9. Arrhythmia Management

9.1 Acute management of AF

The acute management of AF is centred on the following domains:

1. Determination if AF is the primary concern (“primary AF”) or secondary to another acute medical illness (“secondary AF”). AF in the setting of critical illness has been associated with an increased risk of death (see section 11.5).^[505],[506] Unfortunately, there is a paucity of high-quality evidence on whether or how to treat AF patients in the setting of critical illness,^[507],[508] and there is a wide variety of reported approaches to AF management in this setting.^[509] In the ED setting, results of a retrospective study suggested that rate and rhythm control efforts in patients with AF secondary to acute medical illness (predominantly sepsis and acute HF) might be associated with higher rates of adverse events.^[510]

2. Determination of hemodynamic stability, defined as AF causing hypotension, ACS, or pulmonary edema. Acute unstable AF should be treated with synchronized direct current cardioversion (DCCV), however, instability solely due to AF is rare, therefore, an underlying precipitant should also be aggressively sought and managed.^[511]

3. Determination of an arrhythmia management strategy, defined as rate vs rhythm control. In patients with established AF multiple RCTs have shown no significant difference in cardiovascular outcomes between patients treated with a strategy with rate control vs rhythm control.^{[512]–[514]} In patients with newly diagnosed AF (ie, within a year) an initial strategy of rhythm control has been associated with reduced cardiovascular death and reduced rates of stroke.^[515]

4. Determination of the need for hospitalization. Most patients with AF can be safely discharged home after acute management. However, hospitalization might be required for highly symptomatic patients with AF in association with acute medical illness or complex medical conditions, in highly symptomatic patients in whom adequate rate control cannot be achieved, or in those who require monitoring or ancillary investigations not readily available in the outpatient setting.

5. Determination of the need for OAC. There is evidence that OAC prescription in the ED results in improved longterm use.^{[364],[516],[517]} As such, it is of paramount importance that OAC be initiated as soon as time allows in patients who undergo cardioversion for new-onset AF (see section 8.4.1), as well as in patients at risk of stroke (see the “CCS algorithm” in Fig. 8) whether or not attempts at rhythm control are made.

6. Early follow-up. Patients discharged from the ED with AF benefit from early follow-up. Ideally this should occur within a week of discharge, because early follow-up has been associated with lower rates of readmission and death.^[518] In addition, cardiology assessment within 3 months of hospital discharge for new-onset AF has been associated with lower rates of death, stroke, and major bleeding.^[519]

A general overview of rate and rhythm management of AF is provided in Figure 16, and the approach to the management of AF in the acute care setting is provided in Figure 17.

9.1.1 Acute rate control

In the setting of recent onset AF, the rate control agent and the formulation chosen will be influenced by clinical circumstance (eg, the presence of HF or hypotension) and patient comorbidities (eg, known LV dysfunction, reactive airways disease, hypotension, history of MI, or angina; Supplemental Table S10). Options include oral or I.V. β-blockers, oral or I.V. nondihydropyridine calcium channel blockers (ND-CCBs), I.V. digoxin, and I.V. amiodarone (recognizing that the latter is also a rhythm control agent). I.V. rate control agents might be initially considered, however if the patient is hemodynamically stable oral agents might be preferred. If an I.V. agent is used as the initial therapy, it is important to coadminister an oral rate control agent as soon as possible to maintain rate control/avoid rebound tachycardia as the I.V. formulation wears off.^[520] In patients without contraindications, β-blockers and calcium channel blockers (CCBs) are considered first-line agents for rate control. Only two small RCTs totalling 92 patients have compared I.V. β-blockers with CCBs in patients with recent-onset AF (one of which also included AFL), both showed that I.V. diltiazem was more effective at controlling the heart rate (< 100 beats per minute [bpm]) at 20-30 minutes compared with I.V. metoprolol (RR, 1.8; 95% CI, 1.2- 2.6).^{[521]–[523]} A retrospective study of 110 AF patients showed similar results at 60 minutes, although the success rate with diltiazem in that study (57%) was lower than those reported in the RCTs (90%-95%).^{[521],[522],[524]} In patients with ACS who require acute rate control, β-blockers are the agent of choice. Digoxin or amiodarone might be considered for acute rate control in the setting of decompensated HF, known significant LV systolic dysfunction (defined as LVEF ≤ 40%), or mild hypotension.^[7] However, it is important to recall that I.V. formulations of amiodarone can lower BP.^[525] Moreover, in this population the selective use of I.V. CCBs (and β-blockers) has been safely and successfully used in several randomized and nonrandomized studies, often with an improvement in BP when the recent-onset AF is rate controlled. Jandali performed a retrospective cohort study on the use of I.V. diltiazem in 162 patients with LV systolic dysfunction (LVEF ≤ 50%; with 52 having an LVEF ≤ 30%), and compared them with 473 patients with preserved LVEF (≥ 50%).^[526] There was no difference in the rates of hypotension, intensive care unit (ICU) transfer, or mortality between the patients with LVEF ≥ 50% vs those with LVEF < 50%, or those with LVEF ≥ 30% vs those with LVEF < 30%. Hirschy et al. performed a retrospective cohort study on the use of I.V. metoprolol (14 patients) and I.V. diltiazem (34 patients) in patients with known LV systolic dysfunction (mean LVEF 23% [15-35] vs 25% [15-30], respectively). Successful rate control within 30 minutes occurred in 62% of the metoprolol group and 50% of the diltiazem group (P ¼ 0.49), with no difference in complications or worsening HF.^[527] Goldenberg et al. performed a small double-blind RCT of 37 AF patients with reduced ejection fraction (EF; LVEF 36% ± 14%) and symptomatic HF (New York Heart Association [NYHA] class III [62%] and IV [38%]), in whom the cautious use of I.V. diltiazem resulted in therapeutic response (97%) with self-limited hypotension in 11%, and no patients experiencing worsening of HF.^[528] Although these studies have shown that the use of I.V. CCBs for rate control in highly selected patients can be safe, caution must be used because of the risk of precipitating cardiac decompensation.

9.1.1.1 Acute rate control targets

There have been no RCTs that specifically examined rate control targets in the acute care setting,^[529] nor in patients with paroxysmal AF.^[530]

9.1.2 Acute rhythm control

For stable patients with recent-onset AF who are eligible for cardioversion, the choice to pursue sinus rhythm restoration should be made on the basis of patient symptoms and goals of care, recognizing that early rhythm control has been associated with a lower risk of stroke and cardiovascular death.^[515] Because cardioversion increases the risk of systemic embolism, it is important to start appropriate anticoagulation as soon as time allows for all patients (see section 8.4.1).^[339] For patients with recent-onset AF who are eligible for cardioversion, rhythm control is preferred and can be established via either pharmacological or electrical cardioversion. In general, electrical cardioversion is more effective than pharmacological cardioversion, especially for more prolonged AF episode durations.^{[363],[531]–[534]} Pharmacological cardioversion has the advantage of being immediately feasible in a nonfasting patient, as well as avoiding the delays and risks associated with procedural sedation. However, most pharmacological agents have cautions or contraindications that limiting their use in patients with significant cardiac comorbidities, and their use requires a monitored bed, access to a crash cart, and a dedicated nurse to monitor for potential complications.

9.1.2.1 Pharmacologic cardioversion

Antiarrhythmic medication selection is typically dictated by the patient’s comorbidities as well as physician preference. Characteristics, indications, contraindications, and monitoring details of antiarrhythmic medications used for acute pharmacological cardioversion are presented in Supplemental Table S11.

9.1.2.1.1 Procainamide

Procainamide, a class Ia agent, is the most common I.V. medication used for cardioversion of recent-onset AF in the Canadian ED setting.^[535] Although procainamide can be administered as a 1-g bolus over 30 minutes followed by an infusion of 2 mg/min or a dose of 15-18 mg/kg administered over 60 minutes, most clinical evidence (including safety outcomes) was derived on the basis of a single infusion of 1 g over 60 minutes.^{[531],[536]–[538]} Procainamide is more effective for the conversion of recent-onset AF (50%-60% conversion) than for AFL (30% conversion).^{[531],[536]–[538]} The most common side effect is hypotension (approximately 5%), although premature ventricular contractions, runs of ventricular tachycardia and QRS widening might occur.^{[531],[536],[537]} This medication, like all drugs with class I (Na⁺ channel-blocking) action, should be avoided in patients with Brugada syndrome.^[539]

9.1.2.1.2 Ibutilide

Ibutilide is an I.V. class III agent that has been shown to effectively terminate AFL (50%-75%) and AF (30%-50%), with cardioversion typically occurring within 30-60 minutes.^{[540]–[546]} However, widespread clinical uptake has been limited by a significant risk of torsades de pointes (TdP) and ventricular tachycardia (most frequently nonsustained), each of which occurs in approximately 2%-3% of patients.^{[540]–[548]} Consequently, ibutilide should not be used in patients with a prolonged QTc (> 440 ms), a history of HF (typically defined as clinically symptomatic or NYHA classification > II), or reduced EF, signs of an ACS, and/or low serum potassium or magnesium levels.^{[540]–[545],[547]} Periprocedural I.V. magnesium (typically given pre- and post treatment) appears to improve ibutilide cardioversion rates, with higher doses (eg, 4 g total) more effective than lower doses (1-3 g total).^[549],[550] Patients must be observed with continuous ECG monitoring for a minimum of 4 hours after ibutilide administration.

9.1.2.1.3 Vernakalant

Vernakalant is an atrial-selective antiarrhythmic approved for conversion of AF. In patients treated within 48 hours of AF onset, randomized trials report a conversion rate at 90 minutes ranging from 52% to 69%, which is not significantly better than other active agents (combined comparator of ibutilide and amiodarone).^{[544],[545],[551]–[553]} However, the median time to cardioversion of 10-12 minutes is shorter than the next fastest pharmacological agent (ibutilide, median time to conversion 26 minutes). The major side effects are hypotension and bradycardia after cardioversion.^{[551],[554],[555]} Transient but fairly common side effects include dysgeusia, paresthesia, and nausea.^{[544],[551],[554],[555]} Vernakalant is not effective for the conversion of AFL; and should be avoided in patients with hypotension, severe HF (NYHA classification III/IV), bradycardia, recent ACS, or severe aortic stenosis.^{[552],[555],[556]}

9.1.2.1.4 Amiodarone

With the exception of patients with structural heart disease, amiodarone is not recommended for acute rhythm control because of a delay in conversion (approximately 8 hours).^{[525],[532],[557]} The most common adverse drug reactions with I.V. administration are phlebitis, hypotension, and bradycardia.^[525],[532] Although there is potential for prolongation of the QT interval, the incidence of TdP is rare.^[532],[557]

9.1.2.1.5 Flecainide and propafenone

I.V. flecainide and propafenone are superior to placebo but are not currently available in Canada for acute care cardioversion.^[525],[558] The oral formulations, however, have similar, if slightly delayed, efficacy as their I.V. counterparts.^[558],[559] Three hours after administration of a single dose of oral flecainide, between 57% and 68% of patients will convert.^[532] Success rates with oral propafenone are similar.^[532],[559] Although the time to cardioversion (approximately 2-6 hours) is longer than with I.V. formulations, the major clinical benefit is that patients are able to treat their AF episodes at home (“pill-in-the-pocket”), which reduces the need to visit the ED for recurrences. A key caveat to this approach is that the first treatment attempt must be administered in a monitored environment, to verify efficacy and exclude treatment-related adverse reactions.^{[557],[560]–[563]} A β-blocker or ND-CCB should be given ≥ 30 minutes before administration of a class Ic antiarrhythmic to prevent the risk of 1:1 AV conduction during AFL. One study suggests that rare adverse events can occur even after successful use in a monitored environment^[563]; therefore, clear instructions must be given to these patients about when to seek emergency care (Supplemental Table S12). It is important to note that flecainide and propafenone should not be used in patients with structural heart disease, including a history of ischemic heart disease.

9.1.2.2 Electrical cardioversion

In patients with hemodynamically stable recent-onset AF in whom sinus rhythm restoration is desired, there is a wealth of observational data supporting the safety and efficacy (approximately 90%) of direct-current cardioversion (DCCV).^{[363],[531],[533],[535],[564]–[567]} Since the late 1990s patient sedation and cardioversion have typically been performed by emergency physicians in the Canadian ED setting.^{[531],[535],[565]} Adverse events attributable to DCCV such as bradyarrhythmia, acute HF, and skin burns are rare.^{[533],[566],[567]} The time to patient discharge using DCCV is shorter than when using pharmacological cardioversion.^[534],[568] Importantly, AF patients who receive DCCV in the ED rate their care as more effective compared with those who receive only rate control, however, their QOL scores at 30 days were not different than those treated with only rate control.^[569] Biphasic shocks are preferred over monophasic because less energy is required.^[570] Pad placement (anterolateral vs anteroposterior) does not seem to influence cardioversion efficacy.^[538],[571] In obese patients, using paddles and applying force might improve success rates with DCCV over adhesive pads.^[572],[573] Pretreatment with antiarrhythmic drugs (eg, ibutilide and amiodarone) has been shown to improve the effectiveness of DCCV.^[363],[574]

9.2 Long-term rate control

9.2.1 Agents

Pharmacotherapy for long-term AF rate control revolves around agents with negative dromotropic properties such as β-blockers and ND-CCBs (verapamil and diltiazem). The choice of a specific rate-controlling regimen should be on the basis of patient’s characteristics and the drug’s efficacy/side effect profile (Supplemental Table S10; Fig. 18).^[575] In patients without significant LV dysfunction (LVEF > 40%), β-blockers and ND-CCBs are first-line options. There are no randomized long-term data to support choosing a β-blocker over an ND-CCB. Several retrospective studies of AF patients have shown conflicting results when rates of hospital admission after using β-blockers vs CCBs were compared: one showed no difference whereas another showed that use of CCBs was associated with a higher rate of hospitalization compared with use of β-blockers.^[520],[576] In the longer-term, β-blockers might be more effective at slowing ventricular rates at rest and during exercise, however, their use is associated with a higher risk of adverse effects, notably fatigue and exercise intolerance.^{[577]–[580]} Moreover, there is emerging evidence suggesting that CCBs might have favourable dose-response characteristics for AF rate control vs β-blockers, such that they might be preferred in patients with a preserved LVEF and without another indication for a β-blocker.^{[522],[580]–[582]} Specific patient characteristics might favour the use of one pharmacological class (eg, ND-CCBs with hypertension or reactive airway disease, vs β-blockers with CAD). Caution should be used when β-blockers are used with ND-CCBs. In patients with significant LV systolic dysfunction (LVEF ≤ 40%), maximally tolerated doses of evidence-based β-blockers (extended-release metoprolol succinate, bisoprolol, carvedilol) remain first-line therapy for rate control, although the benefits of adrenergic blockade, in addition to that provided by the control of the ventricular response rate, are uncertain.^{[583]–[588]}

Monotherapy with digoxin is generally ineffective in younger patients because of its inability to control ventricular rate during exertion or stress. Moreover, digoxin has a narrow therapeutic window, with observational evidence suggesting potential harmful effects when used for ventricular rate control.^{[589]–[592]} However, a recent meta-analysis of 28 trials of digoxin for AF rate control showed no increase in all-cause mortality vs control intervention (RR, 0.82; 95% CI, 0.24-11.5).^[593] As such, digoxin remains a reasonable choice for selected older or sedentary patients with HF and for those with inadequate rate control while receiving maximally tolerated doses of a β-blocker/ND-CCB. Although there is no direct evidence to support digoxin concentration monitoring for AF rate control, it might be reasonable to monitor patients at risk of digoxin-related adverse events (eg, female sex with low body weight and impaired renal function), at the clinician’s discretion, aiming for trough levels between 0.5 and 0.9 ng/mL.^[594] Furthermore, it is important to note that coadministration of ND-CCBs and amiodarone will decrease digoxin clearance, resulting in a propensity toward toxicity.

Amiodarone is a class III antiarrhythmic with complex pharmacological properties and potential serious adverse effects. However, selected patients such as the critically ill or those with side effects from, or contraindication to, first-line agents might benefit from amiodarone for rate control after careful consideration of alternative agents, alternative approaches (eg, transition to rhythm control), and risk/benefits of continued amiodarone therapy.^[595],[596]

9.2.2 Targets

The goals of ventricular rate control are the reduction of AF-related symptoms and the prevention of adverse cardiovascular events, rather than the achievement of a specific heart rate target. It is known that overly aggressive rate control is associated with adverse outcomes (eg, risk of symptomatic bradycardia with subsequent pacemaker implantation) and increased frequency of medical encounters. Conversely, overly lenient rate control might lead to HF (eg, tachycardia-mediated cardiomyopathy). In addition, specific populations might need stricter HR targets (eg, patients with cardiac resynchronization therapy [CRT], HF, tachycardia-mediated cardiomyopathy, mitral stenosis, stable angina), whereas others might do well with a more lenient target. As such, the intensity of AF rate control beyond the target HR of ≤ 100 bpm should be individualized on the basis of clinical characteristics and coexisting cardiovascular diagnoses. Most of the evidence to guide clinical decision-making for ventricular rate control targets has been acquired in patients with preserved LVEF.^{[512]–[514],[598]} Previous guidelines have recommended heart rate targets of < 80 bpm at rest and < 110 bpm with exercise, because these targets were used in the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study.^[512] However, retrospective analyses of the Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation (RACE) and AFFIRM studies suggested that cardiovascular morbidity, mortality, and QOL did not differ between those achieving or not achieving the prespecified heart rate target (but still maintaining a resting heart rate < 100 bpm).^[599] The prospective randomized Ratecontrol Efficacy in Permanent Atrial Fibrillation: a Comparison between Lenient Versus Strict Rate-control-II (RACE-II) trial598 showed that lenient rate control (resting heart rate target < 110 bpm) was noninferior to strict rate control (resting heart rate target < 80 bpm, < 110 bpm with exercise), with fewer medications, lower medication doses, fewer adverse events, and reduced health care utilization. However, it is important to recognize that the mean heart rates achieved were 76 ± 14 bpm in the strict group and 85 ± 14 bpm in the lenient group, with very few of those patients randomized to lenient rate control having resting heart rate of > 100 bpm.^[600] As such, there remain some questions as to the ideal heart rate target. A recent analysis of the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBITAF) registry and a retrospective analysis of the AFFIRM and Atrial Fibrillation and Congestive Heart Failure (AF-CHF) studies provides further insight. In both studies a U-shaped relationship between the resting heart rate and adverse outcomes was observed, with increased adverse event rates at the 2 extremes of resting heart rates. Although the combined AFFIRM/AF-CHF analysis suggested lower rates of mortality for those with resting heart rates in AF between 76 and 89 bpm, and lower rates of hospitalization for those with resting heart rates in AF between 76 and 114 bpm, the optimal heart rate in the ORBIT-AF registry was noted to be approximately 65 bpm.^[601],[602] On the basis of the best available evidence, a heart rate-target of < 100 bpm at rest appears to be associated with an acceptable risk/benefit profile in patients without significant LV systolic dysfunction (LVEF > 40%). Because of the paucity of data to guide heart rate targets in patients with significant LV systolic dysfunction it is reasonable to target a heart rate of < 100 bpm, although clinically driven targets should be individualized to symptoms and hemodynamics as per in other populations.

9.2.3 Atrioventicular junction ablation and pacing

Pharmacological agents achieve satisfactory heart rate control in most patients; however, a subset of patients fail to achieve adequate control of ventricular rate despite maximally tolerated/combination doses of rate-controlling agents. Implantation of a permanent pacemaker followed by AVJ ablation can be a useful treatment strategy to achieve definitive rate control in this population with permanent AF. An AVJ ablation strategy ensures reliable control of the ventricular rate, regularization of the RR intervals, and discontinuation of rate/rhythm control drugs in most patients.^[603] In patients with HF and CRT devices, AVJ ablation optimizes the delivery of biventricular pacing (see section 9.2.3.1). Compared with medical therapy, AVJ ablation and pacemaker implantation results in significant improvements in symptoms and QOL despite no significant changes in exercise capacity or functional status (eg, treadmill exercise or VO2 max).^[604],[605] In general the greatest improvement in QOL is observed in patients who are able to discontinue rate-limiting pharmacotherapies.^[606] It is important to note that these studies were performed in patients with permanent AF refractory to pharmacological rate control. As such, there are no data to support AVJ ablation and pacemaker implantation as a first-line treatment (ie, prior to attempts at pharmacological rate control). Patients with paroxysmal/persistent AF should be considered for a rhythm control strategy including catheter ablation (section 9.4) prior to pursuing AVJ ablation given the fact that AVJ ablation is irreversible, rendering patients pacemaker dependent, further complicating the management of device infection and lead failure, and increasing the risk of pacing induced cardiomyopathy.^[603] In patients with permanent AF who undergo AVJ ablation, a single-chamber right ventricular pacemaker is preferred for those with normal ventricular function. In those with normal ventricular function CRT implantation before AVJ ablation has not been shown to result in substantial benefit over the single-chamber right ventricular pacemaker. In a meta-analysis of 4 trials comparing denovo CRT vs right ventricular pacemaker in patients with AF treated with AVJ ablation, CRT therapy resulted in only a small improvement in QOL (Minnesota Living with Heart Failure Questionnaire, 2.72 fewer points; 95% CI, 1.45-3.99) and LVEF (þ2.6%; 95% CI 1.69%-3.44%), at the expense of a trend toward increased procedure-related complications (RR, 1.96; 95% CI, 0.71-5.45; P ¼ 0.2).^[607] Although there are emerging data on leadless pacemaker implantation and conduction system pacing (His bundle/left bundle branch area pacing) in the context of AVJ ablation, further studies are required before these strategies can be routinely recommended in preference to conventional single chamber right ventricular pacemakers.^{[608]–[610]} For patients with LV systolic dysfunction being considered for AVJ ablation for permanent AF, a CRT device is preferable to a single-chamber right ventricular pacemaker. In susceptible patients chronic right ventricular pacing might lead to progressive LV dysfunction and HF. Biventricular pacing has been shown to significantly lower incidence of death or HF hospitalization (HR, 0.74; 95% CI, 0.60-0.90) compared with right ventricular pacing in patients with LVEF ≤ 50%.^[611]

9.2.3.1 HF, AF, and biventricular devices

HF and AF often coexist, with up to 25% of patients with a CRT device having coexisting AF.612 In this population, the outcomes of CRT are suboptimal with a larger proportion of “non-responders” and a higher mortality rate vs patients in sinus rhythm.^[603] Therapeutic failure is largely due to the irregularity of the AF interfering with the ability of the device to deliver optimal CRT.^[613],[614] Management options to optimize CRT include increasing the lower rate limit, uptitrating atrioventricular nodal-blocking drugs, and/or AVJ ablation. A meta-analysis of 6 studies that included 768 patients with AF, symptomatic HF, LVEF  35%, and a CRT device reported that AVJ ablation was associated with lower all-cause mortality (RR, 0.42; 95% CI, 0.26-0.68; P < 0.001), lower cardiovascular mortality (RR, 0.44; 95% CI, 0.24-0.81; P ¼ 0.008), greater improvement in NYHA class (0.34; 95% CI, 0.56 to 0.13; P ¼ 0.002), and a reduction in appropriate/ inappropriate defibrillator shocks, compared with medical therapy for ventricular response rate control.^[615],[616] These potential benefits of AVJ ablation have to be weighed against the implications of rendering a patient pacemaker dependent. The Cardiac Resynchronisation Therapy and AV Nodal Ablation Trial in Atrial Fibrillation Patients (CAAN-AF; NCT01522898) and the Resynchronization/Defibrillation for Ambulatory Heart Failure Trial in Patients With Permanent AF (RAFT-Perm AF; NCT01994252) trials should provide additional insight into the role of CRT in patients with AF and HF.

9.3 Long-term pharmacologic rhythm control

9.3.1 Antiarrhythmic drugs

A strategy of sinus rhythm maintenance using long-term antiarrhythmic drug therapy is preferred for those with recently diagnosed AF (ie, within a year), and might be considered for other symptomatic patients with established AF (Fig. 16). Because long-term antiarrhythmic therapy might not completely suppress AF, the focus of rhythm control should be on symptom relief, improving functional capacity and QOL, and reducing health care utilization while balancing potential adverse drug effects. Moreover, a recent study showed that an initial rhythm control strategy for patients with recently diagnosed AF was associated with decreased cardiovascular mortality and a reduced incidence of thromboembolic events compared with rate control alone.^[515] Efficacy and safety data of common antiarrhythmic drugs used for rhythm control have been summarized in several systematic reviews.^{[617]–[619]} In a meta-analysis of 59 RCTs, the pooled recurrence rates of AF was 64%-84% at 1 year in control participants. Antiarrhythmic therapy reduced recurrence rates to 20%-50%. The most effective drug was amiodarone (OR, 0.22; 95% CI, 0.16-0.29 for recurrence vs placebo).^[617] Proarrhythmic events (ventricular or bradyarrhythmia) were significantly more frequent with sotalol (OR, 6.44; 95% CI, 1.03-40.24; P ¼ 0.047) and propafenone (OR, 4.06; 95% CI, 1.13-14.52; P ¼ 0.035), but were not significantly more frequent for flecainide (OR, 6.77; 95% CI, 0.85-54.02; P ¼ 0.067) or amiodarone (OR, 5.45; 95% CI, 0.69-42.93; P ¼ 0.095). Antiarrhythmic drugs have not been associated with a beneficial effect on mortality, and long-term use of sotalol and amiodarone have been associated with increased mortality (OR, 4.32; 95% CI, 1.59-11.70; P ¼ 0.013 and OR, 2.73; 95% CI, 1.00-7.41; P ¼ 0.049, respectively).^[617] A single study showed stroke risk reduction with dronedarone use compared with placebo (OR, 0.69;95% CI, 0.57-0.84) however, this finding has not been confirmed by other studies.^[620],[621] The antiarrhythmic drug doses, contraindications, or precautions are summarized in Supplemental Table S11. The initial choice of antiarrhythmic therapy is primarily driven by safety and tolerability, because these agents have a relatively similar efficacy (Fig. 19). If the initial drug does not achieve the desired results, an alternative antiarrhythmic may be used or catheter ablation may be considered. When the decision is made to abandon pharmacologic rhythm control and favour a rate control strategy, the antiarrhythmic drug should be discontinued.

9.3.1.1 Flecainide and propafenone

Class Ic antiarrhythmic agents, such as flecainide and propafenone, are use-dependent sodium channel-blocking drugs.^[622] The concomitant use of AV nodal blocking agents is recommended with class Ic antiarrhythmic drug therapy because of the potential risk of AF organization into AFL, with the potential of 1:1 AV conduction and rapid ventricular rate. Class Ic agents should be avoided in patients with: (1) preexisting advanced AV block (second- or third-degree AV block) or significant conduction system disorders (left bundle branch block, or right bundle branch block when associated with left hemiblock); (2) LV systolic dysfunction (LVEF ≤ 40%); (3) significant LV hypertrophy; (4) severe hepatic or severe renal impairment (CrCl < 35 mL/min); and (5) ischemic heart disease (active ischemia or history of MI). Because the use of flecainide has been associated with increased mortality when administered to suppress ventricular ectopy in the context of recent MI,^[623] a formal ischemia assessment (eg, stress test) should be considered before initiation of class Ic antiarrhythmic drugs in patients older than 50 years of age or with significant atherosclerotic risk factors. In addition, it is reasonable to consider annual assessment of symptoms of CAD for patients receiving long-term class Ic antiarrhythmic use, with formal stress testing being performed if significant symptoms are present. An ECG should be performed at baseline and after initiation to monitor for PR and QRS interval prolongation. An increase in QRS duration > 25% compared with baseline increases proarrhythmia risk.^[624]

9.3.1.2 Sotalol

Sotalol is a drug with reverse use dependence I_kr inhibition and is also a β-blocker. At lower doses the β-blocker effects predominate whereas the class III effects emerge with higher doses. The major risk of sotalol is QT prolongation and TdP. Sotalol should be avoided in patients with: (1) preexisting QTc prolongation (congenital or acquired long QT syndromes); (2) high-degree AV conduction disorders; (3) severe renal impairment (dose adjustment required for CrCl 40-60 mL/min; avoid with CrCl < 40 mL/min); (4) significant LV systolic dysfunction (LVEF ≤ 40%); or (5) significant risk factors for TdP (eg, women aged older than 65 years who are receiving diuretics or those with renal insufficiency). An ECG should be performed at baseline and 48-72 hours after outpatient initiation to monitor for QT interval prolongation.

9.3.1.3 Amiodarone

Amiodarone is a multichannel blocker and a nonselective β-blocker. A loading regimen of 10-12 g is recommended and then maintenance of ≤ 200 mg daily. Amiodarone has a long half-life and long-term use increases the risk for numerous extracardiac toxicities that affect the skin, thyroid, pulmonary, liver, and neurological systems. It should not be used as a firstline therapy when another drug might be an option. Amiodarone should be avoided in patients with: (1) high-degree AV conduction disorders; (2) active hepatitis or significant chronic liver disease; (3) pulmonary interstitial abnormalities; (4) preexisting QTc prolongation (congenital or acquired long QT syndromes); (5) hypersensitivity to the drug components, including iodine; or (6) concomitant use of strong cytochrome P450 3A4 (CYP3A4) inhibitors (eg, ketoconazole, cyclosporin, clarithromycin, ritonavir). Surveillance investigations are recommended with the use of amiodarone (eg, liver and thyroid tests every 6 months, annual chest radiograph).^[625] Amiodarone inhibits CYP3A, CYP2C9, and P-glycoprotein drugs, which requires close monitoring and dose adjustment of other affected medications (eg, digoxin, VKA, HMG-CoA reductase inhibitors). Patients should be counselled to be diligent in use of sun protection because of photosensitivity and report any symptoms indicative of pulmonary toxicity (eg, persistent nonproductive cough), optic neuropathy (changes in visual acuity and decreases in peripheral vision), or hepatic injury.

9.3.1.4 Dronedarone

Dronedarone resembles amiodarone but removal of iodine and the addition of a methane-sulfonyl group results in shortening the half-life (approximately 24 hours) and less tissue accumulation.^[626] Dronedarone is the only antiarrhythmic drug shown to reduce hospitalization and cardiovascular mortality in patients with paroxysmal or persistent AF with at least 1 additional risk factor for death.^[620] Dronedarone should be avoided in patients with: (1) permanent AF; (2) HF with recent decompensation requiring hospitalization or LV systolic dysfunction (LVEF ≤ 40%), owing to the observed increase in mortality observed with dronedarone use in this population597; (3) high-degree AV conduction disorders; (4) patients with previous lung or liver injury related to previous use of amiodarone; (5) preexisting QTc prolongation; or (6) severe hepatic impairment. Because of the risk of hepatotoxicity it is recommended that liver function tests be performed every 3 months for the first year of use, then every 6 months thereafter.

9.3.2 Pill-in-the-pocket antiarrhythmic drug therapy

In selected patients with symptomatic, infrequent, longer-lasting episodes of AF, the use of intermittent class Ic antiarrhythmic therapy (“pill-in-the-pocket”) shortly after symptom onset to restore sinus rhythm might be an alternative approach to daily antiarrhythmic use.^{[557],[560]–[563]} Indications, contraindications, and monitoring details are presented in Supplemental Table S12.

9.3.3 Trial of cardioversion

Many patients with persistent AF can present with the insidious onset of vague symptoms, such as fatigue or decreased exercise tolerance. In this population it can be difficult to determine the contribution of AF to their clinical presentation. A trial of sinus rhythm using cardioversion can often be useful in this population with established AF to determine if these nonspecific symptoms are secondary to AF. If so, then rhythm control might be the preferred treatment strategy (Fig. 16).

9.4 Catheter ablation of AF

Catheter ablation of AF has emerged as an important therapeutic modality for this common arrhythmia. The goal of catheter ablation is to eliminate the triggers and substrate responsible for the initiation and maintenance of AF. Ablation is performed as a percutaneous procedure in which catheters are inserted through venous access into the heart using fluoroscopic and electroanatomical mapping systems to identify regions of interest. These regions can then be ablated using thermal energy, namely radiofrequency or cryothermy.^[627] Irreversible electroporation (pulsed electrical fields) is a recent modality that might achieve the same effect, without the risk of thermal energy-related complications.^[628] Most of the triggers for AF originate within the PVs of the left atrium.^[629] Therefore, the “cornerstone” of all AF ablation procedures is ablating around the PV to electrically isolate them (Fig. 20). PV isolation (PVI) for paroxysmal AF is associated with success rates ranging from 60% to 80% after 1 procedure. For persistent AF, single-procedure success rates are lower (50%-70%), and it is unclear if ablation beyond PVI is required.^[630] Although the success rates seem suboptimal, these are often determined on the basis of the total elimination of AF. However, in many cases substantial clinical benefit can be achieved without total elimination of AF.^[631] When considered quantitatively, AF ablation substantially reduces the burden of AF (ie, time spent in arrhythmia) by > 98%.^[627] The CABANA trial (2204 patients) is the largest trial of catheter ablation vs medical therapy.^[394] In this trial patients with previous antiarrhythmic drug failure (47%) as well as antiarrhythmic-naive patients (53%) were enrolled. Although in the intention to treat analysis the trial did not show a difference between ablation and medical therapy for the primary end point of death, stroke, bleeding, and cardiac arrest, there was a significant reduction in the secondary end point of mortality and cardiovascular hospitalization in the ablation group (HR, 0.83; 95% CI, 0.74-0.93; P ¼ 0.001). Moreover, ablation was associated with a significant reduction in AF recurrence (HR, 0.52; 95% CI, 0.45-0.60; P < 0.001) with the per-protocol analysis showing that ablation was associated with a significant reduction in the primary end point (HR, 0.73; 95% CI, 0.54-0.99; P ¼ 0.046).

9.4.1 Procedural considerations for catheter ablation of AF

AF ablation is a complex procedure that requires a high degree of operator expertise. Recent balloon-based technologies have helped to make the procedures shorter and more consistent, but there remains evidence that such procedures should only be performed by individuals with appropriate experience and technological support. Recent evidence has shown that the rate of complications is directly related to the number of procedures performed at an institution and/or by the operator.^[632],[633]

9.4.2 Catheter ablation of AF after a trial of antiarrhythmic drugs

The primary indication for AF ablation is to achieve rhythm control in patients in whom antiarrhythmic drug therapy has already failed. Two systematic reviews have been performed, one before (2272 patients in 17 RCTs)^[634] and one including CABANA (4464 patients in 18 RCTs).^[635] Overall, the quality of studies was high and risk of bias low. Most patients included in these trials were patients in whom at least 1 antiarrhythmic drug had already failed. Despite the large difference in sample size, outcomes were remarkably consistent with reasonably large reductions in cardiovascular hospitalizations (RR, 0.63; 95%CI, 0.46-0.87; P ¼ 0.01, and RR, 0.56; 95% CI, 0.39-0.81; P ¼ 0.0001, respectively) and AF recurrences (RR, 0.44; 95% CI, 0.31-0.61; P ¼ 0.001, and RR, 0.42; 95% CI, 0.33-0.53; P < 0.00001, respectively) among patients who underwent catheter ablation, compared with antiarrhythmic therapy. All-cause mortality was also reduced in both meta-analyses but was driven by studies that enrolled patients with HF and reduced EF; sensitivity analysis showed no statistically significant reduction in all-cause mortality when these studies were removed from the analysis (RR, 0.48; 95% CI, 0.23-1.01; P ¼ 0.05, and RR, 0.67; 95% CI, 0.23-1.99; P ¼ 0.47, respectively). In patients without HF, mortality rates were too low to assess the potential for mortality benefit. Stroke rates were very low across all study populations and no effect of catheter ablation was shown on the rate of stroke (0.68% vs 1.23% for medical therapy; RR, 0.56; 95% CI, 0.26- 1.22; P ¼ 0.14). Likewise, major bleeding was not significantly more frequent after catheter ablation (3.4% vs 2.5% for medical therapy; RR, 1.55; 95% CI, 0.83-2.91; P ¼ 0.17). One RCT specifically addressed QOL as the primary outcome in patients in whom previous antiarrhythmic drug therapy had failed.636 Similar to contemporary trials,^[631] it showed important improvements in AF-specific and general QOL driven predominantly by improvements in AF symptoms and AF-related hospitalizations.

9.4.3 Catheter ablation of AF as first-line treatment

Early intervention for AF can prevent progression to persistent AF and avoid some of the long-term risks of the arrhythmia including stroke and HF. Furthermore, studies have shown that success rates for catheter ablation are higher when used earlier in the disease.^[637] Therefore, several clinical trials have addressed whether AF ablation should be used as first-line therapy before the use of antiarrhythmic drugs. Three previous randomized studies suggested a benefit of first-line ablation over antiarrhythmic drugs but these were small.^{[638]–[640]} More recently the large randomized Early Agressive Invasive Intervention for Atrial Fibrillation (EARLY-AF) trial demonstrated that first-line cryoballoon catheter ablation resulted in a signficant reduction in arrhythmia recurrence and AF burden, and a significant improvement in quality of life relative to first line antiarrhythmic drugs.^[641]

9.4.4 Candidates for AF ablation

Similar to long-term antiarrhythmic drug therapy, the decision to pursue a strategy of sinus rhythm maintenance should be aimed primarily at reduction of patient symptoms to improve QOL and reduce health care utilization. Patients who have truly asymptomatic AF are not generally considered candidates for ablation, however ablation might be pursued in those in whom AF is thought to adversely affect LV function even in the absence of overt symptoms. A specific subgroup that deserves mention are patients with HF, in whom AF is a causative or major contributing factor. A number of studies and systematic reviews have shown clinically important improvements in HRQOL, exercise tolerance, and LV function associated with catheter ablation over pharmacological rate control or rhythm control strategies.^[642],[643] In addition, 2 RCTs, including the Catheter Ablation vs Standard Conventional Therapy in Patients With Left Ventricular Dysfunction and Atrial Fibrillation (CASTLE-AF) trial, have recorded reduced hospitalizations in patients with HFrEF who underwent catheter ablation compared with medical therapy.^[644],[645] A recent systematic review showed that catheter ablation in this population was associated with a statistically significant reduction in all-cause mortality (5.3% vs 7.9% for medical therapy; RR, 0.69; 95% CI, 0.54-0.88; P ¼ 0.003; 9 studies, 3576 patients).^[635] Although a low level of bias was seen, the mortality benefit was driven by a single study (CASTLE-AF) that exclusively enrolled patients with HFrEF.^[645] A recent observational study showed that catheter ablation was associated with a long-term reduction in all-cause mortality and HF rehospitalizations.^[646] There are no data showing a mortality benefit in patients with HF and preserved EF. The large multicentre Canadian randomized ablation-based Randomized Ablation-Based Atrial Fibrillation Rhythm Control vs Rate Control Trial in Patients With Heart Failure and High Burden Atrial Fibrillation (RAFT-AF) study has closed enrollment and is due to report later this year (NCT01420393). These data will provide considerable insight into the effect of catheter ablation on HF with preserved EF and HFrEF. First-line therapy might be considered in younger patients who wish to avoid the risks of long-term antiarrhythmic agent use. Patients might also have cardiac or noncardiac absolute or relative contraindications to antiarrhythmic drugs. Some patients with tachy-brady syndrome are unable to tolerate drug therapy because of bradycardia complications in the absence of a pacemaker. If the AF can be successfully ablated, then antiarrhythmic therapy and the need for permanent pacing might be avoided.^[647]

9.4.5 Catheter ablation of AFL

Although AF and AFL might coexist in the same patient, AFL is an arrhythmia that is distinct from AF. In contrast to the apparent disorganization of AF, AFL usually involves a single macroreentrant atrial circuit rotating around a large central functional or anatomic obstacle (valves, veins, scar). AFL can be classified as “typical” (cavo-tricuspid isthmus-dependent) or “atypical” (eg, non-cavo tricuspid isthmus-dependent right or left AFL). Atypical AFL arising from scar related to previous heart surgery (eg, atriotomy or prosthesis) or catheter ablation and is also known as “incisional” flutter. Catheter ablation of typical right AFL is preferred to pharmacological therapy because of the relatively high success rate and relatively low rate of periprocedural complications. Ablation is typically performed with the patient under conscious sedation, using the femoral approach, guided by fluoroscopy or 3-D electroanatomic mapping. The goal is to create a complete ablation line along the cavo-tricuspid isthmus between the tricuspid annulus and the inferior vena cava to achieve bidirectional electrical conduction block. This intervention has an excellent acute and long-term success rate 90%, which is more effective than pharmacological rhythm control.^{[400],[648],[649]} In addition, ablation improves QOL, reduces symptoms, and might also avoid the development of a tachycardia-mediated cardiomyopathy.^[648] The acute complication rate is 2.6%, most commonly related to vascular access problems.^[400] Major complications, like complete heart block or pericardial effusion, are uncommon.^[649] In patients with coexisting AF and AFL it is perfectly cromulent to consider a hybrid treatment approach, in which AFL ablation is performed to prevent recurrent AFL while antiarrhythmic drug therapy is used to control AF.^[650] This approach has been effective for patients who derive clinical benefit from class Ic antiarrhythmic treatment of AF but manifest recurrent AFL.

9.5 Arrhthmia surgery

Surgical treatment for AF can be accomplished as a “standalone” procedure or performed coincident with a planned surgical procedure (valve replacement/repair and/or CABG). A number of factors need to be considered when contemplating surgical AF ablation therapy, including the indication for the procedure (eg, the potential benefits of sinus rhythm), the likely efficacy of the procedure (considering local surgeon/institutional experience), and the safety of the procedure. Moreover, in the context of adjuvant surgical ablation (eg, surgical ablation at the time of another surgery) the type of concomitant cardiac surgery (eg, mitral valve, which necessitates left atriotomy, vs CABG surgery, which does not) needs to be considered. Recent studies have shown that LA ablation in addition to valve surgery and/or CABG was beneficial in terms of sinus rhythm maintenance.^[651],[652] In the first study, 224 patients who underwent CABG and valve surgery were randomized to LA cryoablation or control; 60.2% of cryoablation patients showed sinus rhythm according to Holter monitoring at 1 year, vs 35.5% (P ¼ 0.002) of patients without ablation.^[651] In the second study 63.2% of 260 mitral valve surgery patients randomized to ablation (either PVI alone or biatrial ablation [secondary randomization]) were in sinus rhythm at 1 year vs 29.4% of those randomized to no ablation (P < 0.001).^[652] A recent systematic review combined 9 small RCTs and showed a large effect on the maintenance of sinus rhythm at 12 months in 481 patients (OR, 10.41; 95% CI, 5.30-20.44) as well as beyond 12 months (OR, 11.61; 95% CI, 4.53-29.79; 4 studies, 154 patients).^[653] Although no heterogeneity was observed among these trials, no individual study randomized more than 95 patients. Despite the high reported rates of sinus rhythm maintenance after surgical ablation, the effect on other outcomes is controversial. Although there is evidence for an improvement in HRQOL outcomes with a surgical procedure that includes concomitant surgical AF ablation, in several RCTs the magnitude of improvement was observed to be similar to patients who did not undergo ablation at the time of cardiac surgery.^{[654]–[656]} Moreover, surgical ablation appears to have no effect on the incidence of late stroke/TIA (6 RCTs; OR, 1.01; 95% CI, 0.41-2.49; P ¼ 0.98) or long-term survival (15 RCTs; OR, 0.91; 95% CI, 0.59-1.41; P ¼ 0.67).^[657] A significant concern with surgical ablation relates to the possibility of increased postoperative morbidity secondary to the additional surgical procedure. In a large meta-analysis, concomitant surgical AF ablation was not associated with an increase in the incidence of perioperative morbidity (which included deep sternal wound infection, pneumonia, reoperation for bleeding, and renal failure) or perioperative ICU length of stay.^[657] Likewise, there was no difference in the rate of readmission within 30 days, short-term survival (< 30 days), or perioperative stroke/TIA (< 30 days).^[653],[657] However, as highlighted in the two aforementioned RCTs, surgical ablation was associated with a significant increase in the rates of pacemaker implantation (6% vs 1% in Budera et al.^[651] and 21.5% vs 8.1% in Gillinov et al.^[652]).

There has been recent interest in hybrid AF ablation, achieved through collaboration between surgeons and electrophysiologists. This procedure encompasses surgical ablation (usually minimally invasive) coupled with percutaneous endocardial ablation of any residual gaps 6-8 weeks later. Similar to all invasive procedures, centre expertise and operator experience are important determinants of success. Comparative meta analyses suggest a modestly improved freedom from antiarrhythmic drug use after hybrid AF ablation in comparison to traditional Cox-Maze procedures, but former approach is associated with a higher rate of complications.^[657],[658] Some have advocated for the hybrid approach to involve surgical bilateral PVI with LAA closure coupled with endocardial ablation protocols, although this approach has not been tested in an RCT. Often a bilateral thoracoscopic intervention is needed in this setting. Other hybrid approaches might include unilateral thoracoscopic PVI encircling box lesions and posterior LA wall epicardial ablation lesion sets without concomitant LAAO.^[657] Although stand-alone surgical ablation presently comprises a small percentage of AF ablations, the ability to offer these newer approaches in a minimally invasive fashion without sternotomy or cardiopulmonary bypass is potentially attractive. Nevertheless, these approaches will need to be assessed in outcome studies.