A. Newly Diagnosed Atrial Fibrillation
A. HEMODYNAMICALLY UNSTABLE PATIENT
If the patient is hemodynamically unstable—usually as a result of a rapid ventricular rate or associated cardiac or noncardiac conditions—hospitalization and immediate treatment of atrial fibrillation are required. Intravenous beta-blockers (esmolol, propranolol, and metoprolol) or calcium channel blockers (diltiazem and verapamil) are usually effective at rate control in the acute setting. Urgent electrical cardioversion is only indicated in patients with shock or severe hypotension, pulmonary edema, or ongoing MI or ischemia. There is a potential risk of thromboembolism in patients undergoing cardioversion who have not received anticoagulation therapy if atrial fibrillation has been present for more than 48 hours or is of unknown duration; however, in hemodynamically unstable patients the need for immediate rate control outweighs that risk. An initial shock with 100–200 J is administered in synchrony with the R wave. If sinus rhythm is not restored, an additional attempt with 360 J is indicated. If this fails, cardioversion may be successful after loading with intravenous ibutilide (1 mg over 10 minutes, repeated in 10 minutes if necessary).
B. HEMODYNAMICALLY STABLE PATIENT
If the patient has no symptoms, hemodynamic instability, or evidence of important precipitating conditions (such as silent MI or ischemia, decompensated heart failure, pulmonary embolism, or hemodynamically significant valvular disease), hospitalization is usually not necessary. In most of these cases, atrial fibrillation is an unrecognized chronic or paroxysmal condition and should be managed accordingly (see Subsequent Management, below). For new-onset atrial fibrillation, thyroid function tests and echocardiography to assess for occult valvular or myocardial disease should be performed.
In stable patients with atrial fibrillation, a strategy of rate control and anticoagulation is appropriate. This is true whether the conditions that precipitated atrial fibrillation are likely to persist (such as following cardiac or noncardiac surgery, with respiratory failure, or with pericarditis) or might resolve spontaneously over a period of hours to days (such as atrial fibrillation due to excessive alcohol intake or electrolyte imbalance). The choice of agent is guided by the hemodynamic status of the patient, associated conditions, and the urgency of achieving rate control. In the stable patient with atrial fibrillation, a beta-blocker or calcium channel blocker (orally or intravenously) is usually the first-line agent for ventricular rate control. In the setting of MI or ischemia, beta-blockers are the preferred agent. The most frequently used agents are either metoprolol (administered as a 5 mg intravenous bolus, repeated twice at intervals of 5 minutes and then given as needed by repeat boluses or orally at total daily doses of 25–200 mg) or, in unstable patients, esmolol (0.5 mg/kg intravenously, repeated once if necessary, followed by a titrated infusion of 0.05–0.2 mg/kg/min). If beta-blockers are contraindicated, calcium channel blockers are immediately effective. Diltiazem (10–20 mg bolus, repeated after 15 minutes if necessary, followed by a maintenance infusion of 5–15 mg/h) is the preferred calcium blocker if hypotension or LV dysfunction is present. Otherwise, verapamil (5–10 mg intravenously over 2–3 minutes, repeated after 30 minutes if necessary) may be used. Rate control using digoxin is slow (onset of action more than 1 hour with peak effect at 6 hours) and may be inadequate and is rarely indicated for use in the acute setting. Similarly, amiodarone, even when administered intravenously, has a relatively slow onset and is most useful as an adjunct when rate control with the previously cited agents is incomplete or contraindicated or when cardioversion is planned in the near future. Care should be taken in patients with hypotension or heart failure because the rapid intravenous administration of amiodarone may worsen hemodynamics. Amiodarone should not be used in the acute setting if long-term therapy is planned with other antiarrhythmic agents.
Up to two-thirds of patients experiencing acute onset (shorter than 36 hours) of atrial fibrillation will spontaneously revert to sinus rhythm without the need for cardioversion. If atrial fibrillation has been present for more than a week, spontaneous conversion is unlikely and cardioversion may be considered for symptomatic patients. Importantly, if the onset of atrial fibrillation was more than 48 hours prior to presentation (or unknown), a transesophageal echocardiogram should be performed prior to cardioversion to exclude left atrial thrombus. If thrombus is present, the cardioversion is delayed until after a 4-week period of therapeutic anticoagulation. In any case, because atrial contractile activity may not recover for several weeks after restoration of sinus rhythm in patients who have been in atrial fibrillation for more than 48 hours, cardioversion should be followed by anticoagulation for at least 1 month unless there is a strong contraindication. Patients without valvular heart disease, diabetes, hypertension, or other risk factors for stroke, may not require long-term anticoagulation.
If immediate cardioversion is not performed, adequate rate control can usually be achieved with beta-blockers or nondihydropyridine calcium channel blockers. In older patients, who often have diminished AV nodal function and relatively limited activity, modest rate control can often be achieved with a single agent. Many younger or more active individuals require a combination of two agents. Choice of the initial rate control medication is best based on the presence of accompanying conditions: Hypertensive patients can be given beta-blockers or calcium blockers (see Tables 11–9 and 11–7). Patients with CHD or heart failure should receive a beta-blocker preferentially, whereas beta-blockers should be avoided in patients with severe chronic obstructive pulmonary disease (COPD) or asthma. Long-term use of digoxin is associated with an increase in mortality in patients with chronic atrial fibrillation and is rarely indicated. In symptomatic patients, a resting heart rate of less than 80 beats/min is targeted. In asymptomatic patients without LV dysfunction, a more lenient resting heart rate of up to 100–110 beats/min is reasonable. Ambulatory monitoring to assess heart rate during exercise should be considered in all patients with a goal not to exceed maximum predicted heart rate (220 – age). Adequacy of rate control should be evaluated by recording the apical pulse rate both at rest and with an appropriate level of activity (such as after brisk walking around the corridor or climbing stairs).
For patients with atrial fibrillation, even when it is paroxysmal or occurs rarely, the need for oral anticoagulation should be evaluated and treatment initiated for those without strong contraindication. Patients with lone atrial fibrillation (eg, no evidence of associated heart disease, hypertension, atherosclerotic vascular disease, diabetes mellitus, or history of stroke or TIA) under age 65 years need no antithrombotic treatment. Patients with transient atrial fibrillation, such as in the setting of acute MI or pneumonia, but no prior history of arrhythmia, are at high risk for future development of atrial fibrillation and appropriate anticoagulation should be initiated based on risk factors. If the cause is reversible, such as after coronary artery bypass surgery or associated with hyperthyroidism, then long-term anticoagulation is not necessary.
In addition to the traditional five risk factors that comprise the CHADS2 score (heart failure, hypertension, age 75 years or older, diabetes mellitus, and [2 points for] history of stroke or TIA), the European and American guidelines recommend that three additional factors included in the CHA2DS2-VASc score be considered: age 65–74 years, female sex, and presence of vascular disease (Table 10–12). The CHA2DS2-VASc score is especially relevant for patients who have a CHADS2 score of 0 or 1; if the CHA2DS2-VASc score is greater than or equal to 2, oral anticoagulation is recommended, and if CHA2DS2-VASc score is 1, oral anticoagulation should be considered, taking into account risk, benefit, and patient preferences. Female sex is a relatively weak factor, however, and the European guidelines have removed it from their risk assessment, so that oral anticoagulation is indicated for men who are CHA2DS2-VASc of 2 and women who are CHA2DS2-VASc of 3. (The use of warfarin is discussed in the section on Selecting Appropriate Anticoagulant Therapy in Chapter 14.) Unfortunately, studies show that only about half of patients with atrial fibrillation and an indication for oral anticoagulation are receiving it, and even when treated with warfarin, they are out of the target INR range nearly half the time. One reason for undertreatment is the misperception that aspirin is useful for prevention of stroke due to atrial fibrillation. In the 2016 European guidelines, aspirin is given a class III A recommendation, indicating that it should not be used because of harm (and with no good evidence of benefit). Cardioversion, if planned, should be performed after at least 3–4 weeks of anticoagulation at a therapeutic level (or after exclusion of left atrial appendage thrombus by transesophageal echocardiogram as discussed above). Anticoagulation clinics with systematic management of warfarin dosing and adjustment have been shown to result in better maintenance of target anticoagulation.
Table 10–12.CHA2DS2-VASc Risk Score for assessing risk of stroke and for selecting antithrombotic therapy for patients with atrial fibrillation. ||Download (.pdf) Table 10–12. CHA2DS2-VASc Risk Score for assessing risk of stroke and for selecting antithrombotic therapy for patients with atrial fibrillation.
|CHA2DS2-VASc Risk Score || |
|Heart failure or LVEF ≤ 40% ||1 |
|Hypertension ||1 |
|Age ≥ 75 years ||2 |
|Diabetes mellitus ||1 |
|Stroke, transient ischemic attack, or thromboembolism ||2 |
|Vascular disease (previous myocardial infarction, peripheral artery disease, or aortic plaque) ||1 |
|Age 65–74 years ||1 |
|Female sex (but not a risk factor if female sex is the only factor) ||1 |
|Maximum score ||9 |
|Adjusted stroke rate according to CHA2DS2-VASc score |
|CHA2DS2-VASc Score ||Patients (n = 7329) ||Adjusted stroke rate (%/year) |
|0 ||1 ||0% |
|1 ||422 ||1.3% |
|2 ||1230 ||2.2% |
|3 ||1730 ||3.2% |
|4 ||1718 ||4.0% |
|5 ||1159 ||6.7% |
|6 ||679 ||9.8% |
|7 ||294 ||9.6% |
|8 ||82 ||6.7 % |
|9 ||14 ||15.2% |
|CHA2DS2-VASc score = 0: recommend no antithrombotic therapy |
|CHA2DS2-VASc score = 1: recommend antithrombotic therapy with oral anticoagulation or antiplatelet therapy but preferably oral anticoagulation |
|CHA2DS2-VASc score = 2: recommend oral anticoagulation |
Four DOACs—dabigatran, rivaroxaban, apixaban, and edoxaban—have been shown to be at least as effective as warfarin for stroke prevention in patients with atrial fibrillation and have been approved by the FDA for this indication (Table 10–13). These medications have not been studied in patients with moderate or severe mitral stenosis, and they should not be used for patients with mechanical prosthetic valves. The term “nonvalvular atrial fibrillation” is no longer used in the American or European guidelines since most patients with other types of valvular heart disease have been included in trials of DOACs, which are equally effective in these patients.
Table 10–13.Direct-acting oral anticoagulants for stroke prevention in patients with nonvalvular atrial fibrillation. ||Download (.pdf) Table 10–13. Direct-acting oral anticoagulants for stroke prevention in patients with nonvalvular atrial fibrillation.
| ||Dabigatran ||Rivaroxaban ||Apixaban ||Edoxaban |
|Class ||Antithrombin ||Factor Xa inhibitor ||Factor Xa inhibitor ||Factor Xa inhibitor |
|Bleeding risk compared to warfarin || |
Less intracranial bleeding
Higher incidence of gastrointestinal bleeding
Less intracranial bleeding
Higher incidence of gastrointestinal bleeding
Substantially lower risk of major bleeding
Less intracranial bleeding
Lower risk of major bleeding
Less intracranial bleeding
|Dosage || |
110 mg twice daily
150 mg twice daily
|20 mg once daily (give with food) ||5 mg twice daily ||60 mg once daily |
|Dosage adjustments ||75 mg twice daily for creatinine clearance1 15–30 mL/min (approved in the United States but not tested in clinical trials) ||15 mg once daily for creatinine clearance1 < 50 mL/min || |
2.5 mg twice daily for patients with at least two of three risk factors:
Age ≥ 80 years
Body weight ≤ 60 kg
Serum creatinine ≥ 1.5 mg/dL
30 mg once daily for creatinine clearance1 ≤ 50 mL/min
FDA recommends not to use if creatinine clearance1 > 95 mL/min
Dabigatran (studied in the RE-LY trial) is superior to warfarin at preventing stroke at the 150 mg twice daily dose, and it is noninferior at the 110 mg twice daily dose, although this dose is not approved for treatment of atrial fibrillation in the United States. Both doses result in less intracranial hemorrhage than warfarin but also in more gastrointestinal bleeding than warfarin. When oral anticoagulation with either dabigatran or warfarin was stopped for elective or emergency procedures or surgery in the RELY trial, the risk of bleeding was numerically lower with dabigatran than with warfarin. Dabigatran is 80% renally metabolized. The creatinine clearance should be calculated before initiating therapy. The lower dose of 75 mg twice a day is recommended for patients with creatinine clearances 15–30 mL/min, although clinical practice guidelines recommend avoiding any of the DOACs in patients with an estimated creatinine clearance less than 30 mL/min since these patients were excluded from the clinical trials. There is no widely available test to accurately measure the effect of dabigatran, although the aPTT is affected by dabigatran and a normal aPTT suggests little if any dabigatran effect. Patients may be converted from warfarin to dabigatran by stopping the warfarin and beginning dabigatran once the INR is less than or equal to 2.0, and this is a reasonable approach for transition from warfarin to any of the DOACs. Neither dabigatran nor any of the DOACs should be used in patients with mechanical prosthetic heart valves where the medications are less effective and riskier.
Rivaroxaban is noninferior to warfarin for stroke prevention in atrial fibrillation (in the ROCKET-AF trial). In that trial, rivaroxaban was compared to warfarin for stroke prevention for patients with high-risk features of thromboembolism, with half of the patients in the ROCKET-AF trial having a history of stroke. Rivaroxaban is dosed at 20 mg once daily, with a reduced dose (15 mg/day) for patients with creatinine clearances between 15 and 50 mL/min. It should be administered with food, since that results in a 40% higher drug absortion. Similar to dabigatran, there is substantially less intracranial hemorrhage with rivaroxaban than warfarin.
Apixaban is more effective than warfarin at stroke prevention while having a substantially lower risk of major bleeding (in the ARISOTLE trial) and a lower risk of all-cause mortality. The apixaban dosage is 5 mg twice daily or 2.5 mg twice daily for patients with two of three high-risk criteria (age 80 years or older, body weight 60 kg or less, and serum creatinine of 1.5 mg/dL or more). Apixaban is associated with less intracranial hemorrhage and is well tolerated. Apixaban was also shown to be superior to aspirin (and better tolerated) in the AVERROES trial of patients deemed not suitable for warfarin.
Edoxaban, 60 mg once a day, is noninferior to warfarin for stroke prevention with lower rates of major bleeding and lower rates of hemorrhagic stroke (studied in the ENGAGE-AF trial). Edoxaban carries a boxed warning in FDA labelleling that it should not be used in patients whose creatinine clearance is more than 95 mL/min because it is less effective in this population. The dose is decreased to 30 mg/day for patients whose creatinine clearance is less than or equal to 50 mL/min.
These four DOACs have important advantages over warfarin, and therefore they are recommended preferentially over VKAs. In practice, these medications are often underdosed. They should be used at the doses shown to be effective in the clinical trials as shown in Table 10–13. Even though labeled for “nonvalvular” atrial fibrillation, the DOACs are safe and effective for patients with moderate or severe valvular abnormalities, with the exception of moderate or severe mitral stenosis. In part because of lower rates of intracerebral hemorrhage, DOACs have particular advantage over warfarin in the elderly and the frail, including patients with history of falls. For patients who fall, oral anticoagulation should generally be used, except for patients who are suffering head trauma with falls.
There are some patients with atrial fibrillation, however, who should be treated with VKAs. These patients include those who have mechanical prosthetic valves, advanced kidney disease (creatinine clearance less than 25 mL/min), or moderate or severe mitral stenosis, and those who cannot afford the newer medications. Apixaban may be a reasonable option for patients with creatine clearance less than 25 mL/min, although the randomized trial data are limited. Patients who have been stable while receiving warfarin for a long time, with a high time in target INR range, and who are at lower risk for intracranial hemorrhage will have relatively less benefit with a switch to a newer medication. It is important to note, however, that most patients who have intracranial hemorrhage while taking warfarin have had a recent INR below 3.0, so that good INR control does not ensure avoidance of intracranial bleeding. One way to reduce bleeding for patients taking oral anticoagulants is to avoid concurrent aspirin, unless the patient has a clear indication, like recent MI or coronary stent. Even then, use of oral anticoagulant plus clopidogrel without aspirin, or with only a brief period of “triple” therapy and then discontinuation of aspirin, may be a reasonable approach, as has been shown in clinical trials comparing rivaroxaban and dabigatran with warfarin.
There are some important practical issues with using the DOACs. It is important to monitor kidney function at baseline and at least once a year, or more often for those with impaired kidney function. Each of the medications interacts with other medications affecting the P-glycoprotein pathway, like oral ketoconazole, verapamil, dronederone, and phenytoin. To transition patients from warfarin to a DOAC, wait until the INR decreases to about 2.0. Each of the medications has a half-life of about 10–12 hours for patients with normal kidney function. For elective procedures, stop the medications two to three half-lives (usually 24–48 hours) before procedures with low to moderate bleeding risk (ie, colonoscopy, dental extraction, cardiac catheterization), and five half-lives before procedures like major surgery. Discontinuation times should be extended in patients with impaired renal function, particularly with dabigatran. There are no practical tests to immediately measure the effect of the medications, although a normal aPTT suggests little effect with dabigatran, and a normal prothrombin suggests little effect with rivaroxaban. For rivaroxaban and apixaban, chromogenic Xa assays will measure the effect, but may not be readily available. For bleeding, standard measures (eg, diagnosing and controlling the source, stopping antithrombotic agents, and replacing blood products) should be taken. If the direct-acting medication was taken in the prior 2–4 hours, use activated oral charcoal to reduce absorption. If the patient is taking aspirin, consider platelet transfusion. Antidotes should be considered for life-threatening bleeding or for patients with need for immediate surgery, or both. For cardioversion, the DOACs appear to have similar rates of subsequent stroke as warfarin, as long as patients have been taking the medications and adherent for at least several weeks. Like with warfarin, there appears to be a 1.5- to 2-fold increased rate of bleeding associated with the use of aspirin in combination with the DOACs. Even patients with atrial fibrillation and stable coronary disease taking a DOAC at least 1 year from most recent coronary stent or coronary bypass surgery appear to have greater risk than benefit from the use of aspirin. Therefore, aspirin should not be used with the DOACs unless there is a clear indication, such as coronary stents or acute coronary syndrome within the prior year.
A patient with severe bleeding while taking dabigatran may be treated with the reversal agent idarucizumab, which is a humanized monoclonal antibody approved by the FDA for rapid reversal of the anticoagulation effects, for use in the event of severe bleeding or the need for an urgent procedure. This treatment is widely available in the United States. Andexanet alfa, an intravenous factor Xa decoy, is approved for reversal of factor Xa inhibitors. Four-factor prothrombin complex concentrate may partially reverse the effects of these agents. Due to the short half-life of the DOACs (10–12 hours with normal kidney function), supportive measures (local control, packed red blood cells, platelets) may suffice until the medication has cleared.
Each of the DOACs appears to be safe and effective around the time of electrical cardioversion. In each of these trials, and in one modest-sized prospective randomized trial of rivaroxaban that specifically addressed cardioversion, the rates of stroke were low (and similar to warfarin) with the DOACs when given for at least 3–4 weeks prior to cardioversion. An advantage of the DOACs is that when stable anticoagulation is desired before elective cardioversion, it is achieved faster than with warfarin.
Devices to exclude the left atrial appendage have been shown to protect against stroke, although they are not as effective as warfarin to prevent ischemic stroke; the most commonly used approved devices are the Watchman (United States and Europe) and Amulet (Europe), which are options for patients who are unsuitable for long-term anticoagulation.
B. RATE CONTROL OR RHYTHM CONTROL
After assessing stroke risk and initiating anticoagulation where appropriate, two main treatment strategies for long-term management of atrial fibrillation exist: rate control or rhythm control, although they are not mutually exclusive. Historic trials comparing these strategies (namely the AFFIRM and RACE trials) used pharmacologic rhythm control with antiarrhythmic medications, including amiodarone, which are known to be only moderately effective with significant potential side effects. Nonetheless, the strategy of rate control and long-term anticoagulation was associated with no higher rate of death or stroke compared to antiarrhythmic medication therapy in these trials. Of note is that exercise tolerance and quality of life were not significantly better in the rhythm control group. Rate control should therefore be considered background treatment in nearly all patients with atrial fibrillation, regardless of whether rhythm restoration is eventually pursued, and may be considered the primary treatment in patients with minimal to no symptoms related to atrial fibrillation, or elderly or frail patients who would derive little to no benefit from maintenance of sinus rhythm.
The decision to pursue rhythm control is often individualized, based on symptoms, the type of atrial fibrillation (paroxysmal or persistent), comorbidities (such as heart failure), as well as general health status. As first treatment, elective cardioversion following an appropriate period of anticoagulation (minimum of 3 weeks) is generally recommended in patients in whom atrial fibrillation is thought to be of recent onset or when there is an identifiable precipitating factor. Similarly, cardioversion is appropriate in patients who remain symptomatic from the rhythm despite efforts to achieve rate control.
In cases in which elective cardioversion is required, it may be accomplished electrically or pharmacologically. A number of factors influence the success of electrical cardioversion. Biphasic energy waveform and anteroposterior electrode placement provide superior effectiveness. Pharmacologic cardioversion with intravenous ibutilide may be used (1 mg over 10 minutes, repeated in 10 minutes if necessary) in a setting in which the patient can undergo continuous ECG monitoring for at least 4–6 hours following administration. Pretreatment with intravenous magnesium (1–2 g) may prevent rare episodes of torsades de pointes associated with ibutilide administration. In patients in whom a decision has been made to continue antiarrhythmic therapy to maintain sinus rhythm (see next paragraph), cardioversion can be attempted with an agent that is being considered for long-term use. For instance, after therapeutic anticoagulation has been established, amiodarone can be initiated on an outpatient basis (400 mg twice daily for 2 weeks, followed by 200 mg twice daily for at least 2–4 weeks and then a maintenance dose of 200 mg daily). Because amiodarone increases the prothrombin time in patients taking warfarin and increases digoxin levels, careful monitoring of anticoagulation and medication levels is required.
Other agents that may be used for both cardioversion and maintenance therapy include dofetilide, propafenone, flecainide, and sotalol. Dofetilide (125–500 mcg twice daily orally) must be initiated in hospital due to the potential risk of torsades de pointes and the downward dose adjustment that is required for patients with renal impairment. Propafenone (150–300 mg orally every 8 hours) should be avoided in patients with structural heart disease (CAD, systolic dysfunction, or significant LVH). Flecainide (50–150 mg twice daily orally) should be used in conjunction with an AV nodal blocking medication, especially if there is a history of atrial flutter and should be avoided in patients with structural heart disease. Sotalol (80–160 mg orally twice daily) should be initiated in the hospital in patients with structural heart disease due to a risk of torsades de pointes; it is not very effective for converting atrial fibrillation but can be used to maintain sinus rhythm following cardioversion.
In patients treated long-term with an antiarrhythmic agent, sinus rhythm will persist in 30–50%. The most commonly used medications are amiodarone, dronedarone, sotalol, propafenone, flecainide, and dofetilide, but the latter four agents are associated with a clear risk of proarrhythmia in certain populations; dronedarone has less efficacy than amiodarone, and amiodarone frequently causes other adverse effects. Therefore, after an initial presentation of atrial fibrillation, it may be prudent to determine whether atrial fibrillation recurs during a period of 6 months without antiarrhythmic medications. If it does recur, the decision to restore sinus rhythm and initiate long-term antiarrhythmic therapy can be based on how well the patient tolerates atrial fibrillation. The decision to maintain long-term anticoagulation should be based on risk factors (CHA2DS2-VASc score, Table 10–12) and not on the perceived presence or absence of atrial fibrillation as future episodes may be asymptomatic.
B. Recurrent and Refractory Atrial Fibrillation
1. Recurrent paroxysmal atrial fibrillation
Patients with recurrent paroxysmal atrial fibrillation are at similar stroke risk as those who are in atrial fibrillation chronically. Although these episodes may be apparent to the patient, many are not recognized and may be totally asymptomatic. Thus, extended continuous ambulatory monitoring or event recorders are indicated in those in whom paroxysmal atrial fibrillation is suspected. Long-term anticoagulation should be considered for all patients except in those who are under 65 years of age and have no additional stroke risk factors. For select patients with symptomatic but rare (a few times a year) episodes of atrial fibrillation, an effective treatment strategy is on-demand pharmacologic cardioversion, termed pill-in-the-pocket treatment. Patients without coronary or structural heart disease may be given flecainide (200–300 mg) or propafenone (450–600 mg) in addition to a beta-blocker or nondihydropyridine calcium channel blocker as a single dose at the onset of symptoms. It is recommended that the first such treatment take place in a monitored setting (eg, the emergency department or hospital) to evaluate safety and effectiveness. For more frequent, symptomatic arrhythmic episodes, daily antiarrhythmic agents are first-line therapy; however, they are not often successful in preventing all paroxysmal atrial fibrillation episodes and long-term tolerability is poor. However, dofetilide has been shown to be as effective as amiodarone in maintaining sinus rhythm in certain patients and does not have as many untoward long-term effects.
2. Refractory atrial fibrillation
Atrial fibrillation should generally be considered refractory if it causes persistent symptoms or limits activity despite attempts at rate or rhythm control. This is much more likely in younger individuals and those who are active or engage in strenuous exercise. Even in such individuals, combined use of a beta-blocker and a nondihydropyridine calcium channel blocker can prevent excessive ventricular rates, though in some cases may cause excessive bradycardia during sedentary periods. If antiarrhythmic or rate control medications fail to improve symptoms, catheter ablation of foci in and around the pulmonary veins that initiate and maintain atrial fibrillation may be considered. Pulmonary vein isolation is a reasonable therapy for individuals with symptomatic paroxysmal or persistent atrial fibrillation that is refractory to pharmacologic therapy and for select patients (younger than 65 years or with concurrent heart failure) as first-line therapy. The primary benefit of catheter ablation is an improvement in quality of life. In the CABANA trial, there was no difference in the primary endpoint of death, disabling stroke, serious bleeding, or cardiac arrest in patients randomized to catheter ablation versus medical therapy as first treatment for symptomatic atrial fibrillation. Ablation is successful about 50–70% of the time but repeat ablation may be required in up to 20% of patients. The procedure is routinely performed in the electrophysiology laboratory using a catheter-based approach and adverse event rates are low when performed by experienced operators. Surgical ablation can also be performed via a subxiphoid approach thorascopically, via thoracotomy, or via median sternotomy in the operating room as a stand-alone or adjunct procedure. Finally, in symptomatic patients with poor rate control and deemed inappropriate for pulmonary vein isolation, radiofrequency ablation of the AV node and permanent pacing ensure rate control and may facilitate a more physiologic rate response to activity, but this is used only as a last resort.