Drug-induced QT-interval prolongation and proarrhythmic risk in the treatment of atrial arrhythmias Free

Abstract

Introduction

From the early twentieth century, drug therapy has played an important role in the management of atrial arrhythmias. Quinidine was the first antiarrhythmic used to successfully restore and maintain sinus rhythm in atrial fibrillation (AF). Subsequently, a large number of other drugs have become available. Although the efficacy of many of these agents is impressive, side effects are a frequent occurrence. Amongst the most worrying side effects are QT-interval prolongation and risk of proarrhythmia, including torsade de pointes (TdP).

Pharmacological treatment for atrial fibrillation

Pharmacological cardioversion of AF can be achieved using a number of drugs with different pharmacological properties, including disopyramide, procainamide, quinidine (all class Ia), flecainide, propafenone (both class Ic), dofetilide, ibutilide, sotalol, and amiodarone (all class III). Currently, the most commonly used drugs for chemical cardioversion are flecainide, sotalol, and amiodarone. Little difference is observed between the route of administration for cardioversion rates, although intravenous administration results in faster conversion. Indeed, in patients with recent onset AF, successful cardioversion is reported in up to 80% of cases with oral therapy, rising only to 90% with intravenous administration. ¹

Unfortunately, recurrence of AF is common, often requiring long-term drug therapy to improve maintenance of sinus rhythm. For most current antiarrhythmic agents, the relapse rate is at least 50% during the first year, ^2–5 although slightly better figures are seen with dofetilide ⁶ and amiodarone. ^{7 , 8} A number of studies have also demonstrated that flecainide and propafenone are effective drugs for preventing AF recurrence. ^9–11 The effectiveness of flecainide is comparable to quinidine, but with fewer side effects. ¹² In contrast, propafenone is more effective for maintenance of sinus rhythm than quinidine and as effective as sotalol. ^{13 , 14} Generally, however, class Ic drugs are preferred to class Ia drugs in view of their better safety profile. ^{12 , 13}

The success of electrical cardioversion for AF has been quoted as between 75 and 93%, although this depends on left atrial size and co-existing structural heart disease, and ultimately on the duration of AF. ^15–17 Where there is some concern about a successful restoration of sinus rhythm (for example, previous cardioversion failure or early recurrence of AF), concomitant amiodarone or sotalol can be used pre-cardioversion to improve the success of electrical cardioversion. ¹⁸ Such an approach is advocated by the ACC/AHA/ESC guidelines on AF management. ² The frequency of recurrence of AF after electrical cardioversion is high, and maintenance therapy with antiarrhythmic drugs such as amiodarone or sometimes β-blockers is somewhat useful to prevent AF relapses. ¹

β-blockers are very effective at controlling ventricular rate and also may reduce the risk of AF recurrence following successful cardioversion (whether spontaneous, pharmacological, or electrical) and are currently used as first-line prophylactic agents in paroxysmal AF. β-blockers have also been shown to reduce the frequency of post-operative AF, although sotalol (which also has class III effects) appears to be the most effective in this setting. As AF commonly coexists with hypertension or congestive heart failure, β-blockers may also be part of conventional therapy in such patients.

Rate-limiting, non-dihydropyridine calcium channel blockers (diltiazem, verapamil) are frequently used to optimize rate control where β-blockers are contraindicated or ineffective. An intravenous β-blocker (for example, esmolol or metoprolol) or rate-limiting calcium antagonists (diltiazem, verapamil) are indicated where urgent pharmacological rate control is required. Intravenous amiodarone is a useful alternative in situations where the administration of β-blockers or calcium antagonists is not feasible, such as in the presence of heart failure.

All current class Ia, Ic, and III antiarrhythmic drugs have significant side effects. This includes non-cardiovascular effects (e.g. pulmonary fibrosis and thyroid dysfunction with amiodarone), and of particular importance, the risk of life-threatening ventricular proarrhythmia including TdP in up to 5% of patients. ^{19 , 20} Most of these antiarrhythmic drugs prevent or terminate AF by altering the function of potassium or sodium channels within the atrial cells. Blockade of potassium channels may prolong ventricular repolarization — and hence, the refractory period — resulting in QT-interval prolongation. Given the risk of severe proarrhythmia, the safety profile of many current antiarrhythmic drugs is far from ideal.

Mechanisms of antiarrhythmic drug-induced QT prolongation and proarrhythmia

Blockade of ionic currents

The QT interval represents the cellular ventricular action potential and is the net result of co-ordinated function of various ionic currents. Na ⁺ and Ca ²⁺ inward currents are primarily responsible for the action potential upstroke and depolarization, whereas outward K ⁺ currents in combination with a reduction in depolarizing currents are predominantly responsible for the myocyte repolarization. Furthermore, the same outward K ⁺ currents lead to restoration of negative myocardial intracellular polarity at rest.

Very high membrane resistance and low current flow characterize the plateau phase of the action potential. Thus, abnormalities of the depolarizing and repolarizing currents can dramatically change the duration of the plateau and, therefore, the duration of the action potential. Drug-induced increases in depolarizing currents and/or decreases in repolarizing currents will prolong the ventricular action potential duration and thus the QT interval. ²¹ However, the repolarization phase of the action potential is especially important in QT-interval prolongation, resulting predominantly from attenuated outward movement of potassium ions.

A variety of different K ⁺ channel subtypes are found in the heart, some of which are presented exclusively in atria. The two main subtypes responsible for ventricular repolarization are the so-called ‘rapid’ ( I_Kr ) and ‘slow’ ( I_Ks ) potassium currents. Blockade of either of these delayed rectifier K ⁺ currents is associated with lengthening of the action potential. However, I_Kr is often more susceptible to drug effects that may manifest clinically as a prolonged QT interval and the emergence of other T- or U-wave abnormalities on the surface electrocardiogram (ECG). Moreover, pharmacological inhibition of the current I_Kr appears to be responsible for the proarrhythmic effect of antiarrhythmic, as well as non-antiarrhythmic, drugs.

Wang et al . ²² showed that an early ultra rapid component of the delayed rectifier ( I_Kur ) contributes significantly to repolarization of the human atrial action potential. As I_Kur is present in atrial (but not in ventricular) myocytes in man, it is a potential target for the development of drugs that prevent atrial re-entrant arrhythmias without a risk of ventricular proarrhythmia. ²³

Early afterdepolarization

The prolongation of repolarization may promote action potential instability with increased beat-to-beat variability of duration. Subsequently, this may result in activation of premature inward depolarization currents, known as an early afterdepolarization (EAD). ²⁴ EADs are generally considered to result from reactivation of the voltage-dependent Ca ²⁺ current with secondary depolarization of the cell. ²⁵ However, other mechanisms such as increased late sodium current and potassium blockade have also been proposed. ^{26 , 27} Regardless of underlying mechanism, when EADs have sufficient amplitude, they may trigger another action potential and promote triggered activity. As a result EADs, especially when accompanied by the presence of a notably increased dispersion of repolarization (see below), may induce re-entry and may be responsible for initiation of a tachycardia.

Dispersion of repolarization

Repolarization of ventricular cardiomyocytes is further complicated by temporal and spatial inhomogeneity of the action potential. Significant regional heterogeneity of the action potential profile and duration across the left ventricular wall can be demonstrated, reflecting different functional expression of ion channels of cells in different transmural regions of the left ventricular wall. ²⁸ For example, EADs are easily induced in a subset of myocardial cells from the mid-ventricular myocardium, known as M cells, and in the His-Purkinje network. ²⁹ In response to I_Kr blockade, M cells demonstrate more pronounced action potential prolongation compared with subendocardial or subepicardial cells. ²⁹ Consequently, this leads to differential recovery of activated cells and areas of functional refractoriness in the mid-myocardial layer, which may provoke re-entrant arrhythmia and TdP. It is possible that EADs may be responsible for initiation of a tachycardia, whereas dispersion of repolarization may be responsible for its perpetuation.

Separation of the epicardial action potentials from that of M cells during the plateau phase has been suggested to be represented on the ECG by the beginning of the upright T wave. ³⁰ Under normal conditions, this separation is gradual such that the precise start of the T wave is difficult to determine. Final epicardial repolarization is proposed to correspond with the peak of the T wave, whereas final repolarization of the M cells is consistent with the end of the T wave. Therefore, the descending limb of the T wave represents transmural dispersion of repolarization (TDR), although it has recently been suggested to be a marker of total dispersion of repolarization time. ³¹ This phenomenon may in part explain why the descending limb of the T wave is so vulnerable. For example, a premature electrical impulse might potentially provoke functional transmural re-entry leading to the development of TdP.

The time interval between the peak and the end of the T wave has been demonstrated to be a clinically useful index of TDR in assessing arrhythmic risk. ^32–34 Furthermore, the heterogeneous effect of a blocker or activator of an ion channel may exaggerate the differences in action potential shape and duration favouring development of re-entry arrhythmias.

Structural heart disease can reduce the repolarization reserve and thus alter impulse conduction. This is often associated with impaired function of K ⁺ and Ca ²⁺ channels and altered intracellular handling of ions, which may represent electrical remodelling, particularly characteristic of heart failure. The next result, once more, is evolving action potential prolongation. ³⁵ In fact, the presence of structural heart disease is well established as a potent risk factor for the development of drug-induced polymorphic ventricular tachyarrhythmia and TdP.

Antiarrhythmic agents and QT prolongation

The risk of proarrhythmia has been predominantly demonstrated with class Ia, class Ic, and class III agents. However, incidences strongly depend on the presence of predisposing conditions, in particular, high drug dosage and concomitant use of other medicines, which can prolong QT interval.

Class I antiarrhythmic agents

Selzer and Wray ³⁶ first described a ‘specific toxic effect of quinidine’ and introduced the term ‘quinidine syncope’. Later, it was demonstrated that such syncope was most frequently caused by quinidine-induced ventricular tachycardia (often TdP) with an incidence of 0.5–4.4%. ³⁶ Meta-analysis of randomized trials evaluating the role of quinidine in the maintenance of sinus rhythm after cardioversion from AF demonstrated that this drug was associated with a significant increase in mortality (2.9% in the quinidine group vs. 0.8% in the control group), possibly as the result of quinidine-induced proarrythmia. ³⁷

Class Ia drugs are known to induce pro-arrhythmogenic effects including both ventricular tachycardia and TdP. The risk of proarrhythmia appears to be lower with disopyramide than with quinidine. ³⁸ The frequency of flecainide (class Ic) associated ventricular proarrhythmia is low in the treatment of AF and only a few cases of serious proarrhythmia have been reported in patients who have no or minimal structural heart disease. ³⁹ However, in patients treated with flecainide or encainide after myocardial infarction, the mortality rate was higher than observed in the control group. ⁴⁰ Therefore, given the relatively high risk of proarrhythmia in patients with coronary artery disease or left ventricular dysfunction, class Ic drugs should not be used in this situation.

Interestingly, TdP can occur at low therapeutic or subtherapeutic doses of class Ia drugs. ⁴¹ This phenomenon is reported for most of the available class Ia agents, including quinidine, disopyramide, and procainamide. ⁴² Such trends do not occur with the class Ic agent flecainide, where the incidence of TdP appears to be dose dependent. ³⁹ One possible explanation for this discrepancy is that the blockade of sodium channels by class Ia drugs seems to have less QT-prolonging effect at higher drug concentrations. Although class Ic drugs have a less dramatic effect on repolarization and appear to be safer than class Ia agents in terms of life-threatening proarrhythmia, individual cases of TdP have also been reported for propafenone, flecainide, and ajmaline. ^{39 , 43–45}

It is important to remember that quinidine-like class Ic drugs can enhance atrioventricular node conduction with increasing ventricular response in patients with AF or flutter, predominantly due to inherent vagolytic properties. In some patients, particularly in the presence of renal impairment, these drugs can induce atrial flutter with 1:1 atrioventricular-nodal conduction and haemodynamic instability. Thus, the concomitant use of atrioventricular-nodal blocking agents, such as β-blockers or rate-reducing calcium antagonists, is recommended.

Class III antiarrhythmic agents

Potent blockade of I_Kr occurs with class III drugs (sotalol, amiodarone, and dofetilide), and with a number of other agents, and results in QT prolongation. ⁴⁶ Selective I_Kr potassium-blocking antiarrhythmic drugs ( Table  1 ), such as sotalol, prolong QT interval and induce TdP in a dose-dependent manner up to a plateau resulting from complete potassium current blockade. ^{44 , 45 , 47} The incidence of sotalol-induced TdP is 0.3% for a daily dose of 80 mg, ~1% for patients taking between 160 and 240 mg/day, and up to 5–7% for a daily dose of 480–640 mg. ^{45 , 48} The risk is much higher in women and those with renal or congestive heart failure, sustained ventricular tachycardia, and with concomitant use of diuretics and hypokalaemia. ⁴⁹ Of note, drugs that prolong ventricular repolarization by blocking outward potassium currents generally demonstrate a reverse rate dependence, showing greater effects at slower rather than higher stimulation rates.

While all class Ia drugs demonstrate comparable risk of TdP, there is a marked discrepancy in the frequency of serious proarrhythmia within class III antiarrhythmic agents. For example, despite similar effects on QT prolongation with sotalol and amiodarone, the incidence of TdP is much lower with amiodarone compared with sotalol. A literature review of the incidence of TdP with amiodarone found only 0.7% in 17 uncontrolled studies (2878 patients), and no proarrhythmia was reported in a further 7 studies (1464 patients). ⁵⁰ In the Canadian Amiodarone Myocardial Infarction Arrhythmia Trial (CAMIAT) and in the European Myocardial Infarction Amiodarone Trial (EMIAT), amiodarone treatment was associated with a proarrhythmia rate of < 1% — less than observed in the control groups and probably as a result of its multi-channel inhibitory effects. ^{51 , 52} Indeed, a recent meta-analysis of amiodarone trials proved that amiodarone significantly reduced the risk of arrhythmic death and resuscitated cardiac arrest in patients with heart failure or after myocardial infarction. ⁵³ The risk of TdP with amiodarone predominantly occurs in patients with other concomitant risk factors, such as hypokalaemia or bradycardia.

The relatively new class III antiarrhythmic drugs, dofetilide and ibutilide, also possess the risk of excess QT prolongation and developing TdP. For example, in one randomized, double-blinded trial, infusion of ibutilide resulted in polymorphic ventricular tachycardia in 8.3% of treated patients, whereas all 86 placebo-treated patients were free of this complication. ^{54 , 55} Similar to sotalol, dofetilide has a dose-dependent effect on QT prolongation and TdP. ⁵⁶ The incidence of TdP in the summary basis of approval was 0–10.5% depending on dose. ⁵⁶ As a rule, severe proarrhythmia occurred during the first 3 days of dofetilide treatment initiation.

New antiarrhythmics under development

As the life-threatening complications of available antiarrhythmic drugs are predominantly related to effects on ventricular electrophysiological modalities, the search for safer approaches has focused on identifying agents that specifically target the atria. Such targets include the ultra rapid delayed rectifying potassium current ( I_Kur ), which is partly blocked by drugs such as AZD7009 ⁵⁷ and AVE0118. ⁵⁸ In pre-clinical studies, these drugs caused minimal ventricular proarrhythmias and were effective for restoring sinus rhythm. However, all of these antagonize other channels and are therefore best described as ‘mixed ion-channel blockers’. Preliminary results have been encouraging, with a recent clinical study demonstrating the efficacy of intravenous AZD7009. Here up to 70% of patients converted to sinus rhythm from persistent AF. ⁵⁹ Although some patients exhibited QT prolongation, particularly at higher dosages, only one patient exhibited (asymptomatic) non-sustained ventricular tachycardia. Despite these promising results, AZD7009 development has been discontinued based on non-cardiovascular safety findings in clinical studies.

Given that the class III antiarrhythmic, amiodarone, was found to be effective and relatively safe in AF, a group of drugs each with some properties similar to amiodarone has been developed (e.g. dronedarone, azimilide, and tedisamil). These agents block outward potassium currents and consequently delay atrial and ventricular repolarization. Importantly, it was hoped that by blocking various other ion channels, these novel agents would retain the efficacy and safety (in terms of proarrhythmia) of amiodarone, but without the associated toxicity. There is some evidence, however, that this may not be the case with dronedarone. ⁶⁰

Azimilide blocks both the rapid ( I_Kr ) and slow ( I_Ks ) components of the delayed rectifier potassium current, distinguishing it from other potassium channel blockers such as sotalol, dofetilide, and ibutilide. Several randomized, placebo-controlled clinical trials have demonstrated the efficacy of azimilide in prolonging symptom-free interval in patients with AF or atrial flutter. ^{61 , 62} The preliminary results from the ALIVE (Azimilide Post-infarction Survival Evaluation) study have demonstrated that 100 mg of azimilide has a low incidence of TdP even in high-risk post-myocardial infarction patients with left ventricular systolic dysfunction. ^{63 , 64} Azimilide also reduced the risk of symptomatic AF recurrence by 40% compared with placebo, ⁶¹ although in other studies such efficacy was not demonstrated. This agent prolongs the QT interval by 4–42% at doses up to 200 mg/day. ⁶⁵ Infrequent, but serious adverse events, including severe neutropenia and TdP, were reported. These complications occurred in up to 1 and 1.5% of patients, respectively — predominantly in patients with bradycardia, pauses, or hypokalaemia. A recent analysis of cumulative evidence from 19 clinical studies in the azimilide database demonstrates that the risk of TdP with azimilide is low (1%) and less than that observed with selective I_Kr potassium channel blockers such as dofetilide and ibutilide. ⁶⁶

Dronedarone is structurally similar to amiodarone but lacking the iodine moiety — a feature of amiodarone that has been linked to many non-cardiac side effects (including pulmonary toxicity, ocular effects, thyroid disease, and hepatic dysfunction). The Dronedarone Atrial Fibrillation Study After Electrical Cardioversion (DAFNE) trial demonstrated the efficacy and safety of dronedarone in preventing AF recurrence after cardioversion in 199 patients. ⁶⁷ In the EURIDIS (European trial in atrial fibrillation or flutter patients receiving dronedarone for the maintenance of sinus rhythm) and ADONIS (American–Australian–African trial with dronedarone in atrial fibrillation or flutter patients for the maintenance of sinus rhythm) trials, dronedarone was effective in preventing AF recurrence and was shown to reduce the ventricular response during AF relapse. There was no evidence of proarrhythmia (including TdP), heart failure exacerbation, or thyroid, pulmonary, or other organ toxicity. The mortality rate was low (1.0%) and not significantly different from placebo (0.7%) during the 12-month follow-up. ⁶⁸ However, the Antiarrhythmic Trial with Dronedarone in Moderate-to-Severe Congestive Heart Failure Evaluating Morbidity Decrease (ANDROMEDA) was stopped prematurely because of a trend towards increased risk of death in the dronedarone group, but this numerical increase in mortality was not statistically significant. In 2006, the United States Food and Drug Administration issued a non-approvable letter based on safety concerns. Consequently, the drug manufacturer withdrew an application for licensing to the European Agency for the Evaluation of Medicinal Products.

Tedisamil was originally developed as an anti-anginal agent and Phase III studies in patients with coronary artery disease have demonstrated its efficacy in this setting. Additionally, tedisamil has been shown to have significant class III antiarrhythmic properties with multiple ion-channel effects. Its efficacy and safety in cardioverting AF was recently proved in a multicentre, double-blinded, randomized, placebo-controlled study. ⁶⁹ In this trial, 41 and 51% of patients receiving the lower and higher doses, respectively, cardioverted to sinus rhythm, with two cases (1.8%) of possible proarrhythmia (one TdP and one monomorphic ventricular tachycardia) observed.

Is antiarrhythmic drug-induced QT-interval prolongation a strong predictor of proarrhythmic potential?

The incidence of TdP is not proportional to the extent of QT prolongation. The highest rates of TdP are observed with use of selective I_Kr blockers, such as sotalol and dofetilide. These drugs preferentially prolong subendocardial and endocardial repolarization, leading to dose-dependent QT prolongation, together with a markedly pronounced increase in TDR. ^{70 , 71} Clinical use of these drugs is therefore associated with very high incidence of TdP. ^{72 , 73}

As already noted, blockade of sodium channels at higher concentrations of class Ia agents may attenuate repolarization delay and risk of TdP, implying a favourable effect from modulation of multiple ion channels. The effects of drugs that affect several ion channels on the development of TdP are complex. For example, amiodarone is a potent antiarrhythmic drug that can significantly prolong the QT interval, however, the risk of proarrhythmia is very low for amiodarone and it rarely induces TdP, even in those who have previously developed TdP as a complication of other QT-prolonging agents. ⁷⁴ The safety of amiodarone may be attributed to its multi-channel inhibitory effects, which include sodium, potassium, and calcium currents in the heart. Long-term use of amiodarone reduces TDR by producing a greater prolongation of action potentials in the epicardium. ⁷⁵ In addition, its ability to inhibit inward currents may prevent the occurrence of EAD. ⁷⁶ Still, there is some debate about a possible protective role of certain opposing inward currents such as late I_Na and I_Ca , which may not be inhibited until much higher amiodarone concentrations than those required to block I_Kr .

In summary, the development of TdP under QT-interval prolongation is mainly dependent on two prerequisites: the genesis of EAD that functions as a trigger and enhanced TDR that facilitates EAD propagation and serves as a functional re-entrant substrate to maintain TdP and is not directly related to the degree of QT-interval prolongation. Development of a drug with multiple points of action may prevent potentially dangerous disturbance of the electrophysiological myocardial balance, and thereby reduce undesirable proarrhythmic effects, even despite QT prolongation. For example, ranolazine, a novel anti-anginal drug with antiarrhythmic properties, has pharmacological effects attributable to the preferential blocking of the late sodium current ( I_NaL ) relative to peak sodium current ( I_Na ). Ranolazine demonstrates some action potential prolonging effects, ^{77 , 78} and studies have reported a small degree of QT prolongation in both experimental models ⁷⁷ and human subjects. ^{79 , 80} Multi ion-channel blockade, particularly the ability to block I_NaL , has been suggested as the underlying mechanism whereby this drug prolongs the QT interval without any apparent increase in TdP. ⁷⁷ It is important to note that ranolazine, even at high doses does not induce EADs, ⁸¹ and no TdP has been observed either in experimental models ^{77 , 82} or clinical studies. ^{79 , 80}

How do we minimize the risk of proarrhythmia of antiarrhythmic drugs currently in use?

Hypokalaemia, hypomagnesaemia, congenitally prolonged QT intervals, bradycardia, congestive heart failure, female sex, and pauses associated with the conversion of AF to sinus rhythm are known risk factors for development of TdP ( Table  2 ). The risk of QT prolongation and TdP may be increased further by concomitant use of a wide range of medications, especially those which interfere with the hepatic metabolism of antiarrhythmic drugs. ⁸³ It is therefore advisable that these drugs are not prescribed to patients taking other substances known to promote QT prolongation (such as erythromycin) and that a 12-lead ECG be requested after initiating treatment.

There are reports of ‘cross-reactivity’ when TdP developed in the same patient following administration of different antiarrhythmic agents. ‘Cross-reactivity’ was described with different class Ia and class III drugs. However, amiodarone has been used safely in patients with a history of drug-induced TdP. ⁷⁴

Some electrocardiographic findings can also indicate higher probability of TdP development. For example, QT-interval dispersion represents the difference between the maximum and the minimum measured QT interval on the same 12-lead ECG — whereas normal QT-interval dispersion is 50 ± 15 ms; in patients with TdP, the average QT dispersion is 100 ± 40 ms. ⁸⁴ QT dispersion is independent of length of QT interval and may serve as an independent risk predictor of drug-induced TdP. Of note, QT dispersion was demonstrated to be significantly less for amiodarone than sotalol and class Ia drugs, even despite similar absolute QT-interval prolongation. ⁸⁵

Initiation of TdP is typical when a pause (long RR interval) is noted on the ECG and then followed by premature ventricular contraction with markedly prolonged repolarization. This situation may predispose to EADs and triggered activity. ‘TU- or U-wave alternans’ is another risk factor for TdP, which is characterized by a significant beat-to-beat change in the TU axis.

Given the data with sotalol treatment, where TdP usually occurs within 5 days of initiation, ⁸⁶ it is probably safer to initiate sotalol in hospital where facilities for continuous ECG monitoring and cardiac resuscitation exist. The risk of TdP can be reduced by dose adjustment relative to creatinine clearance and by monitoring the ECG for excessive increases in the QT interval. Sotalol is contraindicated in patients with a QT interval > 450 ms or creatinine clearance < 40 mL/min.

Future directions

In conclusion, despite the large number of antiarrhythmic agents that are currently available, modern cardiology is still waiting for the introduction of new efficient and safe drugs for the treatment of atrial arrhythmias. The ideal antiarrhythmic agent must efficiently cardiovert AF patients and prevent relapses without proarrhythmic potential. To achieve this, it seems that such drugs should be atrial selective, should have multi ion-channel effects, should not increase TDR, should not produce EADs, and should not exhibit reverse use-dependency.

Conflict of interest : G.Y.H.L. has received funding for research, educational symposia, consultancy, and lecturing from different manufacturers of drugs used for the treatment of atrial fibrillation. He was Clinical Advisor to the Guideline Development Group that wrote the United Kingdom National Institute for Health and Clinical Excellence (NICE) Guidelines on atrial fibrillation management (www.nice.org.uk). E.S. is a European Society of Cardiology research fellow.

References