Clinical review: Clinical management of atrial fibrillation – rate control versus rhythm control

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia in the critically ill and is associated with adverse outcomes. Although there are plausible benefits from conversion and maintenance of sinus rhythm (the so-called 'rhythm-control' strategy), recent randomized trials have failed to demonstrate the superiority of this approach over the rate-control strategy. Regardless of approach, continuous therapeutic anticoagulation is crucial for stroke prevention. This review addresses the findings of these studies and their implications for clinical management of patients with atrial fibrillation.


Introduction
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia in the community [1]. It is characterized electrocardiographically by irregular fibrillatory waves, usually associated with an irregular ventricular response, which manifests clinically as an irregular pulse. The presence of rapid, uncontrolled AF may be associated with severe symptoms and haemodynamic compromise, necessitating urgent intervention. In addition to its direct haemodynamic effects, AF is associated with a prothrombotic state and is a major risk factor for stroke and thromboembolism [2]. Overall, this arrhythmia also appears to be an independent predictor for death [3].
In addition, AF is the most common arrhythmia in postoperative patients [4] (particularly following cardiac surgery [5]) and in critically ill patients [6]. In these patients, as with patients in the community, AF is associated with adverse outcomes [7]. Hence, regardless of the clinical setting, AF identifies patients at substantial risk for morbidity and mortality.

Classification of atrial fibrillation
Recent guidelines suggested that AF be classifed on the basis of the temporal pattern of the arrhythmia [8]. AF is considered recurrent when a patient develops two or more episodes. These episodes may be paroxysmal if they terminate spontaneously (defined by consensus as 7 days) or persistent if electrical or pharmacological cardioversion is required to terminate the arrhythmia. Successful termination of AF does not alter the classification of persistent AF in these patients. Longstanding AF (defined as over 1 year) that is not successfully terminated by cardioversion, or when cardioversion is not pursued, is classified as permanent.
Regardless of the eventual classification, patients with AF should be assessed for symptomatic and haemodynamic compromise, which will guide subsequent management (Fig. 1), identification and correction of associated comorbidities and/or precipitants, and assessment of the patient's thromboembolic risk (Fig. 2)

Atrial fibrillation with haemodynamic compromise
The haemodynamic compromise in AF may result from loss of atrial contribution to ventricular filling (and therefore preload and stroke volume) and/or from rate and irregularity of the ventricular response. Under normal circumstances, atrial contraction contributes 20-30% of ventricular stroke volume. This atrial contribution increases with age and in conditions associated with impaired ventricular relaxation, such as hypertensive heart disease and hypertrophic cardiomyopathy. Consequently, loss of this atrial contribution may result in more considerable haemodynamic insult in these patients. The irregularity of the ventricular response and rate-related shortening of the diastolic filling interval results in further reduction in cardiac output [10,11].
Early cardioversion may be necessary in patients with evidence of haemodynamic compromise (acute pulmonary oedema, worsening angina, or hypotension) in relation to uncontrolled AF. Synchronized, direct current cardioversion is more effective and preferable to pharmacological cardioversion under these circumstances. Successful cardioversion depends on the energy used, the output waveform, paddle configuration, and the presence of underlying heart disease and co-morbidities. An initial energy of at least 200 J is recommended, although lower energies may be sufficient for Treatment algorithm for atrial fibrillation. a Consider anticoagulation and early cardioversion if symptomatic. b Refer to Table 5. INR, international normalized ratio; LV, left ventricular. devices that deliver biphasic waveforms. The anteroposterior paddle position is also associated with greater probability of success.
The underlying precipitant or contributory factors (Table 1) should also be addressed at the same time because attempts at cardioversion and maintenance of sinus rhythm without correcting the underlying precipitant are futile or even detrimental to the patient. There is evidence that atrial mechanical activity may not recover concurrently with electrical activity ('atrial stunning') after cardioversion. The lack of atrial contractility predisposes to the development of new thrombi and risk for thromboembolic complications [12]. As such, anticoagulation (heparin in the acute setting) should be administered in the absence of contraindication and continued for at least 4 weeks following cardioversion. This, however, should not delay immediate cardioversion of haemodynamically compromised patients [8].

Atrial fibrillation without haemodynamic compromise: rate control or rhythm control
The necessity for cardioversion is less well established in haemodynamically stable patients with AF. Despite this, the management of these patients has traditionally been dominated Available online http://ccforum.com/content/8/4/271

Figure 2
Risk stratification for anticoagulation treatment. INR, international normalized ratio. by a drive to restore and maintain sinus rhythm -the socalled 'rhythm-control' strategy. Better exercise tolerance, quality of life, improved survival and lower risk for stroke with eventual discontinuation of anticoagulation have been cited as the rationale for this approach [13]. However, limited efficacy and adverse effects associated with the use of antiarrhythmic therapy represent serious drawbacks to this approach. The 'rate-control' approach offers an alternative strategy by employing simpler and generally less toxic ratelowering drugs, but this is intuitively less appealing and inconvenienced by the need for close attention to anticoagulation ( Table 2).
The first of four studies (Table 3) [14][15][16][17] that compared the two management strategies, the Pharmacological Intervention in Atrial Fibrillation (PIAF) trial [14] recruited 252 patients and randomly assigned them to either the rate-control or rhythmcontrol arm. Diltiazem and the class III antiarrhythmic agent amiodarone were the main agents used in the former and latter groups, respectively. After 12 months of follow up, 10% in the rate-control group and 56% of the rhythm-control group were in sinus rhythm, and rhythm control was associated with better exercise tolerance (as assessed by 6-min walk test) but increased number of hospitalizations. There was no difference in quality of life between the groups.
In the Strategies of Treatment in Atrial Fibrillation (STAF) trial [15], 200 patients were randomly assigned to either rate control or rhythm control, and were followed up for a mean of about 20 months. Like the PIAF trial, patients in the rhythmcontrol arm were hospitalized significantly more often, usually for repeated cardioversion or antiarrhythmic therapy. However, that study was limited by lower than expected event rates, with nine primary end-points (a composite of death, cerebrovascular event, cardiopulmonary resuscitation and systemic emboli) in the rhythm-control arm, as compared with 10 in the rate-control arm (although this difference was not statistically    [16] tested the hypothesis that rate control was not inferior to rhythm control. In total, 522 patients were randomly assigned to either rate control with digitalis, nondihydropyridine calcium channel blocker and/or β-blocker, or to rhythm control. Regimens of sotalol, followed by flecainide or propafenone, and then amiodarone were used in a stepwise algorithm to maintain sinus rhythm. Anticoagulation could be discontinued if sinus rhythm was maintained for at least 1 month. After more than 2 years of follow up, sinus rhythm was maintained in 39% of the patients in the rhythm-control group as compared with 10% in the rate-control group, with no significant difference in the primary composite end-point of death from cardiovascular causes, heart failure, thromboembolic complications, bleeding, need for pacemaker implantation and serious adverse events from antiarrhythmic therapy. This suggested that rate control is not inferior to rhythm control. The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study [17] was the largest study to date comparing these two treatment strategies. In total, 4060 patients were enrolled in the study and followed up for a mean of 3.5 years. Digoxin, β-blockers and calcium channel blockers were used in the rate control arm, and amiodarone and sotalol were the most commonly used antiarrhythmic agents in the rhythm-control arm. At 5 years, about 35% of the rate-control group were in sinus rhythm, as compared with about 60% of those in the rhythm-control group. There was no significant difference in the primary outcome of overall mortality but patients in the rhythm-control group were significantly more likely to be hospitalized and suffer adverse drug effects. The majority of strokes occurred when anticoagulation either was stopped or was subtherapeutic.
Taken together, these studies suggest that the rate-control strategy would be an acceptable primary approach to patients with recurrent, persistent AF. Although there are clear differences in the patient populations studied, there is no evidence to suggest that cardioversion of AF in critically ill patients in the absence of haemodynamic compromise is associated with a better outcome. Therefore, restoration of sinus rhythm should no longer be deemed imperative in asymptomatic and haemodynamically stable patients. The choice of rate-lowering drug, however, may vary between patient populations. An approach to the management of newly diagnosed AF is outlined in Fig. 1.

Pharmacological rate control
The aims of heart rate control are to minimize symptoms associated with excessive tachycardia and to prevent tachycardia-associated cardiomyopathy. Historically, digoxin has been the pharmacological agent of choice but it has limited efficacy in patients who are in a hyperadrenergic state such as thyrotoxicosis, fever, acute volume loss, postoperative state and during exertion [18]. Digoxin monotherapy may therefore be of little value in the critically ill, although it may be adequate for the older, sedentary patient.
Other agents include β-blockers and the nondihydropyridine calcium channel blockers such as diltiazem and verapamil (Table 4). β-Blockers are effective in reducing ventricular rate at rest and on exertion (hyperadrenergic state) [19] and may be of additional benefit in patients with concomitant coronary artery disease. Diltiazem and verapamil are also effective ratelowering agents both at rest and during exercise [20]. These rate-lowering agents may therefore be more effective than Available online http://ccforum.com/content/8/4/271 digoxin in controlling ventricular rate during AF in the critically ill. Indeed, diltiazem has been shown to be superior to digoxin in controlling ventricular rate during acute AF [21].
Rate control, however, may not always be achieved with a single drug. In the study conducted by Farshi and coworkers [22], the combination of digoxin and atenolol was found to be more effective than the digoxin-diltiazem combination or monotherapy with digoxin, atenolol or diltiazem in controlling ventricular rate over 24 hours and during exertion. The combination of diltiazem and digoxin was also significantly more effective than digoxin monotherapy. Combination therapy should therefore be considered for AF that is uncontrolled with a single agent. The combination of diltiazem and digoxin is probably preferable to verapamil, in view of the latter's negative inotropic effect and potential interaction with digoxin.
β-Adrenergic receptor and calcium channel blocking classes of rate-lowering agents, however, are negatively inotropic and may precipitate or aggravate pulmonary oedema in patients with left ventricular dysfunction (Table 4). In addition, the associated blood pressure lowering effects may also limit their use in critically ill patients. Under these circumstances, amiodarone and digoxin are two pharmacological options, although β-blockers may be considered in stable heart failure. Studies suggest that amiodarone is haemodynamically well tolerated [23] and may be of at least equal efficacy in controlling ventricular rate as compared with diltiazem, with less hypotensive effect [24]. In addition, amiodarone is an effective agent for pharmacological cardioversion [25] (see below). Alternatively, a nonpharmacological approach (mainly atrioventricular node ablation coupled with pacing) can be considered. A detailed discussion on nonpharmacological interventions for AF is beyond the scope of this review.

Electrical and pharmacological cardioversion
With a number of the potential benefits (Table 2) apparently dispelled by the results of the studies described (Table 3), cardioversion -electrical or pharmacological -now appears less crucial. Nonetheless, restoration and maintenance of sinus rhythm should still be considered in certain groups of patients. Patients with symptomatic AF, particularly if symptoms persist despite rate control, for example, may be candidates for 'rhythm control'. It is also reasonable to consider (elective) cardioversion, with initial rate control and adequate anticoagulation in patients presenting with AF for the first time, particularly in those who are at low risk for recurrence ( Fig. 1 and Table 5).
Electrical or pharmacological cardioversion carries similar risks for thromboembolic complications. Therefore, a period (at least 3 weeks) of therapeutic anticoagulation is recommended before either form of cardioversion. Alternatively, transoesophageal echocardiography may be employed to guide cardioversion. The absence of thrombus in the left atrium (and appendage) suggests a low risk for thromboembolic complications [26]. Earlier cardioversion (with shorter term therapeutic anticoagulation) may be attempted in these cases.
Pharmacological cardioversion tends to be most effective for recent onset AF, which is generally defined as lasting for less than 1 week. Although a significant proportion of patients revert to sinus rhythm spontaneously within 48 hours, antiarrhythmic therapy increases the likelihood of cardioversion to up to 90% if administered early enough and in adequate doses [27]. Vaughan-Williams class Ia, Ic and III antiarrhythmic drugs are associated with increased conversion to sinus rhythm. One small study [28] suggested that intravenous amiodarone may be more effective than quinidine in restoring sinus rhythm, but a recent meta-analysis [29] failed to demonstrate superiority of one drug class over another. Class Ic and III may be preferable to class Ia antiarrhythmic drugs, however, in view of their better safety profile (Table 5).
In a direct comparison study [30], amiodarone appeared superior to sotalol (another class III anti-arrhythmic agent) and propafenone (a class Ic agent) in maintaining sinus rhythm -  a finding that was supported by the more recent AFFIRM study [31].
All antiarrhythmic therapies carry with them potential toxicity and proarrhythmia, particularly in the presence of ischaemic and structural heart disease ( Table 6). These risks should be considered in the individualization of antiarrhythmic therapy.
Regardless of the approach (rhythm or rate control), it is now clear that thromboembolic prophylaxis should be considered in all patients with AF that is not due to reversible causes, and particularly in the presence of risk factors for stroke. Adequate anticoagulation with warfarin is defined as an international normalized ratio in the range 2.0-3.0, in the absence of prosthetic valves or rheumatic valvular heart disease (in which case the international normalized ratio should be maintained in the range 2.5-3.5) [7]. A risk stratification model to guide thromboembolic prophylaxis is outlined in Fig. 2 [32].

Conclusion
Recent prospective studies have shown that, in selected patients, rate control coupled with thromboembolic prophylaxis provides similar benefits to those with rhythm control. The choice of rate-control medication (digoxin, βblockers, calcium channel blockers or amiodarone) should be based on clinical assessment, which includes assessing the presence of underlying heart disease and contraindications.
Available online http://ccforum.com/content/8/4/271 Toxic effects with long-term use 30 mg/kg oral loading dose Dofetilide [33] 125-500 µg orally twice daily based QT interval prolongation, ventricular Class III agent for conversion and on renal function and QT c arrhythmias (in particular torsades de maintenance of sinus rhythm pointes), conduction disturbances also Risk for ventricular tachyarrhythmias recognized Not licensed for use in UK Ibutilide [34] Dependent on patient weight: ≥60 kg, As per dofetilide Intravenous equivalent of dofetilide 1 mg intravenous; <60 kg, 0.01 mg/kg Not licensed for use in UK intravenous AF, atrial fibrillation. Table 5 Predictors of likelihood of cardioversion success and recurrence

Predictor Comment
Duration of arrhythmia Shorter duration associated with higher rates of cardioversion Structural heart disease Valvular heart disease and cardiomyopathy associated with lower rates of cardioversion and higher recurrence rates Left atrial dimension Increased recurrence rates with large left atrial size Although rate control is still often based on digoxin administration, calcium channel or β-adrenergic receptor blockers are generally more appropriate and effective for patients without left ventricular dysfunction, particularly in patients in a hyperadrenergic state. The preferred options for pharmacological rate control in patients with heart failure are digoxin or amiodarone, although β-blockers may be considered for patients with stable heart failure. Nonpharmacological options can also be considered in these patients, typically by multiple ablations to the left atria (around the pulmonary veins) to restore sinus rhythm or specific ablation of the bundle of His with pacemaker implantation for rate control. Pharmacological cardioversion and maintenance of sinus rhythm should still be considered for recurrent, symptomatic AF.