Class I or Class III Agents
for Atrial Fibrillation:

Two medical strategies are used for the treatment of atrial fibrillation (AF): rhythm control and rate control.¹ The former attempts to promote sinus rhythm using atrial defibrillation (cardioversion) combined with drug therapy to prevent recurrence of AF. The latter merely seeks to control ventricular rate without attempting to promote sinus rhythm. Results of two ongoing multicenter trials focusing on the relative merits of these two therapeutic strategies, AFFIRM and RACE, have been preliminarily reported on March 19, 2002 at the Annual American College of Cardiology Meeting. Among 4,060 patients entered into AFFIRM, there were after a follow-up of ~3.5 years, 356 and 306 deaths in the rhythm control and rate control groups, respectively, a nonsignificant trend towards increased mortality with rhythm control. Similarly, in the RACE trial, there appeared to be a trend towards higher morbidity and mortality in rhythm control subgroups. Though these trials failed to demonstrate significant differences for mortality endpoints, they raise once more the question of toxicity from drug therapy for rhythm control. A literature search up to December 2000 reveals no completed randomized trial comparing rhythm to rate control.¹ Further, a large meta-analysis of randomized controlled trials on AF prevention by antiarrhythmic drugs (91 trials up to August 2001) failed to reveal any beneficial effect of antiarrhythmic drug therapy on total mortality.²

In the wake of the CAST trial³ demonstrating the proarrhythmic properties of agents with putative selectivity for sodium channels, the attention switched to drugs inhibiting predominantly potassium channels. However, a trial⁴ with such an agent, d-sotalol, disappointingly showed increased mortality in patients with coronary artery disease (CAD). The electrophysiology of myocardial disease appears to provide a plausible explanation for the failure of potassium channel blockers to improve survival. Both in AF and heart failure, diseased atrial and ventricular myocardium exhibit suppressions of various repolarizing potassium currents (Ito, Ikur, Ikr, others).⁵ The ventricular component of the response is apt to invite QT prolongation and polymorphous arrhythmias (torsades de pointes). Therefore, potassium channel blockers, by prolonging repolarization, do not improve but on the contrary tend to aggravate the acquired long QT syndrome of myocardial disease. Despite the negative experience with d-sotalol, new selective potassium channel blockers were developed and evaluated clinically. In a recent trial (DIAMOND-CHF)⁶ testing dofetilide, a "Apure Class III agent" or HERG/Ikr inhibitor, the authors were apparently satisfied not to have increased mortality. However, considering that arrhythmic death is a dominant cause of mortality from CAD and heart failure,⁷ patients enrolled in the DIAMOND-CHF trial should have exhibited a mortality advantage from effective antiarrhythmic therapy. The dofetilide trialists highlighted beneficial effects on the prevention of AF recurrence. Yet, the accompanying 3.3% incidence of torsade de pointes in the treated group (detected with minimal monitoring) versus 0% in the placebo group was not completely reassuring.

Many have argued that drug toxicity might be minimized using nonselective agents acting on multiple channels. However, two large trials with an agent acting on multiple channels, amiodarone, failed to improve total mortality in patients with CAD and ventricular dysfunction.^8,9 In a recent meta-analysis of 13 mortality trials with amiodarone that included EMIAT⁸ and CAMIAT,⁹ 4 studies showed total mortality reductions of 52%, 55%, 59%, and 61%.¹⁰ Such reductions are surprising in the wake of the large EMIAT and CAMIAT trials showing no significant total mortality reductions. Conversely, among the 13 studies. 3 exhibited increases in total mortality of 19%, 54%, and indeed 94%. The authors computed a 13% mortality reduction for the 13 studies overall, in our view not a meaningful value.¹⁰

It is important to interpret the trials and meta-analyses cited above in a general epidemiological and clinical context. Three large epidemiological studies^11-13 involving a total of 17,437 patients, have revealed that age, CAD, and heart failure are the most potent independent AF risk factors. This strongly suggests that AF is most likely in patient groups at the highest risk for fatal arrhythmias and drug induced proarrhythmia, a fact that requires consideration in the treatment of AF with antiarrhythmic drugs. In contrast to the unconvincing trials and meta-analyses of antiarrhythmic agents, a drug intervention in heart failure not directly targeting voltage dependent ion channels, β-blockade, has repeatedly achieved striking reductions in sudden death (~42% ) and total mortality (~35%).¹⁴ In one of the epidemiological studies, use of β-blockers was one of the few factors associated with an AF risk reduction (-39%).¹² Similarly, ACE inhibitors and statins have decreased coronary disease mortality. We have previously defended the view that interventions acting upstream on the pathways leading to dysrhythmic deterioration might be more rewarding than trying to guess whether an antiarrhythmic drug will, in a given patient, favor anti- or proarrhythmia.¹⁵

Cardiomyocytes from different locations in the adult heart, for instance from different cardiac chambers or different layers of the left ventricular wall, express different phenotypes, reflecting differences in muscle-specific mechanical activities and paracrine influences.⁵ Further, when adult cardiac cells are exposed to pathophysiological conditions, for instance abnormally fast or slow heart rates, they undergo genetic adaptive or maladaptive changes known in cardiac electrophysiology as "electric remodeling." Muscles that lose their physiological contractile activity, undergo very rapid (minutes to hours) genetic adjustments resulting in cellular dedifferentiation demonstrable as suppression of genes expressed in the adult differentiated state and derepression of genes expressed during growth and development ("fetospecific expressions"). Unlike ventricular myocardium, atrial myocardium can undergo marked alterations involving the entire atrial chamber because its functional integrity is not essential for survival. These observations are highly relevant to the pathophysiology of AF. Because changes in ion channel expression are an integral component of genetic adaptation, fibrillating atria may present different molecular targets to antiarrhythmic agents. This explains the well-documented differences in responsiveness and sensitivity to antiarrhythmic drugs between normally contracting and fibrillating atria.¹⁶ Numerous genes, splice variants, and variable heteromultimeric assemblies of subunits of cardiac voltage dependent ion channels allow theoretically a staggering diversity of expressions in response to environmental changes.^5,17 Further, post-lranslational modification such as the phospho/dephospho stales of channel subunits may be important disease-sensilive determinants orion channel function and responsiveness to drugs.^5,17

Recognition of genetic adaptation is essential for the rational developmont of antiarrythmic agents and the understanding of their actions in the clinical setting. Kv1.5 channel expression accounting for the ultrarapid delayed rectifier current (IKur, also called sustained outward current, Iso or Isus) may serve as an example. Kv1.5 is abundantly expressed in normal adult human atrial, but not in normal human adult ventricular muscle, which may suggest that the targeting Kv1.5 might yield atrial cell-selective drugs useful for the treatment of AF.⁴ However, fibrillating human atria exhibit decreased Kv1.5 channel protein expression,⁵ whereas ventricular myocardium undergoing hypertrophy or stress may exhibit fetospecific derepression (upregulation) of Kv1.5. ^18,19 Therefore, we would not be surprised if, by serendipity, efforts to generate potassium channel blockers selective for diseased atria would end up yielding agents selective for diseased ventricles.

Most important is the differentiation between acute and chronic drug effects. Somehow ignoring classical pharmacology that distinguishes between acute and chronic drug efficacy and toxicity, electrophysiologists have for many years attempted to predict chronic drug effects on the basis of single acute laboratory experiments, a maneuver known as "EP-guided" therapy. The clinical pharmacology of amiodarone and procainamide should have warned electrophysiologists not to neglect classical pharmacology.²⁰ According to Kamiya and colleagues,²¹ in cardiac cells that depend on Na-channel for activation, the most consistent acute change of action potential configuration in response to amiodarone is a depression of the maximum upstroke velocity (V max↓, "Class I effect"; no QT prolongation, i.e., no "Class III effect"), presumably by blocking Na-channels in their inactivated state. Why should the putative potassium channel blocking activity of amiodarone (acute in vitro blockade of Ikr and/or Iks with supra-μM IC₅₀'s)^22,23 not become clinically manifest? One might postulate that amiodarone is a prodrug, but no potassium-channel-active metabolite (including desethylamiodarone) with potency exceeding that of the parent compound has been thus far recognized. In a recent study by Kamiya et al. laboratory²² using cells isolated from rabbit ventricular myocardium, it was concluded that amiodarone blocks acutely only the IKr, but not the IKs component of the delayed rectifier. A contradictory conclusion was reached by Balser et al.²³ reporting that in guinea pig ventricular muscle amiodarone blocks acutely the slowly, but not the rapidly activating component of the current. Accepting for a moment Kamiya and colleagues²² conclusion, how meaningful might it be to clinicians treating AF in humans, not in rabbits? The surprising answer is that Ikr as a preferential target for amiodarone's Class III effect might be irrelevant, because this current is not detectably expressed in human atrial muscle according to at least four laboratories.²⁴ This illustrates the importance of distinguishing between different species and between atrial and ventricular cells. If amiodarone is acutely not a potent potassium channel blocker, what mediates its more potent QT-prolonging effects chronically? As suggested by Kamiya,²¹ genetic responses may be important. Shahrara and Drvota²⁵ at the Karolinska Institute recently reported that amiodarone suppressed in mice the expression of thyroid hormone receptors. They further showed in isolated perfused guinea pig hearts that cycloheximide and actinomycin D, inhibitors of protein and RNA syntheses, respectively, completely prevented desethylamiodarone induced prolongation of action potential.²⁶ The prolongation of action potential could be reversed by thyroid hormone (T3). Contrarily, almokalant, an inhibitor with selectivity for HERG/IKr, produced action potential prolongations insensitive to cycloheximide and actinomycin D.²⁶ Results demonstrate incontrovertibly that amiodarone exerts effects on repolarization dependent on gene transcription and not on a direct effect on cell membranes. They illustrate the importance of genetic responses to drug therapy. The fact that changes in cardiac electrophysiology and lipid profile induced by amiodarone mimic those of hypothyroidism have been generally appreciated. In brief, acute amiodarone may act as a mildly hypotensive anesthetic, whereas its chronic effects may depend on thyroid modulation entraining secondary electrophysiological effects. Procainamide further illustrates the importance of differentiating acute from chronic drug effects. Acetylation of procainamide to NAPA (conversion of a Class Ia drug to a dominant Class III metabolite) and T-cell responsiveness to procainamide (autoimmune/Lupuslike response) are both exquisitely dependent on hereditary (genetic) factors.

In clinical practice, Class I, III, and IV agents are often thought of as drugs targeting voltage dependent ion channels selective for Na, K, and Ca ions. Although this simplifying interpretation may have helped to maintain the popularity of the classification by Dukes and Vaughan Williams, it does not conform to their definitions based on hybrid taxonomic criteria (action potential morphology for [Sub-] Class Ia, Ib, Ic, and III, "antisympathetic drugs" for Class II [role of %alpha;-blockers?], and Ca conductance for Class IV).²⁷ One shortcoming of Dukes' and Vaughan Williams' classification is to categorize pleiotropic drugs based on arbitrarily selected acute (in vitro) actions with insufficient regard to long-term clinical actions. To gain insight into the actions of antiarrhythmic agents in current use, study of the genetics and molecular pharmacology of ion channels targeted by these agents is essential.

Structural (crystallography, NMR), genetic (cloning, heterologous expression, site-directed mutagnesis, transgenics), and electrophysiolological studies (patch clamping) have generated abundant new information on the molecular biology of ion channels.^5,17 Voltage dependent cardiac ion channels are assemblies of multiple allosteric sub-units and regulatory or accessory subunits. Specialized domains of the pore-forming alpha sub-units create narrow pathways or "selectivity filters or rings" that can be penetrated only by dehydrated cations possessing narrowly defined size and charge characteristics.^5,17 Organic molecules cannot access the surface of narrow channel pathways and therefore cannot recognize channels directly by the molecular signatures determining ion selectivity. Because antiarrhythmic drugs acting on voltage operated channels are not molecular mimics of physiological ligands (hormones. cytokines) and because they are not peptides addressing specific epitopes, their selectivity for highly conserved voltage operated ion channels is usually limited. Compared to the remarkable ion-channel-active peptide toxins, molecules with highly evolved binding strategies reminiscent of those of antibodies, binding affinities of antiarrhythmic drugs to channel proteins are characteristically several orders of magnitude lower (K_D-values micro- vs nano- to picomolar).¹⁷ As partly illustrated in the so-called "Sicilian Gambit" classification of antiarrhythmics,²⁸ these agents can be demonstrated to exert multitarget effects on ion channel proteins encoded by genes with remarkably preserved basic blueprint.^5,17 Binding of seemingly unrelated antiarrhythmic drugs often involve the same homologous channel domains as inferred from point mutations and may target closely spaced, sometimes even overlapping binding sites. Alpha subunits possess four homologous domains (I-IV), each exhibiting six transmembrane helices (S1-S6). The S6 and S5 helical domains promiscuously contribute to the binding of multiple antiarrhythmic drugs such as dihydropyridine derivatives (nifedipine), local anesthetics (quinidine). potassium channel blockers (dofetilide), 4-aminofYridine, intracellular tetraethyl ammonium, etc.^5,17,29 Binding and mutation studies in intact cells and cell-free systems have shown numerous interactions with different antiarrhythmic drugs. Well-studied examples are the interactive binding of dihydropyridines(nifedipine), phenylalkylamines (verapamil), and benzothiazepines (diltiazem) to L-type calcium channel alpha subunits, and allosteric displacements of dihydropyridines (isradipine) by anesthetics.3o Drug binding can gain in pharmacological specificity by interacting with different conformational states of the channels (rested, open, inactivated states).^5,17,29 For instance, quinidine and flecainide may bind preferentially to open sodium channels, whereas lidocaine may preferentially bind to inactivated sodium channels.¹⁷ Or dofetilide binding to potassium channel alpha subunits (HERG, accounting for rapidly activating inward rectifier, IKr) has been suggested to depend upon sequential conformations, activation allowing access to the inner vestibule (site of channel widening) and inactivation secondarily exposing a submicromolar affinity binding site.²⁹ Here, relatively selective and high affinity binding of the methanesulfonanilide drug appears to depend on an unusual C-type inactivation gating mechanism at the outer pore of HERG/IKr channels. However, even in this instance, the selectivity achieved should not be overestimated.³¹ In brief, looking at antiarrhythmic drugs in current use, several factors may contribute to their limited selectivity for specific ion channels. First, many agents have structural similarities explained by the fact that chemists used common drug prototypes for their development. Second, low molecular drugs cannot access the channel surfaces to recognize them at the sites conferring ion selectivity (antiarrhythmic agents may however exert effects on narrow ion permeation pathways by indirect allosteric interactions as described for effector-modulated enzymes). Third, as nonpeptides, drugs cannot recognize peptide epitopes using high specificity binding strategies as evolved by antibodies and channel-active polypeptide toxins ("neurotoxins"). Fourth, the molecular targets, though products of many genes, are conserved and exhibit similar molecular architecture with predilection target domains (S6 helices).

Testing antiarrhythmic agents on cardiac muscle preparations characteristically exert multiple effects suggestive of interactions with multiple channel targets. These multiple actions or pleiotropic effects, amply documented experimentally, were well recognized and, interestingly enough, even emphasized by Dukes and Vaughan Williams.²⁷ However, for purpose of simplification, certain effects under selected experimental conditions were highlighted to classify (and later reclassify or subclassify) antiarrhythmic agents. Whether dominant effects would be the same during chronic therapy targeting diseased human (atrial) myocardium, is however quite uncertain for the many reasons discussed above. Examples of this uncertainty are provided by fiecainide and propafenone, drugs categorized as Class Ic. Looking at published KD and IC50 values for these two agents for various potassium channels (transient outward, Ilo; ultra-rapid delayed rectifier. Ikur; rapid and slow activating delayed rectifier, Ikr and Iks, others) in different preparations, one notes that they fall in about the same high nanomolar to low micromolar range as those described for Class III agents.^32-39 Thus, on the basis of preclinical pharmacology and considering typical clinical dosages (e.g.. 250 μlg dofetilide vs 300 mg propafenone), one has to conclude that flecainide and particularly propafenone can partly mimic Class III agents. Assuming that pure Class III agents are particularly effective for the treatment of AF as implied by developers of these drugs (DIAMOND-CHF dofetilide trial),⁶ why should propafenone and flecainide not predominantly act by the same mechanism when achieving similar clinical effects on human AF?² Contrarily, as discussed above, amiodarone, a Class III agent, acts acutely as a Class I agent, an observation highly relevant to pharmacological termination of AF with intravenous drugs. Also, effects of amiodarone during acute oral loading, until proved otherwise, might importantly depend on an anesthetic effect. According to Dukes and Vaughan Williams, verapamil is a Class IV agent acting by reducing calcium ion conductance.²⁷ However. calcium channel blockers with the highest selectivity for L-type cardiac Ca-channels, the dihydropyridine derivatives, are widely held to have limited antiarrhyrthmic utility. Contrarily, from the outset after its introduction in Europe, antiarrhythmic actions of verapamil were emphasized.⁴ Briefly after its clinical introduction, voltage clamp experiments demonstrated verapamil to be a potent blocker of delayed rectifier currents (over the same concentrations producing Ca-current suppression),⁴¹ a property confirmed by modern experiments using patch clamping and heterologous expression systems. Verapamil in low micromolar concentrations blocks various K-channels (Kv1.1, Kv1.5 Iks [KvLQT1 + mink], IKr [HERG]) expressed in Xenopus oocytes.⁴² To our knowledge, dihydropyridines and diltiazem fail to exhibit equipotent inhibitory effects on Ca and K channels (for instance, nifedipine's K_D-values [equilibrium dissociation constant] for ventricular myocardial L-type Ca- versus Ikur/Kv1.5 K-channels may be about 0.2 μM versus > 20 JlM (close to aqueous solubility of nifedipine).⁴³

In summary, not only amiodarone as advertised by manufacturers, but most other antiarrhythmic agents exert effects on multiple channels in various test systems. The pharmacotherapeutic significance and relative importance of these effects are more often than not poorly defined in the clinical setting, especially when electrophysiological mechanisms of arrhythmias under consideration remain debatable.

Many trials have been performed to evaluate the efficacy of drugs to maintain patients in sinus rhythm. In a recent meta-analysis of randomized trials, Nichols et al.² collected through a Medline search, 91 English language studies (studies summarized in heart website www.heartjnl.com). Agents evaluated included quinidine, disopyramide, procainamide, propafenone, flecainide, sotalol, amiodarone, dofetilide, and ibutilide. Relying on the Vaughan Williams classification, the authors concluded that Class IA, IB, IC, and III drugs were all superior to placebo in promoting sinus rhythm. No one Class appeared superior in maintaining sinus rhythm for a mean follow-up of 46 ± 136 (SD) days (range 0.04-1,096 days). Overall, the median proportion of patients in sinus rhythm at follow-up was 55% (range 0%-100% !) for patients on active treatment, and 32% (range 0%-90%) for those on placebo. Compared with placebo, there was no evidence that any Class was associated with either increased or decreased mortality (median survival 99%, range 55%-100%). Compared with any other comparative drug, amiodarone was not associated with significant differences in sinus rhythm maintenance or survival. The authors admitted that the analysis lacked statistical power to detect differences between different agents or small differences between Classes of drugs.² The lack of statistical power while compiling no less than 91 studies, illustrates the hopeless variability between studies, reminiscent of the amiodarone meta-analysis discussed above.¹⁰ In a recent meta- analysis of 52 studies on the treatment of post-operative AF with β-blockers, dl-sotalol. or amiodarone, relative superior efficacy of any of these agents was again not demonstrable.⁴⁴

Here, we will briefly focus on one large influential study, the CTAF trial (Canadian Trial of AF).⁴⁵ A total of 403 patients with at least one episode of AF within the past 6 months were randomized to open label treatment with either amiodarone (6 weeks loading/high dose, then maintenance with 200 mg/day) or another agent, either propafenone, 300 mg twice daily, or dl-sotalol, 160 mg twice daily. At entry, the two groups were well matched, with current AF exhibited by 35% and 38% of the amiodarone and propafenone/sotalol patients. After a follow-up of 468 ± 150 days, first AF recurrence was interpreted to have occurred in 35% of amiodarone versus 63% of propafenone/sotalol patients (P < 0.001). Adverse events requiring discontinuation occurred in 18% of amiodarone versus 11% of propafenone/sotalol patients (P = 0.06). The authors concluded that amiodarone was more effective than propafenone or dl-sotalol.⁴⁵ This study is unconvincing because Kaplan-Meier curves were divergent only during the initial period when patients received high dosed amiodarone. During subsequent therapy with maintenance dosages in all patients, the curves were not divergent or even convergent (Fig. 1). Lacking a placebo control curve (essential in the light of the past proarrhythmia experience), the survival statistic is uninterpretable. A loading schedule is described for amiodarone, but not for propafenone or sotalol, raising doubts about dosage bias. However, "loading doses of the drugs" were crucially important, since the investigators elected to enter patients not returned to sinus rhythm within the first 21 days after randomization into the survival statistic as "recurrence on day 1." Propafenone, not only amiodarone, may typically benefit from loading,^46,47 and sotalol should be heart rate and QT interval titrated, paying attention to renal function.⁴⁸ Optimized acute dosing is of pivotal importance to keep the atrial fibrillatory state as brief as possible and prevent so-called atrial remodeling that perpetuates muscular incompetence (prevent "AF from begetting AF)". Optimized dosing of propafenone/sotalol might have improved their acute beneficial effects to match those of amiodarone. In their discussion, the authors state that "other antiarrhythmic agents are less effective and are associated with higher risk" (statement not supported by Nichol et al.² and Crystal et al.⁴⁴). However, the trial did not adequatelyevaluate the chronic toxicity of amiodarone, and its discontinuation rate during the trial was high.

Commonly used Class I and III drugs for the rhythm control of AF have yielded similar variable results.¹ Eclectically tabulating a few studies among the many published to suggest superiority of selected agents without making clear how and why the studies were selected does not conform to principles of evidence-based medicine.

References

1. Mead GE, Flapan AD, Elder AT. Electrical cardioversion for atrial fibrillation and flutter. Cochrane Database Syst Rev 2002; 1;CD002903.

2. Nichol G, McAllister F, Pham B, et al. Meta-analysis of randomised controlled trials of the effectiveness of antiarrhythmic agents at promoting sinus rhythm in patients with atrial fibrillation. Heart 2002; 87:535-543.

3. Echt DS, Liebson PR, Mitchell B, et al. For the CAST Investigators. Mortality and morbidity in patients receiving encainide, flecainide, or placebo. New Engl J Med 1991; 324:781-788.

4. Waldo AL, Camm AJ, Ruyter H, et al. for the SWORD Investigators. Effect of d-sotalol on mortality in patients with left ventricular dysfunction after recent and remote myocardial infarction. Lancet 1996; 348:7-12.

5. Snyders DJ. Structure and function of cardiac potassium channels. Cardiovasc Res 1999; 42:377-390.

6. Torp-Pedersen C, Moller M, Bloch-Thomson PE, Agner E, Carlsen J, Videboek J, et al. for the Danish Investigations of Arrhythmia and Mortality on Dofetilide Study Group (DIAMOND-CHF). Dofetilide in patients with congestive heart failure and left ventricular dysfunction. New Engl J Med 1999; 341:857-865.

7. Zheng ZJ, Croft JB, Giles WH, et al. Sudden cardiac death in the United States. 1989-1996. Circulation 2001; 104:2158-2163.

8. Julian DG, Camm AJ, Frangin Gjanse MJ, et al. for the European Myocardial Infarct Amiodarone Trial Investigators. Randomized trial of effect of amiodarone on mortality in patients with left-ventricular dysfunction after recent myocardial infarction: EMIAT. Lancet 1997; 349:667-674.

9. Camm JA, Conollaly SJ, Roberts R, et al. for the Canadian Amiodarone Myocardial Infarction Arrhythmia Trial Investigators. Randomised trial outcome after myocardial infarction in patients with frequent or repetitive ventricular premature depolarisations: CAMIAT. Lancet 1997; 349:675-682.

10. Amiodarone Trials Meta-Analysis Investigators, Effect of prophylactic amiodarone on mortality after acute myocardial infarction and in congestive heart failure: Meta-analysis of individual data from 6500 patients in randomized trials. Lancet 1997; 350:1417-1424.

11. Benjamin EJ, Levy D, Vaziri SM, et al. Independent risk factors for atrial fibrillation in a population-based cohort - The Framingham Study. JAMA 1994; 271:840-844.

12. Psaty BM, Manolio TA, Kuller LH, et al. Incidence of and risk factors for atrial fibrillation in older adults. Circulation 1997; 96:2425-2461.

13. Wilhelmsen L, Rosengren A, Lappas G. Hospitalizations for atrial fibrillation in the general male population: Morbidity and risk factors. J Intern Med 2001; 250:382-389.

14. Foody JM, Farrell MH, Krumholz HM. Beta-blocker therapy in heart failure: Scientific review. JAMA 2002; 287:883-689.

15. Henry PD, Pacifico A. Altering molecular mechanisms to prevent sudden arrhythmic death. Lancet 1998; 351:1276-1278.

16. Nattel S. Effects of ionic remodeling on cardiac antiarrhythmic drug actions. J Cardiovasc Pharmacol 2001; 38:809-811.

17. Catterall WA. Molecular mechanisms of gating and drug block of sodium channels. Novartis Found Symp 2002; 241:206-218.

18. Guo W, Kada K, Kamiya K, et al. IGF-I regulates K(+)-channel expressIon of cultured neonatal rat ventricular myocytes. Am J Physiol 1997; 277:H2599-25606.
19. Levitan ES, Hershman KM, Sherman TG, et al. Dexamethasone and stress upregulate Kv1.5 K+ channel gene expression in rat ventricular myocytes. Neuropharmacology 1996; 35:1001-1006.

20. Nasir N Jr, Swarna US, Boahene KA, et al. Therapy of sustained ventricular arrhythmias with amiodarone. Prediction of efficacy with serial electrophysiologic studies. Cardiovasc Pharmacol Therapeut 1996; 1:123-132.

21. Kodoma I, Kamlya K, Toyama J. Amiodarone: Ionic and cellular mechanisms of action of the most promising class III agent. Am J Cardiol 1999; 84:20R-28R.

22. Kamiya K, Nishiyama A, Yasui K, et al. Short- and long-term effects of amiodarone on the two components of cardiac delayed rectifier K(+) current. Circulation 2001; 103:1317-1324.

23. Balser JR, Bennet PB, Hondeghem LM, et al. Suppression of time-dependent outward current in guinea pig ventricular myocytes - actions of quinidine and amiodarone. Circ Res 1991; 69:519-529.

24. Lee KS, Lee EW. Ionic mechanism of ibutilide in human atrium: Evidence for a drug-induced Na+ current through a nifedipine inhibited inward channel. J Pharmacol Exp Ther 1998; 286:9-22.

25. Shahrara S, Drvota v. Thyroid hormone alpha1 and beta1 receptor mRNA are downregulated by amiodarone in mouse myocardium. J Cardiovasc Pharmacol 1999; 34:261-267.

26. Drvota V, Blange I, Haggblad J, et al. Desethylamiodarone prolongation of cardiac repolarization is dependent on gene expression - A novel antiarrhythmic mechanism. J Cardiovasc Pharmacol 1996; 32:654-661.

27. Dukes ID, Vaughan Williams EM. The multiple modes of action of propafenone. Eur Heart J 1984; 5:115-125.

28. Task Force of the Working Group on Arrhythmias of the European Society of Cardiology. The Sicilian Gambit: A new approach to the classification of antiarrhythmic drugs based on their actions on arrhythmogenic mechanisms. Circulation 1991; 84:1831-1851.

29. Lees-Miller JP, Duan J, Teng JQ, et al. Molecular determinant of high-affinity dofetilide binding to HERG1 expressed in Xenopus oocytes: Involvement of S6 sites. Molec Pharmacol 2000; 57:367-374.

30. Motoike HK, Bodi I, Schwartz A, et al. A region in the IVS5 of the human cardiac L-type channel Is required for the use-dependent block by phenylalkylamines and benzothiazepines. J Biol Chem 1999; 274:9406-9420.

31. Finlayson K, Pennington AJ, Kelly JS. [3H]dofetilide binding in SHSY5Y and HEK293 cells expressing a HERG-like K+ channel? Eur J Pharmacol 2001: 412:203-212.

32. Slawsky MT, Castle NA. K+ channel blocking actions of flecainide compared with those of propafenone and quinidine in adult rat ventricular myocytes. J Pharmacol Exp Ther 1994; 269:66-74.

33. Yamashita T, Nsksjima T. Hamada E. et al. Flecainide inhibits the transient outward current in atrial myocytes. J Pharmacol Exp Ther 1995; 274:315-321.

34. Wang Z, Fermini B, Nattels S. Effects of flecainide, quinidine, and 4-aminopyridine on transient outward and ultrarapid delayed rectifier current in human atrial myocytes. J Pharmacol Exp Ther 1995; 272:184-196.

35. Follmer CH, Cullinan CA, TJ. Differential block of cardiac delayed rectifier current by class 1c antiarrhythmic drugs: Evidence for channel block and Unblock. Cardiovasc Res 1992; 26:1121-1130.

36. Seki A, Hagiwara N, Kasanuki H. Effects of propafenone on K currents in human atrial myocytes. Br J Pharmacol 1999; 126:1153-1162.

37. Duan D, Fermini B, Nattel S. Potassium channel blocking properties of propafenone in rabbit atrial myocytes. J Pharmacol Exp Ther 1993; 264:1113-1123.

38. Mergenthaler J, Haverkamp W, Huttenhofer A, et al. Blocking effects of the antiarrhythmic drug propafenone on the HERG potassium channel. Naunyn Schiedebergs Arch Pharmacol 2001; 363:472-480.

39. Gross GJ, Castle NA. Propafenone inhibition of human atrial myocyte repolarizing currents. J Mol Cell Cardiol 1998; 30:783-793.

40. Mjasnikov LA. Symposium on Verapamil. Moscow, Nov 24, 1969. Arzneimittelforschung (Drug Res) 1970; 20 (9a):1277-1356.

41. Henry PD. Comparative pharmacology of calcium antagonists: Nifedipine, verapamil, and diltiazem. Am J Cardiol 1980; 46:1047-1058.

42. Waldegger S, Miemeyer G, Morike K, et al. Effect of verapamil enantiomers and metabolites on cardiac K+ channels expressed in Xenopus oocytes. Cell Physiol Biochem 1999; 9:81-89.

43. Zhang X, Anderson JW, Fedida D. Characterization 0f nifedipine block of the human heart delayed rectifier, hKv1.5. J Pharmacol Exp Ther 1997: 281:1247-1256.

44. Crystal E, ConOlly SJ, Sleik K, et al. Interventions on prevention of postoperative atrial fibrillation in patients undergoing heart surgery. Circulation 2002; 106:75-80.

45. Roy D, Talajic M, Dorian P, et al. Amiodarone to prevent recurrence of atrial fibrillation. Canadian Trial of Atrial Fibrillation (CTAF) Investigators. New Engl J Med 2000; 342:913-920.

46. Khan IA. Single oral loading dose of propafenone for pharmacological cardioversion of recent-onset atrial fibrillation (meta-analysis). J Am Coll Cardiol 2001; 37:542-547.

47. Boriani G, Martignani C, Biffi M, et al. Oral loading with propafenone for conversion of recent-onset atrial fibrillation: A review on in-hospital treatment. Drugs 2002; 62:415-423.

48. Pacifico A, Hohnloser SH, Williams JH, et al. for the d,I Sotalol Implantable Cardioverter-Defibrillator Study Group. Prevention of implantable-defibrillator shocks by treatment with sotalol. New Engl J Med 1999; 340:1855-1862.