Methods

Top Methods Results Discussion Author & Article Info References

Our review is based on reports from the world literature that were available as of June 1993. An English-language literature search was done by weekly reviews of Current Contents: Clinical Practice, by a MEDLINE search covering the past 20 years, and by supplementary manual searches of references in review articles. Literature entries were selected by providing key words such as amiodarone, torsade de pointes, proarrhythmia, and aggravation of arrhythmia. We did not consider articles that did not provide an English summary. We excluded review articles and articles with incomplete data with regard to the nature of the arrhythmias treated or the amiodarone-associated proarrhythmic effects. Special effort was made to eliminate obvious duplications in reported studies. For this purpose, published abstracts were included in the analysis only when we found no full-length report by the same investigators. However, some overlap was inevitable because some centers published multiple, discrete reports. As suggested previously [9, 10], two investigators independently evaluated the methods and results sections of all trials. We specifically screened the articles for features such as patient demographics, entry criteria, methods, major study end points, and descriptions of long-term outcome and side effects. Reports with unclear descriptions of the methods used or of the efficacy or side effects of amiodarone were excluded from the analysis. Three categories of articles were evaluated separately: case reports, retrospective studies evaluating amiodarone in at least 50 consecutive patients with an average follow-up of at least 6 months, and prospective controlled trials with amiodarone. In the identified reports, electrocardiographically documented episodes of torsade de pointes, incessant ventricular tachycardia, and ventricular fibrillation (if not previously present) had been considered by the individual investigators to be amiodarone-induced proarrhythmic effects. We included all such events in our analysis. Increases in premature ventricular contractions of any magnitude were not counted as proarrhythmic reactions and thus were not included.

Results

Top Methods Results Discussion Author & Article Info References

Case Reports

Although amiodarone has been used for almost two decades in Europe and since the early 1980s in the United States, we found only 65 case reports of amiodarone-induced torsade de pointes. Since Jorens and colleagues' published report of 59 such cases [11], 6 additional patients with this amiodarone-related proarrhythmic effect have been described [12-17]. Although no definite conclusions with respect to risk factors for the development of torsade de pointes can be drawn from such case reports, torsade de pointes did occur more frequently in women (62% of cases). Known predisposing factors for the occurrence of torsade de pointes, such as hypokalemia, were described in half the cases. Of particular interest is that 20% of the patients were receiving other antiarrhythmic drugs that prolonged the QT interval in addition to amiodarone. Thus, it is possible that torsade de pointes may have been provoked in at least some of these patients by combining class IA antiarrhythmic agents with amiodarone.

Uncontrolled Studies

Using the criteria described in the methods section, we selected 17 reports from the world literature that were published between 1982 and 1993 [18-34]. These studies involved 2878 patients treated with amiodarone for various types of arrhythmias. The most common presenting arrhythmia was sustained ventricular tachycardia or prehospital cardiac arrest caused by ventricular fibrillation in 2021 patients (70%) (Table 1). Most patients had coronary artery disease and a history of myocardial infarction. Dilated or hypertrophic cardiomyopathy was the next most frequently encountered diagnosis. In most of the patients, left ventricular function was depressed; these patients must therefore be considered as being at high risk for future arrhythmic events.

Cases of amiodarone-induced tachyarrhythmias considered to constitute proarrhythmic reactions were reported in 57 patients, or 2.0% of the total study population. The incidence of proarrhythmia ranged from 0% [20-22, 24, 26] to 7% [29] in the individual studies. The nature of the proarrhythmic events in 17 patients was not described (30%) (Table 1). In the remaining 40 patients, 15 (37.5%) developed torsade de pointes. In 8 patients (20% of patients with detailed information concerning the type of proarrhythmia), we considered the occurrence of ventricular fibrillation or the development of incessant ventricular tachycardia in patients with no history of this arrhythmia to be proarrhythmic events. Incessant ventricular tachycardia (n = 11) or sustained ventricular tachycardia (that is, when absent before amiodarone was given) other than torsade de pointes (n = 4) accounted for 37.5% of all described proarrhythmic effects.

Controlled Studies

As of June 1993, the results from seven prospective, placebo-controlled studies had been reported either in full-length articles [35-40] or in abstract form [41] (Table 2). Most patients enrolled in these studies were patients with a history of coronary artery disease and remote myocardial infarction. In these trials, the incidence of amiodarone-induced proarrhythmia was 0%. In addition, in an ongoing placebo-controlled Veterans Affairs trial involving patients with congestive heart failure caused by coronary artery disease or dilated cardiomyopathy, no proarrhythmic events were observed in 674 patients (half of whom received amiodarone) completing the enrollment phase (Singh BN. Personal communication).

Amiodarone in Patients with Previous Drug-induced Torsade de Pointes

Three published reports describe the use of amiodarone in patients with previously documented torsade de pointes induced by antiarrhythmic agents [42-44]. In these studies, data have been reported for 32 patients who were considered to need continued antiarrhythmic drug therapy after recovering from torsade de pointes induced by other antiarrhythmic drugs. Most of these patients (66%) presented with a history of sustained ventricular tachycardia or ventricular fibrillation (Table 3). Torsade de pointes has been observed during initial treatment with class I agents (17 patients with quinidine [53%], 10 with procainamide (31%), and 2 with disopyramide [6%]) or class III compounds (3 patients with sotalol [10%]). During subsequent long-term amiodarone treatment that lasted from 1 to 46 months, one sudden cardiac death occurred. Neither recurrent episodes of torsade de pointes nor a different ventricular tachyarrhythmia were reported in any of the remaining patients. In these patients, the amiodarone-induced average prolongation of the QT interval was of the same magnitude as that observed when torsade de pointes developed while patients received other drugs in all three studies.

Discussion

Top Methods Results Discussion Author & Article Info References

Because of the CAST trials [6, 7, 45, 46] and, more recently, the Electrophysiologic Study versus Electrocardiographic Monitoring [47] and the Cardiac Arrest in Seattle: Conventional versus Amiodarone Drug Evaluation trials (CASCADE) [48], there is an increasing focus on the development of antiarrhythmic compounds [8] either as "pure" class III agents [49] or as those associated with antisympathetic actions producing substantial heart rate slowing (for example, sotalol). The use of these agents is associated with a particular incidence of proarrhythmic reactions, especially torsade de pointes, the precise incidence varying relative to predisposing factors in individual patients. The major exception is amiodarone, which produced an apparently low incidence of torsade de pointes despite considerable lengthening of the QT-QTc interval and a substantial degree of bradycardia [50, 51]. Numerous studies have also documented the beneficial antiarrhythmic effects of the drug [35-41]. Amiodarone is currently being investigated in several controlled clinical trials assessing its effect on mortality in different subsets of patients [52-55]. Clearly, it is clinically important to determine the proarrhythmic potential of the drug compared with its antiarrhythmic potential to place in perspective its overall side-effect profile in the context of its long-term prophylactic use.

Incidence of Proarrhythmic Events in Conventional Drugs and Amiodarone

In the landmark article by Velebit and colleagues [56], aggravation of preexisting or provocation of new arrhythmias such as torsade de pointes was observed in 6% to 16% of treatment attempts with various antiarrhythmic drugs. This study focused mainly on a statistical increase in the number of simple and complex premature ventricular contractions, but it also included new tachyarrhythmias. Subsequent reports confirmed the occurrence of these proarrhythmic reactions, especially when class I antiarrhythmic drugs were used [57-59]. However, after the CAST studies [6, 7], increases in the number of simple and complex premature ventricular contractions no longer appear to be valid for gauging proarrhythmic reactions. The most relevant proarrhythmic reactions of class I antiarrhythmic drugs clinically take the form of incessant ventricular tachycardia or fibrillation; undoubtedly, they affect mortality and compound the difficulties of showing a favorable effect on survival in clinical trials. In the case of class I agents, many predisposing factors for developing proarrhythmic effects have been identified. In a case–controlled retrospective study, Slater and coworkers [60] showed that a history of sustained ventricular tachyarrhythmias and a reduced left ventricular function (left ventricular ejection fraction <35%) reliably identified patients at risk for proarrhythmic events during therapy with quinidine, mexiletine, or encainide. Other findings, such as electrocardiographic measures, drug doses, or drug levels, were not predictive.

In contrast, the proarrhythmic reaction of class III compounds (including class IA agents) usually takes the form of torsade de pointes, which often produces syncope or presyncope, but the arrhythmia may deteriorate into ventricular fibrillation. Torsade de pointes may be responsible for sudden death in certain patients [61]. Amiodarone is a powerful class I agent and a potent class III compound [1, 2]. It might be expected to exhibit proarrhythmic reactions stemming from both types of electrophysiologic properties. We found in the literature a surprisingly low incidence of proarrhythmia for the drug with respect to either torsade de pointes or ventricular tachycardia or ventricular fibrillation, even though in the reviewed studies amiodarone was used mostly in patients with compromised ventricular function who had resistant life-threatening ventricular arrhythmias. In the cited prospective controlled trials, most patients had a history of myocardial infarction [35, 37, 39, 40], depressed left ventricular function [36, 38, 41], the presence of nonsustained ventricular tachycardias [37-40], or all three. Thus, the low incidence of amiodarone-associated proarrhythmia in these patients might reflect the overall risk profile of these persons compared with those with a history of sustained ventricular tachycardia or prehospital cardiac arrest.

Proarrhythmic Events Observed with Amiodarone in Patients with Ventricular Tachycardia or Fibrillation

Our review confirms the generally held clinical notion that torsade de pointes is uncommon in patients with ventricular tachycardia or fibrillation who receive long-term amiodarone therapy. The overall incidence may be less than 1%, which compares favorably with the 3% to 5% incidence with sotalol [62] and the 5% to 8% incidence with quinidine [63]. Amiodarone could be safely given to patients with previous antiarrhythmic drug-mediated torsade de pointes [42-44]. The latter finding is of particular interest because the amiodarone-induced prolongation of the QT interval noted in the electrocardiogram was similar to that previously shown to be associated with torsade de pointes occurring with other drugs that prolonged the QT interval. Therefore, the occurrence of torsade de pointes during therapy with class IA agents such as quinidine or procainamide, or even class III compounds such as sotalol, does not appear to predict the development of torsade de pointes during amiodarone administration.

That amiodarone may also produce incessant ventricular tachycardia as a proarrhythmic reaction is not as widely appreciated. We identified 11 such cases Table 1 in which the temporal sequence of amiodarone administration during the early stages of therapy was consistent with cause and effect. Again, the overall incidence of incessant ventricular tachycardia appears to be low. However, it has been reported in such cases that the arrhythmia may abate and disappear with continued drug administration [64]. The exact mechanism of this phenomenon is unknown. However, such a proarrhythmic reaction might be caused by the early and predominant class I action [2, 65], which might subsequently be attenuated by the stepwise increases in the intensity of the antiadrenergic and class III actions of the drug as a function of time.

In torsade de pointes, the temporal sequence of the development of the arrhythmia and drug administration is not readily correlated, at least from published reports. Unlike incessant ventricular tachycardia, torsade de pointes appears to occur later in association with progressive increase in the QT interval, especially if electrolyte disturbances supervene. It should be stressed that the most reliable criterion for a proarrhythmic reaction with amiodarone is the occurrence of torsade de pointes. Ultimately, the relative frequency of serious life-threatening ventricular arrhythmias induced by an antiarrhythmic drug may need to be gauged in relation to its overall effect on mortality. It is noteworthy that amiodarone might lower arrhythmia mortality consistent with the overall low fatal proarrhythmic potential of the drug.

Factors Responsible for the Low Incidence of Proarrhythmic Events Observed with Amiodarone

The precise mechanism for the low incidence of proarrhythmia with amiodarone remains unclear but may be multifactorial. Amiodarone is an extremely complex compound with a unique electropharmacologic profile and a marked propensity for prolonging cardiac repolarization. The increase in action potential duration that it produces is caused by blocking the delayed rectifier current I_k [66, 67], although the effects of long-term drug therapy on K currents are unknown. Amiodarone can also block the fast sodium current [65] and the slow inward calcium current mediated through the L-calcium channels [68]. In addition, amiodarone has been shown to exert noncompetitive ß-adrenergic antagonism [69]. In particular, the calcium-channel blocking effect may be an important factor with respect to the low incidence of amiodarone-associated torsade de pointes. Most evidence suggests that provocation or perpetuation, or both, of this particular arrhythmia is caused by calcium-dependent early afterdepolarizations [70-72]. The reduced availability of calcium influx in the presence of amiodarone may explain the low likelihood of development of torsade de pointes despite a marked prolongation of the repolarization phase of the action potential [73]. It is noteworthy that amiodarone has been shown to reduce or abolish early afterdepolarizations induced either by barium [73] or electrically [74] in isolated cardiac muscle. Whether such an effect—consistent with a potentially beneficial effect on torsade de pointes—is caused by an inhibition of calcium current is uncertain because the properties of the drug bepridil, a potent calcium channel blocker that prolongs the QT interval, are not consistent with such a notion [50, 51].

The electrophysiologic effects of amiodarone differ in other respects from those of other antiarrhythmic drugs [75]. In contrast to other class III agents, amiodarone produces a much shorter prolongation of repolarization in Purkinje fibers than in ventricular muscle. The converse differential effect occurs for nearly all class III agents, namely, a disproportionate increase in the action potential duration in the Purkinje fibers [51] in which early afterdepolarizations responsible for torsade de pointes appear to be generated [74]. Similarly, whereas with most class III compounds the duration of action potential decreases as the frequency of stimulation increases, with amiodarone there is an increase over the entire range of stimulation frequencies [76]. Finally, increasing evidence suggests that amiodarone might act to selectively inhibit the action of triiodothyronine at the site of the myocardial nuclear receptor [75]. This action results in an electrophysiologic condition that mimics hypothyroidism, a state with an apparently low preponderance for ventricular arrhythmias that is possibly caused by increased homogeneity of repolarization in contiguous myocardial cells [49]. The potential clinical relevance of this phenomenon has already been suggested by Cui and coworkers [77], who found a significant reduction in QT dispersion of the surface electrocardiogram during long-term administration of amiodarone, whereas quinidine tended to increase dispersion [78]. Thus, the electrophysiologic properties of amiodarone that affect the issue of the observed low propensity of the drug to induce proarrhythmic reactions are complex and multifactorial but appear to reduce electrical instability of the myocardium, thereby preventing torsade de pointes and ventricular tachycardia or fibrillation.

Clinical Implications

Recent clinical trials have uniformly shown that amiodarone can reduce mortality within the first year after a myocardial infarction [37, 39, 40, 79]. Larger blinded trials [53-55] such as those currently in progress (for example, the European Myocardial Infarct Amiodarone Trial and the Canadian Amiodarone Myocardial Infarction Arrhythmia Trial) are nevertheless needed to confirm these available results on the beneficial effects of amiodarone on mortality. Effect on mortality is also suggested by uncontrolled data from various reports in which many patients who did not respond to conventional therapy received long-term amiodarone treatment [31-34]. A salutary effect on mortality has also been indicated by the CASCADE trial, in which empiric therapy with amiodarone was found to be superior to conventional (class I) agents administered on the basis of therapy guided by Holter monitoring or electrophysiologic testing [48]. Our review suggests that in all subsets of patients studied, amiodarone has few proarrhythmic effects. Whatever the mechanism that is finally established for such a low degree of proarrhythmia, the clinical significance of the drug's low proclivity to aggravate or induce life-threatening arrhythmias is clear. It may account for the emerging data from controlled trials in which the drug has reduced total and cardiovascular mortality [45]. The potential to prolong survival by controlling life-threatening arrhythmias is greater in a compound in which the antifibrillatory effects exceed the proarrhythmic reaction. In this respect, the net effects on mortality of amiodarone must differ from those of agents such as flecainide [6] or moricizine [7], whose profibrillatory effects appear to be greater than their antifibrillatory actions, at least in certain subsets of patients. The ongoing placebo-controlled trials of amiodarone in patients with heart failure or in patients who survived myocardial infarction, as well as trials comparing implantable devices in patients with ventricular tachycardia or fibrillation, are likely to be crucial in establishing the effect of amiodarone on arrhythmia mortality.