Atrial fibrillation (AF) is the most common arrhythmia in clinical practice, and in many ways it is the most complex to manage. Understanding of this disorder has increased immensely in the past decade, and new therapeutic options are developing at a rapid pace. AF is increasingly recognized as a heterogenous disease that cannot be managed with a “one size fits all” solution. Instead, the approach to the optimal treatment of a patient with AF requires the clinician to address the underlying heart disease and its interactions with potential therapies, the individual’s estimated risk of thromboembolism associated with the arrhythmia, and the presence and severity of arrhythmia-related symptoms. The general physician and cardiologist treating AF should know when to refer the patient for an electrophysiologic intervention, and the cardiac electrophysiologist should know when to stay the course and when to offer one of several invasive approaches.
AF may be classified by the underlying disease, the frequency and duration of its occurrence, or a combination of both. Paroxysmal AF is defined as a self-terminating arrhythmia, usually lasting less than 48 hours and rarely lasting more than 7 days. Persistent AF is an arrhythmia that fails to spontaneously convert but will do so with intervention, usually cardioversion or drug therapy. Permanent AF is an arrhythmia that is resistant to conversion to sinus rhythm and is the loosest term of the three; its definition in many cases depends on the vigor with which the physician pursues sinus rhythm restoration. With the advent of ablation procedures for AF, a new term, long-standing persistent AF, has been introduced to describe patients with a history of continuous AF longer than 1 year in whom an ablation procedure is attempted. To these categories is added the term new-onset AF to represent the initial documented event, since the outcome of a first episode of arrhythmia is not known. These temporal categories are not static and may merge with one another over time, as new-onset or paroxysmal AF become persistent in a significant percentage of patients. The presence or absence of symptoms, their severity, and an assessment of whether they are related to inadequate ventricular rate control or loss of normal AV synchrony is a helpful adjunct to the classification of AF, but it should be recognized that thromboembolic risk is unrelated to the presence or absence of arrhythmia-related symptoms.
Decision for Rhythm or Rate Control
A series of randomized, controlled trials of rhythm control compared with rate control were published 10 years ago. These studies were generally conducted in older patients with AF and clinical risk factors for thromboembolism. The study methods varied, and anticoagulation in the rhythm-control group was not mandated in most instances, although it was used in the majority of patients. In the largest of these trials, the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study, the primary endpoint was death, and the uniform finding of all these studies was that neither strategy was superior to the other in terms of any of the primary endpoints. When a post hoc analysis was done as to outcomes by rhythm, regardless of assigned strategy, the presence of sinus rhythm in the AFFIRM trial favored survival; the presence of coronary disease, diabetes, and/or smoking were adverse factors in terms of mortality. However, to maintain sinus rhythm, antiarrhythmic agents generally had to be used and were associated with a decreased likelihood of survival, which offset any potential benefit of sinus rhythm. Although this observation might be used to argue that an optimal strategy is sinus rhythm maintenance by nonpharmacologic means, the complexity of this retrospective analysis is such that it should only be thought of as a hypothesis-generating observation, not a well-supported conclusion. Indeed, a subsequent analysis of the Comparison of Rate Control and Rhythm Control in Patients with Recurrent Persistent Atrial Fibrillation (RACE) trial data performed by actual rhythm failed to demonstrate any benefit of sinus rhythm over AF, although functional status was slightly better in the AFFIRM trial among patients in sinus rhythm regardless of assigned group.
The trials of rate versus rhythm control are not the last word in this controversy because large groups of patients—including the young and very old, those with congestive heart failure (CHF), and those with highly symptomatic AF—were underrepresented. The role of restoration of sinus rhythm in a group of patients with CHF and an ejection fraction of 35% or less was subsequently investigated in the Atrial Fibrillation and Congestive Heart Failure (AF-CHF) trial. Amiodarone was the predominant drug used to maintain sinus rhythm, and there was no difference in outcome in patients randomized either to the rhythm or to the rate control groups. Despite all these apparently negative outcomes for a strategy of vigorous restoration and maintenance of sinus rhythm, it should be recognized that for many patients, the onset of persistent AF is associated with troublesome symptoms and a diminished overall quality of life despite adequate rate control. Many such patients were likely excluded from clinical trials, and data do support restoration of sinus rhythm for improving quality of life. Thus, when faced with a patient with recent-onset AF, the clinician should take a comprehensive history and have a low threshold for at least one attempt at sinus rhythm restoration to assess whether this leads to an improved sense of well-being. As a consequence, debate is ongoing regarding the relative merits of these two approaches in large numbers of AF patients. Regardless of the unanswered questions left by these trials, an important and consistent finding was that stroke risk was not reduced by the strategy of rhythm control when compared with that of rate control and anticoagulation. Although most of the patients in both groups received warfarin, it was stopped in the earlier trials more often in the group in whom sinus rhythm had apparently been restored; recurrent sustained or paroxysmal AF in this group was most likely responsible for thromboembolic events. Decisions regarding rhythm compared with rate control must therefore not be based on the desire to withdraw or withhold warfarin because recurrence is common and may be both asymptomatic and paroxysmal, rendering its detection on routine examination difficult. Although the decision to attempt to maintain sinus rhythm or choose rate control should be an individualized decision based on a careful assessment of clinical and echocardiographic features in each patient, younger patients with severe symptoms are generally preferred candidates for attempted rhythm control, whereas older patients with minimal symptoms may be equally if not better served by a rate-control strategy.
In patients with new-onset AF, the average ventricular response is 100 to 150 beats/min. The irregular R-R intervals are associated with a marked variation in stroke volume, and this possibly contributes to the uncomfortable sensation of AF with a poorly controlled heart rate. With stimuli that result either in vagal tone withdrawal or an increase in sympathetic tone—such as exertion, fever, hyperthyroidism, or blood loss—the ventricular rate in AF may increase markedly. Thus, a good question for the clinician to ask is whether the patient would have sinus tachycardia in sinus rhythm. If the answer is affirmative, the rapid ventricular response may reflect the presence of one of these conditions, which will need to be corrected before rate control can be achieved.
For some patients, particularly those with paroxysmal AF, a rapid ventricular response is associated with uncomfortable symptoms, most commonly of palpitations or dyspnea. In persistent AF, a disproportionate rise in heart rate with exertion may lead to dyspnea or fatigue.
Pharmacologic Rate Control in Atrial Fibrillation: Optimal Ventricular Rate
Control of the ventricular rate in AF has a twofold aim: elimination of symptoms and improvement of cardiac efficiency. As heart rate increases in AF, stroke volume tends to decrease, but cardiac output is generally maintained over a relatively wide range, probably decreasing when mean heart rate exceeds 110 to 120 beats/min. However, a higher resting heart rate probably is inefficient because it is associated with a higher myocardial oxygen demand. Thus, it is appropriate to attempt to reduce heart rate during AF.
The AFFIRM investigators empirically defined adequate rate control as an average heart rate of 80 beats/min or fewer at rest and either a maximum heart rate of 110 beats/min or fewer during a 6-minute walk or, during 24 hours of ambulatory monitoring, an average heart rate of 100 beats/min or less with no heart rate greater than 110% maximum predicted age-adjusted exercise heart rate. Using this definition, β-blocking agents were the most effective initial drug for rate control, with 70% success if used either alone or with digoxin, compared with 54% for calcium channel antagonists and 54% with digoxin alone.
These results, based on an older population with cardiac disease, document not only the apparent superiority of β-blockers but also that digoxin alone, although not effective in everyone, still has a role to play in heart-rate control in patients with AF. Furthermore, digoxin is synergistic both with calcium channel blockers and β-blockers and may decrease the required dosages of these agents and their associated side effects.
It is important to recognize that tighter heart rate control based on the number of beats per minute does not necessarily translate into improved symptoms. Although calcium channel antagonists are less effective than β-blockade for heart rate control, trials of β-blockers in patients with sustained AF have failed to show an improved exercise tolerance. Moreover, if peak heart rate is excessively blunted, exercise tolerance may decrease. A comparison of more stringent rate control that used criteria similar to the AFFIRM trial, with lenient rate control (resting heart rate <110 beats/min), showed no difference in outcome over a 3-year follow-up; this suggests that tight rate control may not be necessary in all patients.
When considering agents for ventricular rate control in AF, therapies are often divided into β-blocking agents, calcium channel antagonists, and digoxin—as though each of these classes of agents was mutually exclusive ( Table 20-1 ). Although monotherapy may be effective in a significant number of patients, the addition of digoxin and intermediate dose of a β-blocker or calcium channel blocker may produce excellent heart rate control, even when a high-dose single agent has failed. Evaluation of information from the AFFIRM trial shows that physicians frequently use a combination of agents for heart rate control and that combination therapy seems to have a greater likelihood of tight rate control compared with monotherapy with any agent. In a small but well-designed crossover trial, it was shown that the addition of digoxin to either diltiazem or to a β-blocker resulted in superior rate control than when any of these agents was used as monotherapy. Furthermore, rate control could be achieved with a combination that included digoxin at a lower dose than the other agent and with a reduced prevalence of side effects. However, clinicians should recognize that the negative chronotropic effects of the combination may produce significant sinus bradycardia, should sinus rhythm return. If verapamil is chosen, the verapamil-digoxin interaction is likely to necessitate a lower dose of digoxin than would be required if it were used alone or with diltiazem.
|DRUG||CONTROL OF ACUTE EPISODE||CONTROL OF SUSTAINED AF||COMMENTS|
|Calcium Channel Blockers|
|Diltiazem||20 mg bolus followed, if necessary, by 25 mg given 15 min later; maintenance infusion of 5-15 mg/h||Oral controlled-release diltiazem 180-360 mg/day||Long-term rate control may be better with the addition of digoxin.|
|Verapamil||5-10 mg IV over 2-3 min repeated once 30 min later; maintenance infusion rate is not reliably documented||Slow-release verapamil 120-240 mg qd or bid||Use causes elevation in digoxin level; may have more negative inotropic effect than diltiazem.|
|Esmolol||0.5 mg/kg IV, repeated if necessary; follow with infusion at 0.05 mg/kg/min, increasing as needed to 0.2 mg/kg/min||Not available in oral form||Hypotension may be troublesome but responds to drug discontinuation.|
|Metoprolol||5-mg IV bolus repeated twice q2min; no data on maintenance infusion||50-400 mg/day in divided doses||Metoprolol is useful if concomitant coronary disease present.|
|Propranolol||1-5 mg IV given over 10 min||30-360 mg in divided doses or as long-acting form qd||Noncardioselective; use with caution in patients with history of bronchospasm.|
|Digoxin||1.0-1.5 mg IV or PO over 24 h in increments of 0.25-0.5 mg||0.125-0.25 mg/day||Digoxin is renally excreted, with a slow onset of action IV and less effective control than other agents, although it may be synergistic with them. The least effective agent but may be acceptable monotherapy in sedentary patients.|
Nonpharmacologic Approach to Rate Control
Interruption of atrioventricular (AV) nodal function with catheter ablation and concomitant pacemaker implantation is a highly effective strategy for rate control. It is a relatively simple procedure that is associated with a significant and sustained improvement in quality of life. Data suggest, however, that at least in the setting of impaired ventricular function, right ventricular (RV) pacing may aggravate heart failure in some patients. This realization has tempered enthusiasm for this procedure, except in cases for which pharmacologic management and catheter ablation of the atrial arrhythmia prove impossible. An alternative to RV pacing is biventricular pacing, which preserves ventricular synchrony to a greater degree and thus may not have the adverse effects associated with RV pacing.
The strategy of rhythm control involves the restoration and maintenance of sinus rhythm. AF may spontaneously revert to sinus rhythm or may persist indefinitely unless cardioversion is performed. Spontaneous return of sinus rhythm occurs most often within 48 hours of arrhythmia onset. If this does not occur, sinus rhythm can be restored through the use of antiarrhythmic drugs or direct current (DC) electrical cardioversion. Before cardioversion, appropriate precautions must be taken to prevent thromboembolic events.
Intravenous and oral drugs are available for pharmacologic cardioversion ( Table 20-2 ). This approach is most successful when used within 24 to 48 hours of onset of AF, and the use of pharmacologic agents for conversion should be followed by electrical therapy if this approach fails. The currently available intravenous (IV) agents in the United States are limited to procainamide and ibutilide (see Chapter 18 ). The success rate of these agents for the conversion of AF is poorer with procainamide than with ibutilide, and ibutilide is better for the conversion of atrial flutter than it is for AF.
|DRUG||ROUTE OF ADMINISTRATION||DOSAGE *||POTENTIAL ADVERSE EFFECTS|
|Amiodarone||PO||Inpatient: 1.2-1.8 g qd in divided doses until 10 g total, then 200-400 mg/day maintenance or 30 mg/kg as a single dose||Hypotension, QT prolongation, torsades de pointes (rare), GI upset, constipation, phlebitis (IV)|
|PO||Outpatient: 600-800 mg/day in divided doses until 10 g total, then 200-400 mg/day maintenance|
|IV/PO||5-7 mg/kg over 30-60 min, then 1.2-1.8 g/day continuous IV or in divided oral doses until 10 g total, then 200-400 mg/day maintenance|
|Dofetilide||PO||Creatinine clearance (mL/min):||Dose (µg bid):||QT prolongation, torsades de pointes; adjust dose for renal function, body size, and age|
|Flecainide||PO/IV||PO: 200-300 mg † |
IV: 1.5-3.0 mg/kg over 10-20 min †
|Hypotension, atrial flutter with high ventricular rate|
|Ibutilide||IV||1 mg over 10 min; repeat 1 mg when necessary||QT prolongation, torsades de pointes|
|Propafenone||PO/IV||600 mg 1.5-2.0 mg/kg over 10-20 min †||Hypotension, atrial flutter with high ventricular rate|
|Quinidine ‡||PO||0.75-1.5 g in divided doses over 6-12 h, usually with a rate-slowing drug||QT prolongation, torsades de pointes, GI upset, hypotension|
† Insufficient data are available on which to base specific recommendations for the use of one loading regimen over another for patients with ischemic heart disease or impaired left ventricular function; these drugs should be used cautiously or not at all in such patients.
Electrocardiographic (ECG) monitoring should be used when administering these drugs because of the risk of torsades de pointes (TdP) with procainamide and ibutilide—as high as 3% to 5% with ibutilide —and hypotension with procainamide. Patients must be monitored for proarrhythmia for at least 2 hours following the termination of the infusion. It had been shown that the combination of the class 1C agent propafenone administered orally with IV ibutilide increases the likelihood of conversion to sinus rhythm and is well tolerated, but experience with this combination is limited to one publication.
An alternative approach to IV therapy is the use of high-dose oral antiarrhythmic drugs for the conversion of AF to sinus rhythm. Quinidine in an initial dose of 200 mg, repeated at a dose of 200 mg every 2 hours for three separate doses, has a high success rate for conversion, but it has fallen into disfavor because of a high incidence of side effects, including TdP. High-dose oral therapy has been widely studied with high doses of type 1C medications (450 to 600 mg of propafenone or 300 to 400 mg of flecainide). These drugs should be avoided in patients with bundle branch block, structural heart disease, or ventricular preexcitation because of a risk of proarrhythmia, including atrial flutter with 1 : 1 AV nodal conduction, hemodynamic collapse, and ventricular tachycardia.
The initial administration of these drugs should be in a monitored setting with the capability for defibrillation. If found to be successful without adverse effects, oral therapy can subsequently be self-administered by selected patients with a structurally normal heart, the so-called pill-in-the-pocket approach. Oral amiodarone has been extensively studied for the conversion of AF and has been shown to be safe to load during AF in the outpatient setting; it is also associated with a rate of successful cardioversion of up to 80% after 24 hours in AF of 48 hours’ duration or less, although the rate is much lower for longer duration arrhythmia.
Electrical cardioversion can restore sinus rhythm in AF more than 90% of the time. Successful cardioversion is defined as the restoration of sinus rhythm, if even for only one beat, and it is inversely related to the duration of AF and the size of the left atrium. Biphasic defibrillator waveforms are more successful than monophasic waveforms, and anteroposterior electrode positioning is somewhat more successful than anterolateral placement, although this is not a consistent finding. Reversion to AF immediately after a single beat of sinus rhythm or within the subsequent 24 hours occurs frequently, and these phenomena are called immediate recurrence of AF (IRAF) or early recurrence of AF (ERAF). Early recurrences may be prevented by antiarrhythmic drugs such as ibutilide.
If cardioversion fails at the maximum energy output of the device, a second shock at the same energy level a minute or so later may succeed, as transthoracic impedance falls with each shock. Alternatively, a change to another orientation will occasionally be successful. Application of pressure to the anterior electrode will also diminish the transthoracic impedance and improve success rates. Finally, pretreatment with an intravenous or oral antiarrhythmic drug can facilitate cardioversion by lowering defibrillation thresholds and reducing early recurrence of AF.
Maintenance of Sinus Rhythm
Antiarrhythmic drugs are widely prescribed for the maintenance of sinus rhythm. In general, these agents have a similar efficacy, with a 50% rate of recurrence at 1 year compared with 20% to 25% maintenance of sinus rhythm in the absence of antiarrhythmic drugs. Amiodarone has consistently been shown to be more efficacious than other available antiarrhythmic drugs, with up to a 75% rate of AF suppression at 1 year. If one agent fails to maintain sinus rhythm, a second agent may be successful, particularly one from another group of drugs. It is unrealistic to expect complete suppression of AF with any of these agents. Instead, the goal should be a significant reduction in the frequency of AF compared with pretreatment patterns.
The major toxicities of antiarrhythmic drugs include both proarrhythmia and noncardiovascular adverse effects. The noncardiovascular toxicities differ by drug and range from benign changes in taste to life-threatening pulmonary or liver toxicity ( Table 20-3 ). All antiarrhythmic drugs alter cardiac sodium and/or potassium channel function. It is believed that these drugs prevent or terminate AF by prolonging the refractory period (potassium channel blockers) or slowing the conduction (sodium channel blockers) of atrial cells. Prolongation of the refractory period (prolongation of repolarization) results in QT prolongation, which may be excessive if dosing is too high, excretion is reduced, or the patient has either a genetic predisposition to a prolonged QT or a genetically prolonged drug metabolism. Prolongation in conduction such as that produced by the class IC agents causes QRS prolongation, which is more marked at faster heart rates. QT prolongation may result in TdP, and antiarrhythmic drugs that block the delayed rectifier potassium channel (I kr and/or I ks ) may cause TdP in up to 5% of patients. Drug-related TdP is more likely to occur in association with slow heart rates, electrolyte abnormalities (hypokalemia or hypomagnesemia), female gender, prior unrecognized congenital long QT syndrome, and pauses associated with the conversion of AF to sinus rhythm. The concomitant use of drugs that interfere with the hepatic metabolism of antiarrhythmic drugs may also result in QT prolongation, and a reduced urinary clearance of renally excreted medications may result in toxicity. In some instances, such as with sotalol or dofetilide, the risk of TdP is proportional to blood levels and is related to renal excretion ; with quinidine it may be idiosyncratic and not dose related. A metabolite of procainamide, N-acetyl procainamide, prolongs the QT interval, whereas the parent compound has little effect on repolarization. Slow acetylators produce less N-acetyl procainamide and have a lower risk of TdP. Rapid acetylators develop more N-acetyl procainamide and are more prone to TdP.
|DRUG||DOSE (24 H)||CARDIAC TOXICITY||NONCARDIAC TOXICITY||RECOMMENDED MONITORING AND DRUG INTERACTIONS|
|Amiodarone||200-400 mg (initial load 600-1200 mg for 1-2 wk) with lower dosing for smaller and older patients||Bradycardia, TdP (rare)||Pulmonary toxicity, photosensitivity, hepatic toxicity, GI upset and neurologic toxicity (dose related), thyroid dysfunction, rise in INR with warfarin||LFTs: q6mo, PFTs at baseline and only with symptoms of potential toxicity; CXR yearly |
TFTs: baseline, 3 mo, then q6mo; potentiates warfarin, digoxin, dilantin, tricyclics
|Dronedarone||400 mg bid||None; avoid in patients with congestive heart failure or permanent AF||Hepatic toxicity; impairs secretion of creatinine||Monitor for recurrent AF at least q3mo|
|Dofetilide||500-1000 µg||TdP||NS||In-hospital initiation for 72 h, creatinine clearance q3mo QT interval; levels increased by multiple drugs (e.g., verapamil)|
|Sotalol||160-320 mg||TdP, CHF, sinus bradycardia||Bronchospasm||QT interval|
|Flecainide||200-300 mg||Ventricular tachycardia, atrial flutter with 1 : 1 conduction||Dizziness||NS|
|Propafenone||450-900 mg||Ventricular tachycardia, atrial flutter with 1 : 1 conduction||Metallic taste||NS|
|Quinidine||600-1500 mg||TdP, enhanced AV nodal conduction||Thrombocytopenia, fever, nausea, diarrhea||QT interval; platelet count; will increase digoxin levels; can potentiate warfarin|
|Procainamide||1-4 g||TdP||Agranulocytosis, lupuslike syndrome||QT interval; NAPA increased by ethanol|
|Disopyramide||400-750 mg||TdP, CHF||Urinary retention, dry mouth; contraindicated in glaucoma||QT interval|
Ventricular tachycardia may occur in patients taking antiarrhythmic drugs. This complication is well described in patients taking type 1C medications (flecainide and propafenone) who have had a prior myocardial infarction (MI) and have impaired ventricular function. Atrial flutter with a slow atrial rate and one-to-one AV nodal conduction, producing a widened QRS duration with hemodynamic collapse, may also occur, particularly with class 1C drugs. This latter complication can generally be avoided with the addition of AV node–blocking medications.
Bradyarrhythmias develop most often as a result of sinus node suppression or slowing of conduction through the AV node. Both of these complications are more frequently seen in elderly patients with underlying sick sinus syndrome. Ambulatory monitoring and appropriate dose reduction or discontinuation can prevent serious consequences.
Choice of Drug
The toxicity of antiarrhythmic drugs can be reduced by the selection of agents in the context of the patient’s clinical history. In brief, the risk factors for adverse effects include myocardial scarring (most commonly as a result of earlier MI), left ventricular (LV) systolic dysfunction, and possibly LV hypertrophy. The presence of any of these clinical features should be evaluated before choosing an antiarrhythmic drug ( Figure 20-1 ). For patients without structural heart disease, the choice of drugs is wider, but it is still important to evaluate noncardiac comorbidity.
Initiation and Monitoring of Antiarrhythmic Drugs
The toxicity of antiarrhythmic drugs can also be reduced with the appropriate choice of dosage and method of monitoring during the loading phase . For example, unexpected bradycardia can often be avoided by reducing the usual loading dose of amiodarone or by starting with a low dose of sotalol and increasing to therapeutic dose as tolerated. This strategy is particularly important in patients with suspected sinus node dysfunction.
Some controversy exists regarding the need for in-hospital initiation of antiarrhythmic drugs for the treatment of AF, the major concern being the precipitation of TdP. This proarrhythmic effect occurs in drugs that prolong repolarization, and it is almost never seen with flecainide and propafenone. In-hospital monitoring for 72 hours is incorporated into the drug labeling of dofetilide and is therefore mandatory for this drug, regardless of the presence or absence of structural heart disease. Quinidine is rarely used nowadays because of its significant and idiosyncratic propensity for provoking TdP, but it is an effective atrial antiarrhythmic agent. Experimental data suggest that the risk of TdP can be reduced by the concomitant use of verapamil. Two large trials of quinidine, for paroxysmal AF or for maintenance of sinus rhythm after an episode of persistent AF, demonstrated that when quinidine was combined with verapamil, it was as effective as sotalol, but with less TdP risk. As previously noted, a pause associated with the conversion of AF to sinus rhythm may promote the development of TdP. Consequently, in patients with paroxysmal AF, it is advisable to initiate antiarrhythmic drugs with the potential for TdP while the patient is in sinus rhythm.
In contrast to QT-prolonging agents, in-hospital initiation of type 1C agents, which should only be used in patients with a structurally normal heart, and amiodarone, which may be used in any form of heart disease, is generally not required. There is considerable experience with the initiation of amiodarone in the outpatient setting during AF without significant toxicity. QT prolongation is common with amiodarone, but it is not generally associated with a great risk of TdP (<1%), unless the corrected QT interval is significantly prolonged (>500 ms).
Dronedarone has been approved for the treatment of AF and can be started in the outpatient setting. It is similar in chemical structure to amiodarone but lacks the iodine moiety, so thyroid dysfunction is less of a concern. However, it is less effective at maintaining sinus rhythm than amiodarone and should not be used in patients with recently decompensated heart failure because of an increased risk of death. 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) was the first trial to demonstrate prospectively a decrease in cardiovascular hospitalizations in patients with AF, although this has been shown with dronedarone in a retrospective analysis and is probably simply a function of maintenance of sinus rhythm. In a post hoc analysis of the ATHENA trial, stroke incidence was also lower in patients treated with dronedarone. The main side effect of dronedarone is diarrhea, and there may be a previously unrecognized risk of hepatic toxicity associated with this therapy. Data from the Permanent Afibrillation Outcome Study Using Dronedarone on Top of Standard Therapy (PALLAS) trial, which used dronedarone in patients with persistent AF, showed increased toxicity and was stopped. Specifically, there was an increase in stroke, cardiovascular hospitalization, and cardiovascular mortality rate in patients taking dronedarone. At present, it is recommended to only use dronedarone if frequent monitoring (at least every 3 months) fails to show AF recurrence.
The concern with type 1C agents is the conversion of AF to atrial flutter with 1 : 1 conduction to the ventricles and hemodynamic instability. It is therefore strongly recommended that AV node–blocking agents be used in conjunction with a type 1C agent. Particular caution should be exercised for patients who are athletic because the development of atrial flutter during exercise may be associated with 1 : 1 conduction, even in the presence of digoxin or calcium channel blockers. Theoretically, the antisympathetic aspects of β-blockade should be more effective in such patients.
Amiodarone may be initiated in an outpatient setting during AF given the widespread safety data with this practice. If there is no CHF, and if the patient is in sinus rhythm, it may be possible to initiate other antiarrhythmic drugs, with the exception of dofetilide, in the ambulatory setting. To increase the safety of this approach, patients can be monitored with a continuous event recorder. The transmission of a single 30-second tracing daily permits monitoring for bradycardia, QT prolongation, and tachyarrhythmias. This protocol is continued for 10 days and has been shown to be quite effective.
Despite safe initiation of an antiarrhythmic drug, the long-term possibility of proarrhythmic effects still exists. It is therefore important that both physicians and patients are aware of circumstances that may render previously tolerated medications dangerous. Examples of these situations include the initiation of diuretics or certain antibiotics in patients taking QT-prolonging drugs and the development of renal dysfunction in patients receiving renally excreted antiarrhythmic drugs such as sotalol.
Adjunctive Therapy for the Maintenance of Sinus Rhythm
Advances in the understanding of the electrical and mechanical remodeling of the atrium that occurs during AF have led to the evaluation of drugs that are not primarily antiarrhythmic in nature as an adjunct to maintaining sinus rhythm. The calcium channel–blocking agents may blunt electrical remodeling of the atrium, and several trials have evaluated their efficacy in the pericardioversion state. As single agents, diltiazem and verapamil do not appear to prevent recurrence of AF in humans. However, when given for a few weeks before cardioversion and continued for a few weeks after cardioversion of AF in the setting of an antiarrhythmic agent, there appears to be a modest, synergistic benefit. In large clinical trials of losartan versus atenolol for hypertension, AF was reduced in the group treated with losartan, and several small trials seem to demonstrate a benefit of pericardioversion use of these agents, although the results are more consistent with angiotensin receptor blockers (ARBs) than with calcium channel antagonists. Inflammation is believed to play a role in recurrence of AF, although small trials of statins, which reduce C-reactive protein, have shown contradictory results. The Gruppo Italiano per lo Studio Della Sopravvivenza nell’Infarto Miocardio–Atrial Fibrillation (GISSI-AF), the largest trial to date of valsartan compared with placebo, failed to demonstrate a reduction in recurrent episodes of AF in patients taking the ARB valsartan. Thus, if the proposed effect is class specific rather than drug specific, it seems unlikely that ARBs have any role in preventing postcardioversion AF. Currently, no specific recommendations about the use of either calcium channel antagonists or angiotensin-converting enzyme (ACE) inhibitors as an adjunctive therapy can be made.
Nonpharmacologic Approaches to the Maintenance of Sinus Rhythm
The invasive approach to the management of AF is rapidly evolving. As noted above, certain antiarrhythmic agents may transform AF into atrial flutter. Advantage has been taken of this transformation, because atrial flutter may be cured by the application of a line of ablation from the tricuspid valve to the inferior vena cava. This hybrid therapy, defined as the conversion of AF to atrial flutter by an antiarrhythmic drug, most commonly flecainide or propafenone, with subsequent ablation of atrial flutter is most commonly unplanned and is suitable for only a small minority of patients with AF in whom flutter fortuitously occurs.
Percutaneous left atrial ablation is now used widely for the attempted prevention of AF recurrence. Multiple permutations of this procedure exist, but all involve the electrical isolation of the pulmonary veins to prevent atrial premature beats from entering the left atrium and triggering the onset of AF. Some clinicians choose to create additional linear lesions in the left atrium to impair the perpetuation of AF, should it become triggered by a nonpulmonary vein source or an incomplete pulmonary vein isolation (PVI). The optimal ablation procedure and the best choice of patient population for this procedure continue to evolve. Although relatively uncommon, the risks include pulmonary vein stenosis, pericardial tamponade, stroke, and atrial-esophageal fistula formation.
This procedure is associated with a 60% to 70% rate of suppression of AF during a 12- to 24-month follow-up period. Left atrial tachycardia, flutter, or recurrent AF is seen in the remaining patients, and these arrhythmias often require treatment that includes repeat ablation. Five years after AF ablation, the majority of patients have at least one recurrence of the arrhythmia, and the need for repeat procedures is common.
Studies of arrhythmia surveillance after PVI have demonstrated significant rates of asymptomatic AF. It is therefore commonly advised to continue anticoagulation in patients with clinical risk factors for stroke, irrespective of the perceived success of the PVI procedure.
The surgical correlate of the percutaneous PVI procedure is the Maze procedure. The modern Maze procedure is a series of ablations performed on the endocardial surface, most often in conjunction with a coronary artery bypass or valve operation. The left atrial appendage (LAA) is generally oversewn as part of this procedure. Data on patients with preoperative AF undergoing the Maze procedure along with mitral valve surgery reported a 78% to 81% actuarial rate of sinus rhythm at 5 years compared with less than 10% in a group with mitral surgery but no Maze procedure. The efficacy of this procedure has been reported to be substantially higher in patients with paroxysmal AF, and the overall efficacy will vary from center to center, as it is dependent on the experience of the surgeon and the selection of patients. The obliteration of the LAA as part of the Maze procedure carries the potential additional advantage of reducing stroke risk. Nevertheless, caution must be taken in discontinuing warfarin after this procedure, because no data confirm that the combined Maze and LAA obliteration completely abolish thromboembolic risk.
A minimally invasive variant of the Maze procedure involves the creation of lesions around the pulmonary veins through a thoracoscopic or minithoracotomy approach. The LAA can also be obliterated from the epicardial surface. This procedure is under investigation, and the long-term efficacy is awaited.
The early recurrence of AF after the surgical Maze procedure approaches 30%, but these early recurrences do not necessarily predict long-term failure of the primary procedure. To reduce short-term postoperative AF after the Maze procedure, amiodarone is often prescribed for 1 to 3 months after surgery. If AF redevelops after discontinuation of amiodarone, the procedure is recognized as unsuccessful.
Pacing for the Maintenance of Sinus Rhythm
Current guidelines favor the use of dual-chamber compared with single-chamber pacing to reduce the frequency of AF in patients with a history of AF. Efforts to reduce the frequency of AF through alternative-site pacing techniques have demonstrated marginal to no benefit. The one exception maybe the use of postoperative pacing to prevent cardiac surgery–associated AF.
Stroke represents the most devastating complication of AF, and the percentage of strokes attributable to this arrhythmia increases with increasing population age. Thus it is estimated that more than 35% of strokes in patients older han 80 years are directly attributable to AF. Strokes in patients with AF tend to be more severe than those that result from other causes, such as carotid artery stenosis, and they carry a higher mortality rate. Several large trials performed in the late 1980s and early 1990s clearly demonstrated a very significant benefit of warfarin therapy for the prevention of stroke among patients with nonrheumatic AF. Paroxysmal AF was shown to have the same annual stroke incidence as persistent AF, and the development of stroke was shown to be an ongoing risk. These trials also defined several subsets of patients with AF who were at a greater risk of stroke than patients in whom the arrhythmia existed in isolation, so-called lone AF . The risk of stroke is greatest in patients with prior stroke or transient ischemic attack (TIA) at 11% per year; other risk factors include CHF, systolic ventricular dysfunction, hypertension (whether current or treated), older age, and diabetes. The finding on a transesophageal echocardiogram (TEE) of dense left atrial spontaneous contrast, diminished LAA velocities, or complex aortic plaque was associated with a risk of stroke in excess of 13% per year ; many of these features were associated with the clinical features noted earlier, and transesophageal imaging is not mandatory for risk stratification. Using a simple point system known as CHADS 2 , an annual stroke estimate can be assessed ( Table 20-4 ). For patients deemed to be at moderate to high risk for stroke, warfarin or equivalent anticoagulation is indicated, prescribed to maintain an International Normalized Ratio (INR) of 2.0 to 3.0. The role of aspirin therapy in AF is less clear. Low-dose aspirin (81 mg) is not effective, and the efficacy of higher doses (325 mg daily) is controversial. For patients who are deemed unsuitable for warfarin therapy yet are at a high risk of AF-associated stroke, the addition of clopidogrel to aspirin modestly decreases the risk of stroke but incurs an increased risk of major bleeding.