Antiarrhythmic Drugs for All Patients With an ICD?

In the early stages of implantable cardioverter defibrillator (ICD) use, the combination of antiarrhythmic drug concurrent with the device seemed redundant, akin to “a belt plus suspenders.” Before long it was discovered that, over time, most patients eventually required antiarrhythmic medication in addition to the ICD. In this issue of JAMA, the results of the Optimal Pharmacological Therapy in Cardioverter Defibrillator Patients (OPTIC) trial reported by Connolly and colleagues¹ raise the issue of whether antiarrhythmic therapy, plus a β-blocker, should be initiated at the time of initial ICD implantation in an effort to reduce the likelihood of ICD shock. To respond to this question, it is important to examine the interaction of antiarrhythmic drug therapy with the ICD, and the consequences of such a combination.

Antiarrhythmic medication is administered in patients with an ICD for a number of reasons. Most importantly, drug therapy can reduce or eliminate ICD shocks by suppressing ventricular arrhythmias, or by slowing ventricular tachycardia (VT) to such a degree that it can be terminated with programmed antitachycardia pacing. In addition to the suppression of such “appropriate” shocks, antiarrhythmic therapy may suppress the “inappropriate” shocks precipitated by supraventricular arrhythmias (primarily atrial fibrillation). Because ICD shocks have been associated with reduced physical capacity and impaired quality of life,^{2 - 3} suppression of such events may have important clinical benefits.

Antiarrhythmic agents may also be added to suppress symptomatic ventricular and supraventricular arrhythmias that are neither fast enough nor sustained to the degree that they cause the ICD to discharge. A further benefit is that, in some cases, these drugs may reduce the defibrillation threshold (defined as the energy required to defibrillate the ventricles) and as such allow for successful defibrillation when the defibrillation threshold is otherwise increased to a level that does not provide an adequate margin of safety (the difference between the defibrillation threshold and maximum output).

On the other hand, antiarrhythmic drugs can interfere with ICD function. Proarrhythmic effects of such drugs include torsades de pointes and stabilization of otherwise nonsustained VT. Furthermore, the cycle length of VT can become prolonged to a point that the heart rate no longer exceeds the programmed rate cutoff of the ICD, leaving the arrhythmia unrecognized by the device. In addition, the pharmacological agent may increase, rather than decrease, the defibrillation threshold, in the worst case rendering the ICD ineffective in defibrillating the ventricles.

Whenever an antiarrhythmic agent is administered to a patient with an ICD, the effect on device function must be considered. Both animal and clinical studies provide insights to help predict such interactions. As a drug class effect, the category III antiarrhythmic agents (potassium channel blockers), such as sotalol, tend to lower defibrillation threshold, although this finding is variable.^{4 - 6} In a human study,⁷ the defibrillation threshold was lowered acutely by 32% with the addition of D-sotalol (the pure class III enantiomer of D,L-sotalol). This property of lowering the defibrillation threshold may be used clinically to improve the defibrillation threshold when the margin of safety is insufficient.⁸ In contrast with potassium channel blocking drugs, the category IC sodium channel blocking agents (flecainide acetate, propafenone hydrochloride) have shown variable effects in animal and human defibrillation threshold.^{4 - 5} Amiodarone, which shares membrane properties of all the antiarrhythmic drug classes, has been shown to increase the defibrillation threshold in animal models and in some (but not all) human studies.^{4 - 5 ,9} Owing to the often unpredictable effects of antiarrhythmic drugs on ICD function and defibrillation threshold, some clinical electrophysiologists advocate testing of the ICD, and even induction of VT in patients with clinical arrhythmia, before discharging a patient with a new antiarrhythmic agent.⁴ In general, β-adrenergic blocking agents are considered to have a neutral effect on defibrillation threshold.

In routine practice, adjunctive antiarrhythmic therapy is administered to between 49% and 69% of patients who have an ICD.^{10 - 11} In the Multicenter Automatic Defibrillator Implantation Trial (MADIT) I study,¹² which compared standard therapy with the ICD, 46% of the patients enrolled in the ICD group were receiving drug therapy at last contact (mean follow-up, 27 months). In the Antiarrhythmics Versus Implantable Defibrillators (AVID) trial,¹³ in which the investigators were strongly cautioned not to add drugs to patients with ICD (because crossover would weaken the power to determine benefit of drug or ICD), 27% of patients with an ICD were receiving antiarrhythmic drug therapy after 2 years.

Patients receiving an ICD for primary prevention of cardiac arrest may require antiarrhythmic therapy less frequently. In the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) study,¹⁴ which compared prophylactic amiodarone, placebo, or ICD in patients with depressed left ventricular function, 14% of those patients receiving the ICD also received open-label amiodarone during part of the 46 months of follow-up. Similarly, in Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation (DEFINITE)¹⁵ and Defibrillator in Acute Myocardial Infarction Trial (DINAMIT),¹⁶ 2 recent studies of the ICD for primary prevention, amiodarone was added to the ICD therapy in only 3.1% and 8.1% of patients, respectively, during 29 to 30 months of follow-up. The addition of these agents to the therapeutic regimen is almost always reactive, with the drug being initiated in response to an ICD shock or some other symptomatic supraventricular or ventricular arrhythmia.

The effect of empirical therapy with antiarrhythmic drugs in patients with ICDs has been examined previously. The study by Pacifico et al¹⁷ randomized 302 patients (with history of symptomatic life-threatening ventricular arrhythmias and recent ICD implant or generator change) to sotalol therapy vs placebo to study the end point of death or ICD shock. They found a 48% reduction in shock or death (P<.001), with 34% of patients receiving sotalol reaching this end point at 12 months compared with 54% of patients receiving placebo. Inappropriate shocks were reduced by 64% (P = .004). A smaller randomized trial¹⁸ showed no difference between the frequencies of appropriate shocks in patients receiving metoprolol vs sotalol (38% vs 32%, respectively). The incidence of inappropriate shocks and total shocks were not reported.¹⁸

The OPTIC study¹ adds substantial insight into the effect of antiarrhythmic drugs on the frequency of ICD shock. The patients studied received the ICD primarily for secondary prevention, as 294 (71%) of 412 patients had demonstrated spontaneous VT or ventricular fibrillation while other patients had clinical syncope and positive electrophysiology study. A small group of patients had received ICD for depressed ejection fraction (40% or lower) and positive electrophysiology study (without documented sustained clinical ventricular arrhythmia or syncope), and as such could be considered to have received the ICD for primary prevention.

Consistent with current clinical practice, all patients received a β-blocker (alone, in combination with amiodarone, or as the intrinsic property of sotalol) in doses that should provide for similar antiadrenergic effects; this allows evaluation of the contribution of the cardiac ionic channel properties of sotalol and amiodarone. Furthermore, the investigators selected the 2 antiarrhythmic agents most commonly added to ICD therapy and allowed for comparison of these agents. Although enrollment numbers precluded the initial design for comparison of each antiarrhythmic agent with β-blocker alone, the data support the primary end point that administration of either sotalol or amiodarone plus a β-blocker results in less frequent ICD shocks than the β-blocker alone.

Secondary analyses of the OPTIC trial data demonstrate that amiodarone plus β-blocker is better than sotalol or β-blocker alone and strongly suggest that sotalol is better than β-blocker alone. However, the antiarrhythmic regimens were not without problems. Discontinuation was common (18% for amiodarone plus β-blocker vs 24% for sotalol and 5% for β-blocker alone) and the amiodarone regimen was associated with higher frequency of pulmonary and thyroid adverse events (5% pulmonary adverse events, 4.3% hypothyroidism, and 1.4% hyperthyroidism).

Alternative pharmacological therapy for suppression of ICD shocks may be on the horizon. Azimilide, an investigational class III agent, was evaluated for its effect on ICD therapy. Although it resulted in a reduction of total symptomatic ventricular arrhythmia recurrence (all-cause shocks plus symptomatic VT terminated by antitachycardia pacing), azimilide did not significantly reduce shocks overall.¹⁹ New agents that block atrial-specific ion channels (such as I_Kur, the ultrarapid delayed rectifier potassium current) or that affect novel targets (such as connexins) could allow for safer, more effective prophylaxis of atrial fibrillation and the reduction of inappropriate shocks.²⁰

Based on the study by Connolly et al¹ and taken in context with previous studies, should cardiologists advocate empirical antiarrhythmic therapy for patients receiving an ICD? Importantly, the OPTIC study applies primarily to ICDs placed as secondary prevention, in which sustained ventricular arrhythmias have been observed clinically. There are less data to support the use of antiarrhythmic agents in patients with prophylactic or primary prevention ICD therapy and this group appears to have less frequent need for such therapy; thus, empirical antiarrhythmic therapy cannot be recommended for this setting. For patients who receive an ICD for secondary prevention, one could argue for empirical initiation of amiodarone or sotalol. As per the OPTIC study, such therapy would reduce the absolute risk of shock by 28% or 14%, respectively, and as such would provide a substantial benefit in comfort and possibly quality of life.

However, such pharmacological intervention has important trade-offs. Adverse lung and thyroid effects were common among patients receiving amiodarone over just 1 year, and it can be expected that toxicity would be even more common over longer follow-up. In the OPTIC study, these and other adverse effects resulted in the discontinuation of 18% of those patients receiving amiodarone and 24% of those patients receiving sotalol. As noted by the authors, most patients taking β-blocker alone will not receive an ICD shock and could avoid the inconvenience, cost, and risk of antiarrhythmic therapy. At present, all patients with an ICD who tolerate such therapy should receive a β-blocker. Addition of amiodarone or substitution with sotalol cannot be advocated for all patients and should be considered on an individual basis.