecainide-induced atrial flutter. ^51,52 In select patients who wish to avoid taking medication daily, a “pill-in-the-pocket” strategy, where a high dose of flecainide (200-300 mg) is taken at the onset of an episode of AF, has also been validated, ⁵³ but this should only be used if the patient has demonstrated in a monitored setting that he or she can take such high doses of flecainide without significant adverse effects. ²⁴ Other common side effects of flecainide include dizziness, visual disturbances, nausea, dyspnea, and other neurologic symptoms.

Propafenone

Propafenone was initially approved in Europe in 1977 and in the United States in 1989 for treatment of ventricular arrhythmias. It is primarily a sodium channel–blocking drug and shares many properties with flecainide. In addition to its sodium channel–blocking properties, propafenone also blocks potassium channels (including I _Kr, I _to, and inward rectifying potassium channels), calcium channels, and β-adrenergic receptors. Dissociation kinetics are slightly faster than flecaninde ⁵⁴ but are still slower than quinidine or disopyramide.

Standard dosing of propofanone is every 8 hours, although an extended-release preparation can be administered twice daily. Propofanone is well absorbed, although via extensive first-pass hepatic metabolism bioavailability may be as low as ∼10%. The drug is highly protein bound. Metabolism is primarily by hepatic oxidation via CYP2D6, and dose reduction is required in patients with hepatic dysfunction. ⁵⁴ Genetic variations in metabolism are responsible for variations in drug effect and systemic side effects. ^55,56

Like flecainide, propofanone should not be used in patients with structural heart disease and/or coronary artery disease/prior MI, and 1:1 atrial flutter can result from administration during AF. Although propofanone has some intrinsic β-blocking properties, given the risk of rapid atrial flutter, it should usually be used with an additional AV nodal-blocking agent.

Amiodarone

Amiodarone, an iodainated benzofuran derivate, is the most common AAD prescribed for AF, ⁵⁷ despite the fact that it has never been approved for this indication. It was developed in the 1960s and was initially used as an antianginal agent until the 1970s when its antiarrhythmic properties were discovered. ^58,59 Amiodarone exerts a multitude of electrophysiological effects on cardiac tissue. It blocks I _Na, I _Kr, I _Ks, I _Kur, I _CaL, acetylcholine sensitive potassium channels (I _KAch), I _to, α-receptors, and β-receptors ⁶⁰ and therefore has a wide range of electrophysiological effects on all types of cardiac tissues. Amiodarone has an oral bioavailability of approximately 30% to 50% and is highly lipophilic with a very large volume of distribution and an extremely long and highly variable elimination half-life on the order of weeks to months. ^29,61 The electrophysiological effects of oral amiodarone therefore are usually delayed for days (with full effects not expected for weeks) after initiation, although effects can be seen faster with intravenous administration. ⁶¹ Drug levels do not correlate with clinical efficacy. ⁶¹ Amiodarone is hepatically metabolized by CYP3A4 to desethyl-amiodarone which shares similar antiarrhythmic properties with amiodarone. ⁶⁰ Dosage should be adjusted in patients with severe hepatic dysfunction, but dose reduction is not required in patients with renal dysfunction or patients on dialysis.

Due to its prolonged half-life, amiodarone is usually initiated at a total daily dose of 600 to 800 mg (divided into 2-3 doses) to achieve a loading dose of 10 g over a period of 7 to 10 days. The dose is then reduced to maintenance of 200 mg daily, although a dose of 400 mg daily may be used for a few weeks after initial loading in select patients. ²⁴ Of note, this dose is lower than what would typically be used for treatment of ventricular arrhythmias. Slower loading may minimize adverse effects such as gastrointestinal distress and suppression of sinus node and AV nodal conduction. ⁶² Administration with food also increases bioavailability ⁶³ and reduces gastrointestinal side effects. Although some patients may convert from AF to SR during loading, many patients do not, and this has no effect on long-term drug efficacy once appropriate drug levels have been achieved and SR is restored. ⁶⁴

In studies comparing the relative efficacy of AADs in suppressing AF, amiodarone is consistently the most effective AAD available ^46,47,65 although a multitude of side effects limit its widespread use. Amiodarone contains a significant amount of iodine, is structurally similar to the thyroid hormones triiodothyronine (T₃) and thyroxine (T₄), and inhibits both peripheral conversion of T₄ to T₃ and entry of T₃ and T₄ into cells. ⁶⁶ Most patients develop some perturbation in thyroid function tests including an elevation in T₄ and thyroid-stimulating hormone and a reduction in T₃ which are not indicative of hyperthyroidism and do not require treatment. ⁶⁶ Amiodarone can, however, induce both hypo- and hyperthyroidism, and up to one-quarter of patients may develop clinically relevant thyroid dysfunction while on the drug. ⁶⁶ Hypothyroidism is more common in areas with high dietary iodine intake, while hyperthyroidism predominates in areas with iodine deficiency. ⁶⁶ Treatment of hyperthyroidism with amiodarone discontinuation, methimazole, propylthiouracil, steroids, or thyroidectomy in severe cases where continued use of amiodarone is necessary may be required. ⁶⁶ Recurrence of AF while on amiodarone may be a sign of hyperthyroidism even if there are no other clinical signs pointing to thyroid derangement. Drug discontinuation and/or thyroid replacement are the main treatments for amiodarone-induced hypothyroidism. ⁶⁶

Nonspecific gastroenterological symptoms such as nausea are common and may resolve with dose reduction. ⁶¹ Severe hepatotoxicity is rare (<3%), ⁶¹ but patients will frequently have minor abnormalities in transaminases, which can mimic nonalcoholic fatty liver disease. ⁶⁷ This low-level inflammation can, in rare cases, result in cirrhosis if unrecognized and the drug is continued. ⁶¹ Acute amiodarone-induced hepatitis, often after parenteral administration, has also been described. ⁶⁸ Treatment of amiodarone-induced hepatitis is supportive, with discontinuation of the drug.

Pulmonary toxicity occurs in approximately 2% of patients treated with amiodarone. ⁶¹ Patient age and duration of therapy are significant risk factors. ⁶⁹ Preexisting pulmonary disease increases the risk of amiodarone-induced pulmonary toxicity but does not increase total mortality or pulmonary mortality associated with use of amiodarone, and the drug can be used cautiously in patients with mild lung disease. ⁷⁰ Pulmonary toxicity may present at any time point during amiodarone therapy and usually presents with cough with later development of dyspnea and fever. ⁷¹ Chest X-ray reveals diffuse infiltrates, consolidations, and pleural thickening. ⁷² Pulmonary function testing usually reveals evidence of restrictive physiology and impaired diffusing capacity, and computed tomography of the chest can reveal evidence of fibrosis and ground glass opacities. ^71,72 Findings on bronchoscopy are nonspecific for amiodarone toxicity but reveal an increase in inflammatory cells with structurally abnormal “foamy” macrophages. ⁷² Lung biopsy is usually avoided due to a high incidence of postprocedure acute respiratory distress syndrome and high mortality. ⁷² Treatment of amiodarone-induced pulmonary toxicity includes drug discontinuation and treatment with steroids. ⁷¹

Cardiac side effects include bradycardia, which is often dose related. ⁷³ Prolongation of the QT interval occurs in almost all patients, although, despite this, rates of TdP (<0.5%) are significantly lower than for other drugs that prolong the QT interval such as sotalol and dofetilide, likely due to amiodarone’s simultaneous effects on other ion channels. Amiodarone can also cause photosensitivity and rarely causes blue skin discoloration or alopecia. ⁶² Neurologic side effects including neuropathy, tremor, and cognitive impairment are rare. ⁷⁴ Almost all patients develop corneal deposits, which have no clinical significance, but optic neuropathy can rarely develop. ⁷⁵

Amiodarone has multiple clinically significant drug interactions. Amiodarone inhibits CYP3A4, CYP2C9, and P-gP and can increase the drug effect of coadministered warfarin, digoxin, statins (especially simvastatin), and multiple other drugs.

Dofetilide

Dofetilide, one of the newest AADs available for treatment of AF, was approved for use in the United States in the year 2000. It is not approved or available in Europe. Dofetilide blocks the rapid component of the delayed rectifier potassium current (I _Kr) and unlike most other AADs has minimal effects on other ion channels. It therefore prolongs the QT interval, and its main adverse event is TdP. Dofetilide exhibits reverse use dependence, whereby the effect of I _Kr blockade is diminished as the heart rate increases, and the risk of drug toxicity is amplified at low heart rates. ⁷⁶ Drug potency is also affected by extracellular potassium concentration, and hypokalemia and hyperkalemia can increase or decrease drug potency, respectively. ⁷⁷

Dofetilide has high oral bioavailability of >90%. It is taken twice daily, and the starting dose (between 125 and 500 µg) is carefully selected based on renal function to avoid significant QT prolongation and Tdp. During the initial drug loading period, the dose of dofetilide dose may be further adjusted based on the change in QT interval after each dose during required inpatient monitoring (see subsequently). Dofetilide’s half-life is 6 to 10 hours, and approximately 80% of the drug is excreted unchanged in the urine via cationic secretion, while the remainder is metabolized in the liver by CYP3A4 prior to renal excretion. ^76,78 Dofetilide does not inhibit or induce any CYP450 enzymes, ⁷⁸ although there are still many important interactions with common drugs. Coadministration of other drugs that can prolong the QT interval should be avoided. ⁷⁹ Verapmil, cimetidine, ketoconazole, trimethoprim, and hydrochlorothiazide can also significantly increase dofetilide levels and cannot be coadministered. ^78,80

Although dofetilide has demonstrated efficacy and safety when used in patients with reduced LV systolic function and/or coronary artery disease, ^81,82 it remains relatively underutilized, primarily due to the requirement that drug initiation take place during a 3-day inpatient hospitalization. This was mandated because most adverse proarrhythmic events occurred in the first 3 days after drug initiation. Additionally, physicians and pharmacies have to be specially certified to prescribe and dispense the drug, respectively.

In the Symptomatic Atrial Fibrillation Investigative Research on Dofetilide (SAFIRE-D) study, dofetilide was effective at converting AF to SR with a dose-dependent response of 6% to 30%, and 91% of patients converted to SR within 36 hours. Between 40% and 60% of patients on dofetilide remained in SR at 1 year which was significantly better than placebo (25% at 1 year, P = .001). ⁸³ The Danish Investigations of Arrhythmia and Mortality ON Dofetilide (DIAMOND) Study ⁸² randomized 1518 patients with LV dysfunction and symptomatic heart failure to dofetilide or placebo and found no difference in total mortality at 18 months (41% vs 42% in dofetilide and placebo groups, respectively). Dofetilide has also demonstrated safety when used in patients with MI within 1 week and reduced LV function. ⁸¹ Overall, dofetilide has not been associated with an increased mortality risk. ³³ Rates of TdP ranged from ∼1% in the SAPHIRE-D to ∼3% in the DIAMOND trials. Women, patients with severe heart failure, recent MI, and prolonged baseline QT are at increased risk of TdP. Dofetilide has only been studied and approved for use in patients with persistent AF, and efficacy in patients with paroxysmal AF is less clear.

Dronedarone

Although amiodarone is highly effective in maintaining SR, as noted earlier, its myriad toxicities frequently limit its use. Dronedarone, a noniodinated benzofuran derivative with a structure similar to amiodarone (with the addition of a methyl–sulfonyl group and without the iodine moieties), was developed to overcome these limiting toxicities by reducing its interaction with thyroid physiology, decreasing lipophilicity, ⁸⁴ and shortening plasma half-life (∼13-19 hours). ⁸⁵ Similar to amiodarone, dronedarone blocks a wide array of cardiac ion channels, including I _Kr, I _Ks, I _K1, I _KAch, I _Na, I _CaL, α-receptors, and β-receptors. ^84,86 Dronedarone is administered at a fixed dose of 400 mg twice daily and is well absorbed, but extensive first-pass metabolism limits bioavailability. Administration with food can increase bioavailabilty up to 4-fold. ⁸⁴ Dronedarone pharmacokinetics are not significantly influenced by age, gender, weight, or renal function. Serum creatinine usually increases due to inhibition of creatinine secretion and does not reflect impaired renal function. ⁸⁴ Dose adjustments are only recommended in patients with severe hepatic dysfunction. ⁸⁴

Dronedarone is metabolized by and inhibits CYP3A4 and P-gp and therefore is associated with many significant drug–drug interactions. Coadministration of other strong inhibitors of CYP3A4 such as azole antifungals, grapefruit juice, and cyclosporine is contraindicated. ⁸⁴ Dronedarone can also significantly increase levels of statin medications and digoxin, resulting in increased toxicity. Unlike amiodarone, dronedarone has no interactions with warfarin ⁸⁴ and has no significant effect on thyroid function.

In the A Placebo-Controlled, Double-Blind, Parallel Arm Trial to Assess the Efficacy of Dronedarone 400 mg bid for the Prevention of Cardiovascular Hospitalization or Death from Any Cause in patiENts with Atrial Fibrillation/Atrial Flutter (ATHENA) trial, compared to placebo, dronedarone was associated with a reduction in the composite outcome of hospitalization due to cardiovascular events or death (hazard ratio (HR) 0.76, P < .001). ⁸⁷ The Randomized, Double-Blind TrIal to Eval-uate the Efficacy and Safety of DrOnedarone [400 mg bid] Versus AmiodaroNe [600 mg qd for 28 daYS, then 200 mg qd Thereafter] for at Least 6 mOnths for the Maintenance of Sinus Rhythm in Patients with AF (DYONYSOS) study compared amiodarone to dronedarone in 504 patients with persistent AF and found that amiodarone was significantly more effective than dronedarone at preventing recurrence of AF (HR 1.59, P < .0001) but was associated with significantly more adverse thyroid, neurological, ocular, and dermatological side effects at 1 year. ⁸⁸

Despite these favorable initial results, the Permanent Atrial Fibrillation Outcome Study Using Dronedarone on Top of Standard Therapy (PALLAS) study, which randomized 3236 patients with persistent AF and cardiovascular risk factors to placebo or dronedarone used as a rate-controlling agent, was stopped prematurely due to safety concerns: The composite end point of stroke, systemic embolism, MI, or cardiovascular death was significantly more frequent in dronedarone-treated patients (HR 2.3, P = .002). Compared to placebo, the individual outcomes of cardiovascular death (HR 2.1, P = .046), arrhythmic death (HR 3.3, P = .03), stroke (HR 2.3, P = .02), and heart failure hospitalizations (HR 1.8, P = .02) were also significantly increased in the dronedarone group. ⁸⁹

The divergent results from the PALLAS and ATHENA studies could be related to the fact that ATHENA allowed enrollment of patients with paroxysmal AF and higher rates of concomitant digoxin use in PALLAS. However, another study (Antiarrhythmic Trial with Dronedarone in Moderate to Severe CHF Evaluating Morbidity Decrease [ANDROMEDA]) which randomized patients with LV systolic dysfunction (LV ejection fraction ≤ 35%) and New York Heart Association (NYHA) class III or IV symptoms within the prior month to dronedarone or placebo found a similar increase in mortality in dronedarone-treated patients (HR 2.1, P = .03) which was frequently due to worsening heart failure. ⁹⁰ Although approximately 20% of patients in ATHENA had NYHA class II-III heart failure, and subgroup analysis did not suggest an increase in mortality or adverse events in this subgroup of patients, the severity of heart failure in ATHENA was lower than in ANDROMEDA, and this may explain the disparate results. ⁸⁷ Nonetheless, based on increased mortality, dronedarone should be avoided in patients with congestive heart failure, and it should not be used as a rate-control agent in patients with long-standing persistent or permanent AF.

Ranolazine

Ranolazine blocks sodium, potassium, and calcium currents and is approved as an antianginal agent, although evidence has emerged regarding its antiarrhythmic properties for both atrial and ventricular arrhythmias. Ranolazine inhibits peak and late I _Na, the rapidly and slowly delayed rectifier potassium currents (I _Kr and I _Ks, respectively), and is a relatively weak inhibitor of l-type calcium channels (I _CaL). ⁹¹ Ranolazine is administered twice daily, has a half-life of 7 hours, and is predominantly hepatically metabolized by CYP3A4 prior to renal excretion. Dose adjustment is therefore reasonable in both renal and hepatic failure, although the drug manufacturer does not require such adjustment. ⁹²

Data supporting use of ranolazine as an AAD for AF are limited. In a substudy of patients with non-ST-segment elevation acute coronary syndrome in the Metabolic Efficiency With Ranolazine for Less Ischemia in Non−ST-Elevation Acute Coronary Syndromes (MERLIN)-TIMI 36 trial, compared to placebo, treatment with ranolazine resulted in a reduction in AF events over 1-year follow-up (2.9% vs 4.1%, P = .01), and patients with baseline paroxysmal AF had an overall lower burden of AF (4.4% vs 16.1% for ranolazine and placebo, respectively, P = .015). ⁹³ The PHase 2, Proof of Concept, Randomized, Placebo-Controlled, PArallel Study to Evaluate the Effect of Ranolazine and Dronedarone When Given Alone and in CoMbination On Atrial FibrillatioN Burden in Subjects with ParoxYsmal Atrial Fibrillation (HARMONY) study, which compared placebo, ranolazine, dronedarone, and the combination of ranolazine and dronedarone, found that ranolazine alone had no effect on pacemaker detected AF burden over 12 weeks of follow-up, but the combination of ranolazine and dronedarone was superior to placebo in reducing AF burden. ⁹⁴ In the Ranolazine in Atrial Fibrillation Following An ELectricaL CardiOversion (RAFFAELLO) study, administration of ranolazine after cardioversion from AF to SR was ineffective at prolonging time to AF recurrence, although there was a trend toward reduced AF recurrences in patients treated with high-dose ranolazine. ⁹⁵ Based on this limited evidence, ranolazine should not be used as a primary rhythm control agent in AF, and current guidelines do not recommend its use as an AAD for AF. Given the synergistic effects of ranolazine with dronedarone ⁹⁶ or amiodarone ⁹⁷ in suppressing AF, ranolazine may have a role as a second AAD in select patients.

Ranolazine prolongs the QT interval in a dose-dependent manner predominantly via its effect on I _Kr, but its inhibition of the late sodium current is protective against TdP. ^98,99 Although a significant increase in TdP has not been observed, caution should still be exercised when using ranolazine in patients with significant QT prolongation or in patients who are taking other QT-prolonging drugs. ⁹² The most common side effects include headache, dizziness, nausea, and constipation.

Ibutilide

Ibutilide, a methanesulfonamide derivative with similar structure to sotalol, is an intravenous I _Kr blocker that is used only for acute termination of AF. ¹⁰⁰ In addition to its potassium channel–blocking effects, ibutilide activates late inward sodium current and, unlike sotalol, does not exhibit significant reverse use dependence. ¹⁰⁰ It has minimal hemodynamic effects and can be used safely in patients with structural heart disease and prior MI. Despite its efficacy in terminating both AF and atrial flutter, due to its blockade of I _Kr and facilitation of late inward sodium current, ibutilide prolongs action potential duration and the QT interval and use has been associated with an increased risk of TdP (overall rate ∼4%). ¹⁰¹ Patients must therefore be monitored closely for at least 4 hours after the infusion has ended and QTc has returned to baseline. ^{102 -104} Noncardiac side effects other than nausea are rare. ¹⁰¹

Ibutilide is usually administered as a 1-mg infusion over 10 minutes, with a repeat 1-mg infusion 10 minutes after the first infusion completes if necessary. Patients weighing <60 kg should receive a dose of 0.01 mg/kg. Potassium and magnesium should be replete prior to administration, and the drug should be avoided in patients with baseline QT prolongation or use of other drugs that can prolong the QT. ¹⁰⁰ Ibutilide has a large volume of distribution and an average half-life of 6 hours. It undergoes hepatic metabolism via CYP3A4 and renal excretion, and although dose reductions are not required in either hepatic or renal dysfunction, patients with liver disease may metabolize ibutilide more slowly and require a prolonged period of postinfusion monitoring. ¹⁰⁰

Vernakalent

Vernakalent is a new AAD that blocks multiple ion channels but has the unique benefit of only affecting atrial tissue. Its major antiarrhythmic effect in AF is inhibition of the ultra-rapid delayed rectifying potassium channel (I _Kur) which is highly selective for atrial tissue. It also has weak effects on I _to and rate-dependent effects on I _Na. ¹⁰⁵ Vernakalent is available only as an intravenous infusion, and it is metabolized by CYP2D6 ¹⁰⁵ with an elimination half-life of approximately 2 hours. Multiple studies have demonstrated that intravenous administration of vernakalent is able to rapidly terminate AF episodes of less than 7 days duration in about 50% of patients without significant proarrhythmia. ^{106 -108} Common side effects include hypotension, bradycardia, dysgeusia, sneezing, and cough. ¹⁰⁶ Vernakalent has been approved for use only in Europe for the acute termination of AF. Prior to approval in the United State, the Federal Drug Administration requested that an additional phase III study be performed, and during this study a patient died during drug infusion. The study was canceled, and for this reason, the drug has not been approved in the United States. ¹⁰⁹ Development of a sustained-release oral formulation has also been discontinued.

Antiarrhythmic Drugs on the Horizon

Vanoxerine

Vanoxerine is a dopamine transporter antagonist that was initially developed for treatment of depression and Parkinson disease. In addition to its effects on dopamine receptors, it blocks I _Kr, I _Na, and I _CaL in a use-dependent manner. It has no effect on dispersion of repolarization and therefore is thought to have a reduced risk of proarrhythmia. ¹¹⁰ Early animal studies demonstrated efficacy in terminating AF and atrial flutter. ^111,112 There are limited data on use in humans, although in a small placebo-controlled randomized trial of 104 patients with AF lasting <7 days duration, administration of oral vanoxerine was associated with conversion to SR in a dose-dependent manner. Side effects were generally mild and included QT prolongation, bradycardia, gastrointestinal symptoms (nausea and vomiting), and neurologic symptoms (headache and dizziness). ¹¹³ A larger study evaluating use of vanoxerine for conversion of AF and atrial flutter to SR is currently underway (clinicaltrials.gov NCT02454283).

Other Novel AADs Under Development

Many novel AADs have failed to reach market, and multiple novel AADs for the treatment of AF are currently under development. Great effort has been put forth to find drugs that act only or predominantly on atrial tissue to reduce side effects (including the risk of ventricular arrhythmias). Multiple drugs that target I _Kur or acetylcholine-sensitive potassium channels are currently in the early stages of development. ¹¹⁴ Other novel drugs that stabilize ryanodine receptors (K201 ^{115 -117} ) or gap junctions (rotigaptide ^118,119 ) are also under development, but data in humans remain limited at this point in time. It is likely that other novel drug targets will be discovered in the future as well.

Antiarrhythmic Drug Selection

Selection of a specific AAD for use in the treatment of AF is primarily dictated by patient characteristics and safety. As noted previously, some AADs are absolutely contraindicated for use in patients with structural heart disease and reduced LV function. An algorithm for AAD selection based on patient characteristics based on the most recent ACC/AHA/HRS guidelines is presented in Figure 1. In patients with structurally normal hearts, it is reasonable to start with flecainide or propafanone (class IC agents), although sotalol and dronedarone are also reasonable options. Dofetilide is also effective, but due to the requirement that the drug be initiated during a 3-day hospitalization, this tends to be used as a second-line agent. Due to multiple long-term toxicities, amiodarone is generally reserved for patients who have failed other AADs or who are elderly, and its long-term use is usually avoided entirely in young patients.

Figure 1.

Antiarrhythmic drug selection for patients in AF based on characteristics and comorbidities. Adapted from January et al. ²⁴ AF indicates atrial fibrillation.

In patients with structural heart disease and/or coronary artery disease, AAD options are more limited, and class IC agents are contraindicated. In these patients, dofetilide or amiodarone are reasonable first-line options. In patients with significant CAD but without significant heart failure, sotalol and dronedarone can also be used, although these drugs tend to be less effective overall. In patients with LV dysfunction, dronedarone is contraindicated, and sotalol may have some negative inotropic effects. Patients with significant renal dysfunction are usually poor candidates for dofetilide or sotalol. Use of disopyramide is usually restricted to a very small subset of patients with hypertrophic cardiomyopathy, preserved LV systolic function, and a significant LV outflow tract gradient, although side effects can often limit its use, especially in men. Due to its side effects, quinidine is usually reserved for those who fail or have contraindications to other AADs, although as mentioned it can be safely used in patients with heart failure and coronary artery disease. In patients who fail treatment with AADs or who are unwilling to take an AAD, it may also be reasonable to consider catheter-based ablation for rhythm control.

Antiarrhythmic Drugs and Catheter Ablation of AF

In the last 2 decades, catheter ablation (pulmonary vein isolation) ¹²⁰ has emerged as a nonpharmacological option for rhythm control of AF. Multiple studies have compared the efficacies of AAD therapy and ablation in preventing AF recurrence; in general, ablation has demonstrated superiority to AAD therapy in preventing AF recurrence ^{121 -127} although it is invasive and associated with potential procedural complications, ¹²⁸ and multiple procedures are often necessary to obtain a durable result.

Studies have investigated whether catheter ablation and AADs are synergistic in preventing AF recurrence. In a study of 157 patients with AF who had previously failed AAD therapy randomized to catheter ablation and AAD therapy or AAD therapy alone, combined therapy was superior to AAD therapy alone (AF recurrence at 1 year 91% vs 44% for patients treated with AADs alone or combined therapy, respectively, P < .001). ¹²⁹

Recurrence of atrial arrhythmias within the first 3 months after catheter ablation for AF is common. ¹³⁰ Although early recurrence is associated with reduced long-term procedural success, ^130,131 it does not always signify that the procedure was not successful, ¹³² as early recurrence may reflect temporal variability in postprocedure atrial inflammation and maturation of ablation lesions. Suppression of early recurrence of AF after ablation with AADs has been investigated as a way of improving long-term ablation success.

In the Antiarrhythmics After Ablation of Atrial Fibrillation (5A) Study, 110 patients with paroxysmal AF who underwent catheter ablation were randomized to either 6 weeks of AAD therapy or no AAD therapy. The choice of AAD was left to the discretion of the treating physician, and most patients were started on flecainide or propafanone (60%) or sotalol (34%). At 6 weeks postablation, the composite end point of atrial arrhythmias lasting >24 hours; atrial arrhythmias associated with severe symptoms requiring hospitalization, cardioversion, initiation/change in AAD therapy, or AAD intolerance requiring drug discontinuation, was significantly less frequent in patients randomized to AAD therapy (19% vs 42%, P = .005). ¹³³ Short-term use of AADs after ablation, however, did not result in a reduction in AF recurrence at 6 months (72% vs 68% for patients randomized to AAD therapy and no AAD therapy, respectively, P = .84). ¹³¹ Similarly, in a study of 107 patients randomized to ablation combined with AADs for 1 year or ablation alone, long-term continuation of AADs after catheter ablation (71% amiodarone and 25% flecainide), resulted in a reduction in AF recurrence during the first month after ablation (17% vs 35%, P = .02) but no difference in rates of AF recurrence at 1 year after ablation (30% vs 34% for AAD therapy and no AAD therapy, respectively, P = .63). ¹³⁴

Initiation of AADs

There remains controversy about the safety of outpatient initiation of AADs primarily due to concerns about proarrhythmia. In general, all AADs currently used in the treatment of AF, with the important exception of dofetilide (which requires a mandatory 3-day inpatient hospital admission for drug initiation and continuous telemetry monitoring), can be safely initiated in the outpatient setting if used in appropriate patients with appropriate monitoring.

During dofetilide initiation, an ECG is obtained 2 to 3 hours after each dose of dofetilide, and the dose is adjusted (or the drug may be discontinued altogether) based on changes in the corrected QT interval. Continuous telemetry is also required to assess for episodes of TdP, which rarely occur after the first 3 days if the drug is well tolerated by the time of discharge. ⁸⁰ Initiation of sotalol is recommended to occur in the inpatient setting due to concerns about an elevated risk of TdP after conversion of AF, but sotalol was initiated safely in the outpatient setting in the SAFE-T trial. ⁴⁷

For AADs other than dofetilide, data have demonstrated that, in general, outpatient initiation can be performed safely with use of a continuous loop recorder and transtelephonic telemetry for 10 to 14 days. Adverse events are uncommon but can occur up to 9 days after drug initiation, when patients would have already been discharged from the hospital after inpatient admission. ^73,135 In a study of outpatient initiation of multiple AADs in 374 patients who were in SR, there were 3 (0.8%) deaths. Bradycardia was the most common adverse event that occurred in 3.7% of patients, the majority of whom were on amiodarone in addition to other AV nodal-blocking agents, and no other significant adverse events were noted. ⁷³ The safety of outpatient AAD initiation for patients who are in AF is less clear for AADs other than amiodarone. Additionally, in select patients who are in persistent AF and/or have multiple comorbidities, severe LV dysfunction, or significant baseline ECG abnormalities (including significant baseline QT prolongation or sinus bradycardia), inpatient initiation of AADs may be appropriate.

Antiarrhythmic Drug Monitoring

Once initiated, AADs require periodic monitoring. All patients on AADs should have an ECG obtained at regularly scheduled clinic visits to ensure there are no electrocardiographic signs of toxicity. For class IC agents (flecainide and propafanone), we routinely obtain a nonimaging exercise stress test 1.5 to 2 weeks after drug initiation to ensure there is no significant QRS widening during exercise. If the QRS duration increases by more than 30% with exercise, or if the QRS duration increases significantly at rest, the class IC agent should be discontinued.

For patients on chronic amiodarone, an ECG and thyroid and liver function tests should be checked every 6 months, and a chest X-ray should be checked annually. Pulmonary function tests should be checked at baseline, but then only need to be repeated if new pulmonary symptoms develop. It is reasonable to have patients periodically assessed by an ophthalmologist as well. ⁶¹

For patients on sotalol, there are no formal monitoring guidelines, although an ECG and renal function should be assessed periodically. Patients on dofetilide should have their renal function and ECG monitoring every 3 months in accordance with the manufacturer’s guidelines. ⁸⁰ Patients on dronedarone should have an ECG repeated every 3 months to assess for recurrence of AF, as the drug is contraindicated in patients who are in permanent AF. It is also reasonable to periodically monitor liver function tests while on dronedarone, especially during the first 6 months of use. As noted earlier, dronedarone should also be discontinued in patients who develop heart failure. ⁸⁵

Future Directions for Rhythm Control

Multiple older trials evaluating outcomes with rate and rhythm control in AF have not found a significant benefit to selecting a rhythm control strategy. More recently, however, data have suggested that maintaining SR might be beneficial, and rhythm control therefore might be especially useful in select patients.

Current AF guidelines consider the risk of thromboembolism in AF to be similar regardless of the duration and/or frequency of AF episodes (AF burden). ²⁴ Newer data, however, have challenged this long-held tenant of AF management. In a subanalysis of the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET-AF) (comparing anticoagulation with adjusted dose warfarin to rivaroxaban), patients with paroxysmal AF had a significantly lower all-cause mortality (adjusted HR 0.79, P = .006) and risk of stroke or systemic embolism (adjusted HR 0.79, P = .048) than those with persistent AF. Significantly more patients with paroxysmal AF than persistent AF were on AADs (28% vs 9%, P < .0001). ²² A similar reduction in stroke and mortality in patients with lower AF burden was seen in a subanalysis of the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) study (comparing adjusted dose warfarin to apixaban). ²¹ In these studies, all patients were on anticoagulation regardless of AF burden, which might explain why these results differ those in the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) trial ¹⁵ and other trials which found no mortality benefit to maintaining SR. Whether or not treatment with an AAD can reduce this excess risk of thromboembolism and/or death by reducing AF burden, or whether a high burden of AF is simply a marker for increased risk remains to be determined, and answering this question will require large randomized controlled trials.

In addition to potential mortality benefits of maintaining SR, in the future, successful rhythm control and new cardiac rhythm monitoring technology might also allow safe discontinuation of anticoagulation used for thromboembolic prophylaxis in select patients. Studies evaluating AF burden using cardiac implantable electronic devices (pacemakers, implantable cardioverter defibrillators (ICDs), and implantable loop recorders) which allow highly accurate assessment of the frequency and duration of AF episodes have demonstrated that episodes of subclinical AF are common and that AF burden may be an important risk factor for thromboembolic stroke. ¹³⁶ The Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pacemaker Patients and the Atrial Fibrillation Reduction Atrial Pacing Trial (ASSERT) followed 2580 patients with dual chamber pacemakers or ICDs and asymptomatic AF and found that AF episodes lasting longer than 6 minutes were associated with a 2.5-fold increase in the hazard of thromboembolism, and episodes lasting more than approximately 18 hours were associated with a 5-fold increase in the hazard of thromboembolism. The total number of AF episodes had no association with the risk of thromboembolism suggesting that episode duration rather than frequency was the more important determinant of thromboembolic risk. ¹³⁷

Botto et al ¹³⁸ prospectively followed 568 patients with dual chamber pacemakers and found that the risk of thromboembolism increased with higher AF burden across all CHADS₂ scores ²⁴ (a clinical scoring system for thromboembolic risk which ranges from 0 to 6). Patients were then stratified based on CHADS₂ score and duration of AF episodes. The authors found that there were 2 populations with significantly different levels of thromboembolic risk. In patients with low (0) or high CHADS₂ scores (≥3), AF burden had little influence on thromboembolic risk. In patients with intermediate CHADS₂ scores (1-2), AF burden was able to provide further risk stratification for thromboembolic risk, and revealed an equivalent risk of thromboembolism for a patient with a high CHADS₂ score and any burden of AF, and patients with a CHADS₂ score of 1 and a high burden of AF. ¹³⁸ Thus, in patients with relatively low thromboembolic risk, AF burden appears to be a significant risk factor for stroke and systemic embolism, and reduction in AF burden with AADs might allow modification of this risk.

Randomized studies evaluating use of AADs to suppress AF and reduce thromboembolic events do not yet exist. There are data, however, showing feasibility of “pill-in-the-pocket” anticoagulation for patients with low CHADS₂ scores and rare episodes of AF. In the Rhythm Evaluation for Anticoagulation With Continuous Monitoring (REACT.COM) study, 59 patients with a low burden of AF and a Medtronic Reveal (R) implantable loop recorder (Medtronic, Minneapolis, MN, USA) used for automated detection of AF stopped anticoagulation after 60 days without any AF episodes lasting >1 hour and resumed anticoagulation for 30 days (with either dabigatran, rivaroxaban, or apixaban to allow rapid onset of anticoagulation) if they experienced an AF episode lasting >1 hour. Using this approach, there was a 94% reduction in time on anticoagulation compared to chronic anticoagulation. Three patients experienced transient ischemic attack while off anticoagulation, but none of these patients had any AF around the time of their neurologic events. ¹³⁹ A large, randomized control trial evaluating the safety of implantable loop recorder tailored anticoagulation will be starting in 2016 (REACT-AF Trial, Clinicaltrials.gov NCT02488421). If this approach is safe, a future indication for AAD therapy may be to reduce the need for continuous anticoagulation and its associated risks and costs in select patients with appropriate monitoring.

Conclusion

Multiple AADs are available to assist in the maintenance of SR in patients with AF. Although historical data have suggested that rate-control and rhythm control strategies in AF are equivalent, newer data suggest that more aggressive maintenance of SR might be beneficial in terms of a reduction in mortality and stroke. However, until randomized and prospective trial data are available, it remains unclear whether patients with a lower AF burden do better because they are in AF less of the time (thus reducing the burden of AF with AADs would be expected to be beneficial) or if AF burden is simply a marker of more advanced comorbidities and risk that cannot be modified by restoring or maintaining SR. Additionally, rhythm control with AADs and continuous rhythm monitoring might allow select patients to avoid the need for continuous anticoagulation during prolonged periods of SR.

Regardless of the relative importance of rate- and rhythm control strategies, AADs will always have a role in treating AF to reduce bothersome symptoms, to maintain SR in patients with heart failure who benefit from AV synchrony, or in those with a tachycardia-mediated cardiomyopathy due to AF. However, AADs do have side effects and risks, and clinicians need to be aware of the specific indications, contraindications, and important pharmacokinetic and pharmacodynamic interactions for each drug. Newer, more effective, and safer AADs for the acute conversion and/or maintenance of SR are hopefully on the horizon. AADs may also continue to have a synergistic role with catheter ablation of AF. Ultimately, improvement in the safety profile and tolerability of AADs might usher in an era where pharmacological rhythm control is found to be superior to rate control in certain patient populations.