the sicilian gambit - circulationcirc.ahajournals.org/content/circulationaha/84/4/1831.full.pdfthis...

1831

Point of View

The Sicilian GambitA New Approach to the Classification of Antiarrhythmic

Drugs Based on Their Actions on

Arrhythmogenic Mechanisms

Task Force of the Working Group on Arrhythmias of

the European Society of Cardiology

The Queen's Gambit is an opening move in chess that provides a variety of aggressive optionsto the player electing it. This report represents a similar gambit (the Sicilian Gambit) on thepart of a group of basic and clinical investigators who met in Taormina, Sicily to consider theclassification of antiarrhythmic drugs. Paramount to their considerations were 1) dissatisfac-tion with the options offered by existing classification systems for inspiring and directingresearch, development, and therapy, 2) the disarray in the field of antiarrhythmic drugdevelopment and testing in this post-Cardiac Arrhythmia Suppression Trial (CAST) era, and3) the desire to provide an operational framework for consideration of antiarrhythmic drugsthat will both encourage advancement and have the plasticity to grow as a result of the adv-ancesthat occur. The multifaceted approach suggested is, like the title of the article, a gambit. It isan opening rather than a compendium and is intended to challenge thought and investigationrather than to resolve issues. The article incorporates first, a discussion of the shortcomings ofthe present system for drug classification; second, a review of the molecular targets on whichdrugs act (including channels and receptors); third, a consideration of the mechanismsresponsible for arrhythmias, including the identification of "vulnerable parameters" that mightbe most accessible to dr-ug effect; and finally, clinical considerations with respect to antiar-rhythmic drugs. Information relating to the various levels of information is correlated acrosscategories (i.e., clinical arrhythmias, cellular mechanisms, and molecular targets), and a

"spread sheet" approach to antiarrhythmic action is presented that considers each drug as aunit, with similarities to and dissimilarities from other drugs being highlighted. A completereference list for this work would require as many pages as the text itself. For this reason,referencing is selective and incomplete. It is designed, in fact, to provide sufficient backgroundinformation to give the interested reader a starting frame of reference rather than to recognizethe complete body of literature that is the basis for this article. (Circulation 1991;84:1831-1851)

F or two decades, the approach to antiarrhythmicdrug development and administration has fo-cused on the Vaughan Williams classifica-

tion.1-3 This classification, presented in Table 1, hasbeen useful in teaching students and physicians be-cause it is physiologically based, can be learned quickly,and facilitates discussion of the potentially beneficial ordeleterious actions of drugs. The Vaughan Williamsclassification originated at a time when our knowledgeof electrophysiological mechanisms (including the roles

This article appears in similar form in the October issue of theEuropean Heart Journal.

Addresses for reprints: Michael R. Rosen, MD, Professor ofPharmacology and Pediatrics, College of Physicians and Surgeonsof Columbia University, Department of Pharmacology, 630 West168 Street, New York, NY 10032, or Peter J. Schwartz, MD,Professor of Medicine, Istituto di Clinica Medica, Universita diMilano, Via Francesco Sforza, 35, 20122 Milano, Italy.

of receptors and channels) was less extensive than nowand relatively few antiarrhythmic agents were available.In the ensuing years there has been a continuingattempt to fit new concepts about arrhythmias andabout the actions of new agents into its general frame-work. However, the usefulness of this classification forbasic investigators and clinicians is limited because itprovides incomplete links among the actions of antiar-rhythmic agents, the mechanisms of arrhythmias, andthe efficacy of therapy.

Other limitations that have become increasinglyvisible are as follows:

1) The classification is a hybrid, such that classes Iand IV represent block of ion channels; class II,block of receptors; and class III, a change in anelectrophysiological variable (the action potentialduration). In addition, a single class effect can be

by guest on June 26, 2018http://circ.ahajournals.org/

Dow

nloaded from

http://circ.ahajournals.org/

1832 Circulation Vol 84, No 4 October 1991

TABLE 1. Vaughan Williams Classification of Antiarrhythmic Drugs*

Class IDrugs with directmembrane action (Nachannel blockade)

la: Depress phase 0Slow conductionProlong repolarization

Ib: Little effect on phase 0 innormal tissueDepress phase 0 inabnormal fibers

Shorten repolarizationIc: Markedly depress phase 0

Markedly slowconduction

Slight effect onrepolarization

Class IISympatholytic drugs

Class IIIDrugs that prolong

repolarization

Class IVCalcium channel-

blocking drugs

*This is the classification as modified by Harrison and others.1-3

produced by multiple mechanisms. For example,prolongation of action potential duration (i.e., a classIII effect) can result from block of any one of anumber of K' channels or from modification of Na+or Ca2` channel function. Moreover, the multiplesubdivisions of class I create a problem in placementof drugs based on their effects on conduction and onrepolarization. In fact, some drugs have severalcla'sses of actions. As an example, amiodarone, theprototype of class III, has class I, class II, and class IVeffects, and no one or combination of amiodarone'sknown effects explains why it is so effective clinically.

2) The classification essentially describes antitachy-arrhythmic drugs acting via block of channels andcurrents. The possibility that activation of certainchannels or receptors might be antiarrhythmic is notconsidered.

3) The classification is incomplete. For example,a-adrenergic blockers, cholinergic agonists, digitalis,and adenosine are not included. In addition, thepotential roles of agents that modulate gap junctionalconductance, biochemical pumps, or exchangers arenot considered.

4) The classification (excluding class II) is primar-ily based on the effects of drugs on electrophysiolog-ical characteristics of isolated, normal cardiac tissues.However, in diseased tissues, where arrhythmias arelikely to arise, channels and receptors are modified,and the actions of drugs may be modified as well.

5) The classification does not incorporate the con-cept that antiarrhythmic drugs can be effective invarious ways: slowing tachycardias and making thembetter tolerated, terminating established arrhyth-mias, or preventing arrhythmia initiation.

6) In simplifying matters, the classification sug-gests we know more than we do. This can haveserious consequences when physicians or regulatoryagencies assume, without adequate information, thata drug in a given class will have the'favorable and/orunfavorable effects characteristic of other drugs as-signed to that same class.

These limitations and the increasing complexity ofinformation concerning the pathophysiology of ar-rhythmias and the clinical efficacy of drugs make itclear that a more sophisticated framework for under-standing antiarrhythmic drug actions and their rela-tion to clinical drug efficacy is needed. The goal ofthis workshop was to reconsider the actions of anti-arrhythmic drugs and to develop a rational and usefulconstruct whereby basic and'clinical investigators andclinicians might communicate and design researchmore effectively in order to further drug develop-ment, evaluation, and administration.

ApproachWe recognized that a classification of antiarrhyth-

mic actions of pharmacological agents might be con-structed within various frameworks. For example,drug actions can be expressed at the molecular level(i.e., direct effects on-channels and/or receptors).Although our understanding of drug actions at thislevel is far from complete, it does provide oneframework to which new knowledge may be addedand from which mechanistic and therapeutic insightsand ideas concerning drug development may begained. Advances in molecular biology and biophys-ics have identified numerous sites of drug action,both real and potential. For this reason, any work-able classification based on specific direct effectswould require astute powers of discrimination toselect those effects that are physiologically importantand those that are artifacts of the experimentalsystem. While an effective classification of drugsshould include actions on targets at the molecularlevel, this inclusion is more useful for basic researchrather than clinical directions.An alternative and very pragmatic framework

might classify drugs in terms of their effects on' aspecific type of arrhythmia in humans. This wouldhave the advantage of allowing a one-step process ofdrug selection by the clinician. Once a diagnosis ismade, a drug might be selected from the group ofpharmacological agents whose actions are associated


Dow

nloaded from


Working Group on Arrhythmias The Sicilian Gambit 1833

NERVES AND HUMORS )----- TARGETS (- DRUG ACTIONS

A. CHANNELSB. PUMPSC. RECEPTORSD. OTHER

MECHANISMS OF ARRHYTHMIASA. CELLULARB. EXPERIMENTAL ANIMALC. CLINICAL

with the arrhythmia diagnosed. However, there areseveral disadvantages to such a classification. First, ifa grouping were created according to site of origin ofan arrhythmia, this would result in a large number ofdiagnostic categories. Second, drug effects vary mark-edly even for morphologically similar arrhythmias.Susceptibility of a given type of arrhythmia to a drugwill be affected by associated factors including (butnot limited to) electrolyte balance, ischemia, me-chanical factors (stretch), and neural influences. Fur-ther, very few, if any, studies are available that wouldallow an interindividual-comparison of drugs in thesame setting, and different mechanisms might lead tothe same "phenotype" of arrhythmia on the surfaceelectrocardiogram. In summary, no absolute andrigorous classification of drug effects on the basis ofclinical efficacy appears possible at present.Another possibility is that no simple classification is

adequate but that a framework could be constructedthat incorporates the broad spectrum of knowledgeabout the actions of drugs, the mechanism(s) of ar-rhythmias, and the "vulnerable parameters" or targetsin the arrhythmia mechanism a drug might modify.Such information might ultimately be linked to themechanisms underlying clinical arrhythmias since, in-creasingly, clinical diagnostic categories are expressedin terms of arrhythmia mechanisms (e.g., atrioventric-ular [AV] nodal reentrant tachycardia).

Such a framework, based on mechanisms for ar-rhythmias and stressing the interrelations of drugsand their targets, has 'several advantages. It groupsdrugs having various actions according to that actionthat is relevant to countering the arrhythmogenicmechanism. For example, the group of drugs that are

clinically effective for AV nodal reentrant tachycar-dia is representative of three different classes (Ic, II,

and IV) in the Vaughan Williams classification butincorporates, as well, several drugs outside that clas-sification (adenosine, cardiac glycosides, and musca-rinic agonists). In the proposed classification, thesediverse agents might be grouped according to a

common action to depress conduction dependent on

either Na+ current or Ca2+ current. In addition, sucha system could be expanded as understanding ofmechanisms or number of drugs is increased.

A classification based on mechanisms for arrhyth-mias also would encourage the search for clinicalinformation that might clarify the causes of arrhyth-mias in patients. Even now, suppositions about mech-anism are often implicit in drug selection. For exam-ple, the selection of an agent to depress conductionor prolong refractoriness in treating ventricular tachy-cardia is based on the assumption (often unper-ceived) that the mechanism is reentry. A classifica-tion based on mechanisms would require thatassumptions underlying current practices be reeval-uated through suitable and adequate tests. It wo'uldthus require a focus on variables most likely to besuitable for modification by a therapeutic agent. Sucha classification also' would encourage drug develop-ment tightly linked to the evolving understanding ofarrhythmogenic mechanisms. It would stimulate'theeducation of students, trainees, clinicians, and scien-tists about the relations between drug actions and themechanisms of arrhythmias. Finally, it would inter-face readily with concepts and knowledge in clinicalmedicine and in the basic sciences; in fact, the key tothe success of such a classification would depend onthe extent and the success of its interface with a

number of basic and clinical elements. These ele-ments are summarized in Figure 1 and are describedin detail in the following pages. Here, we will firstprovide a description of the targets of drug action(i.e., the sites at which drugs act, including channels,pumps, and receptors) and of the effects of drugs atthe cellular and molecular levels. We then willconsider the effects of drugs in relation to the mech-anisms responsible for cardiac arrhythmias in exper-imental studies. Finally, the application of informa-tion about antiarrhythmic drugs to clinical therapywill be reviewed. Incorporated with the above infor-mation is a multidimensional classification that we

believe describes current knowledge cohesively andprovides a convenient framework for interpretingfuture developments.

Molecular and Cellular Targets of Drug ActionAntiarrhythmic drugs interact with cellular struc-

tures that influence electrophysiological function.These structures include ion channels, pumps/carriers,

FIGURE 1. Elements for a classifi-cation system. See text for discussion.


Dow

nloaded from



receptors, and perhaps some cytoplasmic regulators ofsecond messenger systems. In theory, the actions ofdrugs can be understood from a combination of theidentification of their targets, knowledge of the modi-fication they cause in the function of the target system,and awareness of the role the target system plays indetermining electrophysiological behavior.Our ability to identify drug targets and the nature

of drug-target interactions has increased dramati-cally with the availability of a variety of moleculartechniques. We now know the primary molecularstructures of many ion channels, pumps, and recep-tors, and this provides the opportunity to determinedrug binding sites. This understanding has greatlyincreased our insight into the complex interplay ofionic currents that underlies cardiac electrical behav-ior. The following section identifies the moleculartargets for drug action on electrophysiological func-tion and provides a functional definition of thecharacteristics of drug action on these targets.

Molecular Targets for Drug ActionChannels. Channels are large glycoproteins that span

the membrane bilayer and, under an appropriate stim-ulus, form pores that permit ions to cross the mem-brane rapidly, thereby creating ion currents. A channelis often selective for certain ions (e.g., sodium channelsor potassium channels), although different channelscan incorporate regions in which the amino acid se-quences are quite similar (homologous). Channel pro-teins that have similar sequences and physiologicalproperties are termed "isoforms." Some channels onlyopen after a delay following the stimulus, and somechannels rectify; that is, they conduct ions more effec-tively in one direction across the membrane than theother. Stimuli for channel opening include a change inmembrane voltage; chemical signals, which may actdirectly or by occupying an adjacent receptor; or me-chanical deformation. It appears likely that these stim-uli act by inducing conformational changes in thechannel proteins.The process whereby an ion channel protein re-

sponds to an external stimulus to change conforma-tion is termed "gating." Channel activation is similarto (but may not be synonymous with) channel opening.The rate of change of channel conformation (gatingkinetics) can be very rapid, with major changes inchannel function occurring in less than 1 msec, orquite slow, with changes occurring over seconds. Gat-ing kinetics should not be confused with the kinetics ofantiarrhythmic drug binding to or dissociation fromchannels, which occur at different rates for differentdrugs. Once open, channel proteins may stay openuntil closed by another signal, or they may inactivate,that is, close even in the face of a maintained stimulus.Inactivated channels generally do not reopen whenrestimulated until they have recovered from inactiva-tion, a time- (and often voltage-) dependent process.

Channels can be classified according to the mech-anisms that govern their opening: voltage-gated orligand-gated. A number of channel types have been

purified, cloned, and sequenced from heart and fromother excitable cells.4 They share considerable ho-mology, especially in the intramembranous segments,such that it is not surprising that some channel-activedrugs interact with more than one channel type. Inaddition, channels of one type (e.g., the sodiumchannels) are even more homologous among tissues,so that drugs that interact with one channel type inheart may also interact with that channel type in thenervous system. The progress in determining channelstructure encourages us to believe that it will soon bepossible to determine key structural differences anddrug binding sites and to tailor drugs that are tissue-and channel-specific. A schematic representation ofthe role of channels and pumps/carriers in generatingthe normal transmembrane resting and action poten-tials is illustrated in Figure 2. Knowledge of thestructure and function of channels and pumps/carri-ers is increasing rapidly, and the following tabulationwill thus require a regular update.CHANNELS THAT CARRY INWARD CURRENTS: SO-

DIUM CHANNELS. These channels represent a familyof large proteins with multiple subunits.4 The main(a-) subunit contains about 2,000 amino acids andincludes four major repeated internal sequences thatrepresent internal subunits. Presently, five membersof this family have been cloned and expressed, in-cluding one from heart muscle. In brain and skeletalmuscle, the a-subunit is normally accompanied byother subunits, but it is functional even when alone.The a-subunit in heart is also functional alone,5although some evidence suggests that there may bean additional subunit. Functional evidence points toat least two isoforms of the cardiac sodium channel.

INa is the inward excitatory current carried by Na+through a voltage-activated sodium channel. Thecurrent is activated at threshold to produce rapiddepolarization and to provide the current to driveaction potential or impulse propagation in atrial,His-Purkinje, and ventricular cells. These Na+ chan-nels are sparse or absent in sinoatrial (SA) and AVnodal cells.6

INa-B is a proposed background Na+ currentthrough a voltage-independent channel in SA nodalcells.7 It is offset by an outward K+ current at thebeginning of phase 4, but as the K' current decays, itcontributes to pacemaker behavior.CHANNELS THAT CARRY INWARD CURRENTS: CA2+

CHANNELS. These channels also represent a family ofproteins with multiple subunits. Some of the subunitsfrom various tissues, including a cardiac subunit of theL type, have been cloned and sequenced, and theyshow extensive homology. There are at least two typesof calcium channels in the heart.8 The L type (IC,,-L),which is blocked by verapamil, diltiazem, and thedihydropyridines, predominates in all cardiac tissuesstudied. The T type (ICa-T) has been identified inpacemaker tissues. Ca2' channels are also found inhigh density in vascular smooth muscle, endocrinecells, and the brain.


Dow

nloaded from



Atrial & Ventricular Cells

INalCaa-]Ca-Ty'[INJ _

INalca

Sino-Atrial Node Cells

0 or small

-qw

I Kl

I K

Ito

pump

IKfATP1I

Na-B

0

K

1 _.

I K(ACh)

-pump

FIGURE 2. Currents and channels involved in generating the resting and action potential. The time course of a stylized actionpotential ofatrial and ventricular cells is shown on the left and ofsinoatrial node cells on the tight. Above and below are the variouschannels and pumps that contribute the currents underlying the electrical events. See text for identification of the symbols anddescription of the channels or currents. Where possible, the approximate time courses of the currents associated with the channelsorpumps are shown symbolically without effort to represent their magnitudes relative to each other. The heavy bars for Ic, Ip,,mp,and IK(ATP) only indicate the presence of these channels orpump, without implying magnitude of currents, since that would vary withphysiological andpathophysiological conditions. The channels identified by brackets [INS and IK(ATP)] imply that they are active onlyunder pathological conditions. For the sinoatrial node cells, I'Na and IKI are small or absent. Question marks indicate thatexperimental evidence is not yet available to determine the presence of these channels in sinoatrial cell membranes. Although it islikely that other ionic current mechanisms exist, they are not shown here because their roles in electrogenesis are not sufficiently welldefined.

ICa-L is the calcium current that is activated regen-eratively from a relatively depolarized threshold po-tential to produce depolarization and propagation inSA and AV nodal cells. It is also present in atrial,His-Purkinje, and ventricular cells, where it contrib-utes to the plateau and triggers calcium release fromthe sarcoplasmic reticulum. It is inactivated by bothdepolarization and [Ca2"]j, but it usually lasts longenough to contribute to the overall plateau currents.It is the target for the clinically useful calciumchannel blockers, and it is strongly modulated byneurotransmitters.8

1Ca-T is a calcium current through a different volt-age-gated channel that is activated at potentialsintermediate between thresholds for INa and ICa-L. Itprobably contributes inward current to the laterstages of phase 4 depolarization in SA node andHis-Purkinje cells, and it may also play a role inabnormal automaticity in atria.7 It is almost absentfrom ventricular cells.OTHER INWARD CURRENT CHANNELS. There are two

cationic nonselective channels that under normalphysiological conditions allow current to be carriedby Na+. They are not well characterized, and no

structural information is available yet.

If is an inward current carried by Na+ through a

relatively nonspecific cationic channel that is activatedby polarization to high membrane potentials in SAand AV nodal cells and His-Purkinje cells.9 Thiscurrent generates phase 4 depolarization and contrib-utes to pacemaker function. Its kinetics are fairly slow,and it is strongly modulated by neurotransmitters.

INS is a channel that is gated by [Ca2+]1.10 It is acationic nonselective channel that if activated at theresting potential, produces an inward current carriedby Na+. Under some conditions, it is activated byCa2' release from the sarcoplasmic reticulum during[Ca2"]i overload and contributes to delayed afterde-polarizations (see below).CHANNELS THAT CARRY OUTWARD CURRENTS.

Many functionally different types of potassium chan-nels have been identified in heart and other tissues.Indeed, almost all cells have some type of potassiumchannel. These channels are composed of four sub-units, each about one quarter the size of the sodiumchannel subunit.4 It seems likely that naturally occur-ring potassium channels are composed of differentsubunit types. It has not yet been possible to matchthe cloned channels with the functional currents, so

this discussion will focus on K' currents definedfunctionally.

Ca-L

I Ca-T

I Na/Ca

K(ACh)


Dow

nloaded from



IKI is the K' current responsible for maintainingthe resting potential near the K' equilibrium poten-tial in atrial, AV nodal, His-Purkinje, and ventricularcells." This current, which is also called the "inwardrectifier," shuts off during depolarization (inwardrectification). Its absence from SA node cells isimportant in letting small currents control the pace-maker rate.

IK is a K' current carried through voltage-gatedchannels with slow activation kinetics, giving it thename "delayed rectifier." IK turns on slowly duringthe action potential plateau and is the major currentcausing repolarization. After repolarization it turnsoff slowly enough in the SA node to contribute tophase 4 depolarization. There are probably severalpharmacologically different 'K channels, a subset ofwhich can be modulated by neurotransmitters.12"13

Ito is a K' current that turns on rapidly afterdepolarization and then inactivates. One type of I, isactivated by [Ca2]i, and the other is voltage-activatedand modulated by neurotransmitters.14 Ito can play animportant role in modifying action potential durationand in contributing to the heterogeneity of repolar-ization because of its nonuniform distribution (it ispresent in subepicardial but not subendocardial mus-cle). Similar currents, called IA, are found in nerveand skeletal muscle cells.

IK(ACh) is a K' current whose channel is activated bythe muscarinic (M2) receptor via GTP regulatory (G)protein signal transduction.15 It shuts down some-what during depolarization (inward rectification), butit contributes outward current both at rest and duringthe action potential. It is particularly important in theSA and AV nodes and in atrial cells, where it canproduce substantial hyperpolarization, and in atrium,where it produces marked acceleration of repolariza-tion. This channel also is opened by activation of thepurinergic (adenosine) receptor and so the currentalso is designated 'K(Ado).

IK(ATP) is a K+ current carried through a metaboli-cally regulated channel.'6 This channel is blocked byATP and is strongly activated during hypoxia. Thechannel has been present in all cardiac cells studied,but it has not been identified in the SA node. It maycontribute to shortening of the action potential duringischemia. Experimentally available antiarrhythmicdrugs can either increase or decrease this K+ current,thereby accelerating or prolonging repolarization.

IC is a Cl- current, usually quite small, but which canbe greatly increased by adrenergic receptor activation,favoring repolarization.'7 Because the Cl- concentra-tion is sometimes above what is expected for equilib-rium, the Cl- channel could under those circumstancesgenerate inward current. Because the channel rectifies,the inward current could be quite small, but it mightcontribute to pacemaker depolarization.

IK(Ca) is a K+ current carried through a channel thatis activated by [Ca2+]i. It appears to require very highlevels of [Ca'+]i for activation. Its presence in cardiaccells has been hard to establish, and its physiologicalrole is not yet clear.'1

SOME OTHER IMPORTANT CHANNELS. Cardiac cellsare coupled electrically and chemically to one an-other by large channels called gap junctions or con-nexons.18 These junctions are composed of two mul-timeric complexes, one set for each cell. The two setsalign themselves to form a large pore between thecells that permits passage of ions and small mole-cules. The conductance of gap junctions is regulatedby [Ca'+]i and pHi and perhaps by phosphorylationvia activation of the ,3-adrenergic receptor system.They play a major role in isolating cells damaged byischemic injury or trauma from adjacent more normalcells.A calcium channel in the sarcoplasmic reticulum

can be triggered to release Ca's by calcium entrythrough ICa L. Because it also can be modulated by thedrug ryanodine, it is often called the ryanodinereceptor.'9

Pumps/cariers. ACTIVE TRANSPORT. At least twoATP-dependent pumps in the sarcolemma generateionic current, the Na-K pump (which is blocked bydigitalis) and the calcium pump. Several isoforms ofthe Na-K pump have been identified by differingaffinities for digitalis. A different ATP-dependentcalcium pump is found in the sarcoplasmic reticulum.Both sarcolemmal pumps have been cloned andsequenced, and several isoforms of the Na-K pumpare identified. Some evidence suggests that they aremodulated by phosphorylation through the adrener-gic receptor system. INa K pump is the current generatedby the Na-K pump. Because each cycle transportsthree Na+ out and two K' into the cell, it generatesa small outward current that is relatively constantduring the cardiac cycle.20

CARRIERS. This group of membrane proteins facil-itates exchange of ions or substrates, or pumps themusing energy. It includes the Na/Ca countertransportsystem in the sarcolemma and in the mitochondria,the Na/H exchanger, the Na/K/Cl cotransporter, andthe Cl/HCO3 exchanger. The cardiac Na/Ca ex-changer has recently been cloned, but no selectiveblocking or activating drugs are presently known.Amiloride blocks the Na/K/Cl cotransporter and theNa/H exchanger, as well as several types of channels.

INa/Ca is the current generated by the Na/Ca coun-tertransport system, which exchanges 1 Ca21 for 3Na+.21 It is the chief means of Ca2' efflux through thesarcolemma. The direction of the current depends onthe relation between the Na+ and Ca2' gradients andthe membrane potential. At the resting potential theexchanger generates a small inward current, butupon depolarization the current may show a briefoutward phase and then become inward as [Ca2']irises. During [Ca2]i overload, the sarcoplasmic retic-ulum may release Ca2 spontaneously during dias-tole, causing the Na/Ca exchanger to generate alarger inward current, and thereby contributing tothe generation of delayed afterdepolarizations (seebelow).

Receptors. The autonomic nervous system is animportant modulator of cardiac rhythms. a- and


Dow

nloaded from



13-adrenergic, muscarinic and purinergic receptorsinfluence various channels and pumps, which, inturn, influence the initiation and propagation ofnormal and abnormal cardiac impulses. The receptorsystems in the heart are coupled by G proteins totheir effector systems.22 These G proteins translatethe results of an external stimulus, such as receptoroccupancy, into a physiological response, such asactivation of a kinase and subsequent phosphoryla-tion of target proteins (e.g., ion channels). Alterna-tively, they may link receptors to second messengersystems and pumps. The two best understood recep-tor-effector coupling systems are the 81-adrenergicand the M2 muscarinic.f-ADRENERGIC RECEPTOR-EFFECIOR COUPLING SYS-

TEM. Both 81- and P32-adrenergic receptors exist in thehearts of various species, including humans,23 but thecontribution of the 32 component is not yet clear. Bothreceptor subtypes are coupled to adenylate cyclase viathe G protein GQ. 13-Adrenergic receptor stimulationmodulates L-type calcium channels, the If channel,various potassium channels, Ic, and, under certainconditions, sodium channels.The binding of agonist to 8-adrenergic receptor in

cardiac myocytes increases intracellular cyclic AMPlevels, activating the cyclic AMP-dependent proteinkinase and phosphorylating a peptide associated withthe L-type Ca21 channel, increasing Ca2 current andcontractility.24 The chronotropic effect of ,3-adrener-gic stimulation is also mediated by an increase incyclic AMP that appears to shift the activation curveof the pacemaker current If toward more positivepotentials.25 -Adrenergic stimulation also enhancesNa-K pump function, which would tend to hyperpo-larize cardiac fibers, especially those that are par-tially depolarized as in the setting of acute ischemia.A fast, direct G protein (GQ) coupling between

,3-adrenoceptors and calcium, If, and sodium chan-nels has recently been documented.26 'K and a com-ponent of 'to are enhanced by 13-adrenergic stimula-tion,27 whereas IK1 does not appear to be affected.1013-Adrenergic stimulation enhances voltage-depen-dent K' and voltage-independent Cl- currents.28There appear to be physiological and pathophys-

iological importance of each of these actions, asfollows: The effect on the L-type Ca2 channel could,in its own right, induce early afterdepolarizations andtriggered activity, as well as increase free intracellu-lar Ca2 concentrations, which in turn would poten-tiate delayed afterdepolarizations and resultant trig-gered activity. The effect on If can increase the rate ofimpulse initiation of the sinus node pacemaker, aswell as of latent atrial and ventricular pacemakers.The potassium channel effects tend to acceleraterepolarization and shorten refractoriness, and boththe Ca24 and K' channel effects may combine toaccelerate AV nodal conduction and shorten AVnodal refractoriness. These ion channel effects canaccount for the tachycardias and QT interval short-ening associated with sympathetic stimulation of theintact heart.

Clinically, ,B-adrenergic blocking agents have well-established effects on those arrhythmias that requirethe participation of the AV node. The 18-adrenergicblocking effects result largely from drug-induceddecreases in conduction velocity and prolongation ofrefractoriness in the AV node, actions reflectingblock of catecholamine effects on Ca'4 and K' chan-nels. Limited subsets of patients with various ventric-ular arrhythmias, for example, some ventricular pre-mature depolarizations, exercise-induced ventriculartachycardias, some ventricular tachycardias inducibleby programmed stimulation, and adenosine sensitiveventricular tachycardias, also may respond to ,B-ad-renergic blockade. ,B-Adrenergic blocking drugs alsoreduce the risk of sudden cardiac death in selectedpopulation subgroups (e.g., post-myocardial infarc-tion, long QT syndrome). In post-myocardial infarc-tion patients, the mechanisms of protection are notyet fully understood.

a-ADRENERGIC RECEPTOR-EFFECTOR COUPLING SYS-TEM. There are two and possibly three a,-adrenergicreceptor subtypes in the heart,29 linked via G proteinsto a series of effector processes (Na-K pump,30 Kchannels,31 and phospholipase C29) that modulate im-pulse initiation and repolarization. Information on sec-ond messenger involvement in these processes is in-complete. The Na-K pump-stimulating effect isresponsible for a decrease in the rate of impulseinitiation by automatic fibers outside the sinus node,and the decreases induced in IKI and/or I, result inprolongation of repolarization.30,32

Studies in isolated tissues suggest that a,-adrener-gic receptor subtype stimulation also can be arrhyth-mogenic. a,-Agonists induce triggered rhythms viadelayed or early afterdepolarizations and abnormalautomatic rhythms33,34 studied in settings where[Ca24]0 is elevated, repolarization is prolonged, isch-emia and reperfusion are simulated, or infarction isinduced. In intact cats after coronary occlusion andreperfusion, there is an increase in a,-adrenergicreceptor number and affinity.35 This is associatedwith ventricular tachycardia and fibrillation that startoccasionally during ischemia and invariably within1-3 minutes of reperfusion.For humans, only limited data are available con-

cerning a-adrenergic involvement in cardiac arrhyth-mias. The efficacy of phentolamine against arrhyth-mias in a small group of patients in the immediatepostmyocardial infarction period has been reported.In patients with the congenital long QT syndrome, asubset whose arrhythmias are not blocked by pro-pranolol do respond to left thoracic sympathectomy,leading to the speculation that a-adrenergic mecha-nisms may be involved.36 In summary, the clinicalantiarrhythmic potential of a-adrenergic blockaderemains largely untested.MUSCARINIC RECEPTOR-EFFECTOR COUPLING SYS-

TEM. The M2 receptor has been identified pharmaco-logically as the dominant cardiac muscarinic recep-tor.37 Its density is two to five times higher in the atriathan the ventricles.38 The M2 receptor is coupled


Dow

nloaded from



directly to the ligand-operated potassium channel[LK(ACh)] by the G protein GK.39 The M2 receptor alsoinhibits adenylate cyclase via Gi and in this wayaffects currents that are modulated by the cyclicAMP dependent protein kinase, including 1Ca-L, If,and presumably 1K-40A major aspect of muscarinic action on the heart is

the antagonism of adrenergic effects at the level ofthe adenylate cyclase-cyclic AMP system.40 Whereasf-adrenergic agonists stimulate the cAMP secondmessenger system via the G protein GQ, muscarinicagonists such as acetylcholine inhibit this same sys-tem via Gi.39 The effects of muscarinic activation onCa2' and K' currents contribute to the depression ofconduction in the AV node. The effects of M2agonists to decrease If and to increase conductance ofthe muscarinic K' channel suppress the sinus nodepacemaker.4 M2-receptor activation leads to an in-crease in IK(ACh), which hyperpolarizes and shortensthe action potential in atrial tissues.41

Increased vagal activity (whether due to directstimulation or to pharmacological activation) reducesthe incidence of ventricular fibrillation during acutemyocardial ischemia in intact animals.42 This effect isonly in part dependent on heart rate reduction; italso depends on the antiadrenergic action of musca-rinic activation. Muscarinic blockade with atropinesometimes increases the incidence of ventricularfibrillation during ischemia.

Clinically, vagal activation is effective against ar-rhythmias involving the AV node. There is no directevidence for an antiarrhythmic effect of vagal activa-tion at the ventricular level in human subjects. Thereis indirect evidence that impairment of or a decreasein either vagal tone (heart rate variability)43 or re-flexes (baroreflex sensitivity)44 is associated with in-creased mortality and incidence of sudden deathamong post-myocardial infarction patients. Hence,there is direct evidence that an antiarrhythmic actionof muscarinic agonists occurs at the supraventricularlevel and indirect evidence suggesting that an impor-tant role may be played in ventricular arrhythmiasand sudden death.PURINERGIC RECEPTOR-EFFECTOR COUPLING SYS-

TEM. The cardiac purinergic receptor population isdesignated A1. It is coupled presumably by GK to theligand-operated potassium channel [IK(Ado)].45 Assuch, its actions are thought to reflect operation ofthe same effector coupling pathway as that describedfor the M2 muscarinic receptor. Clinically, a subset ofbenign ventricular tachycardias is terminated by thea1-agonist, adenosine. Of greater importance, reen-trant tachycardias involving the atrioventricular nodeare consistently terminated by adenosine.46 Hence,this receptor-effector system offers an attractive tar-get for therapeutic interventions.

Cytoplasmic regulators ofsecond messengers. We arebecoming increasingly aware of the roles a variety ofmolecules may play in the regulation of secondmessengers. The latter, in turn, exert importantinfluences on ionic currents. For example, agents that

block cyclic AMP-phosphodiesterase may increasecyclic AMP and affect target channels such as theL-type calcium channel. Receptors for angiotensinand a1-adrenoreceptors activate phospholipase C,which in turn activates protein kinase C via diacyl-glycerol and releases inositol trisphosphate. The di-acylglycerol-protein kinase C pathway results inphosphorylation of transsarcolemmal Ca2' channels,and the inositol trisphosphate pathway induces sar-coplasmic reticulum release of Ca2`. Both eventswould increase free intracellular Ca2`. Other exam-ples of regulators are: phospholipase A2 may stimu-late the ligand-activated K(ACh) channel via arachi-donic acid metabolic pathways; long-chain fattyacids18-20 are reported to activate delayed rectifier Kcurrents; and finally, dephosphorylation is promotedby phosphatases and a specific muscle phosphataseinhibitor, okadaic acid, is known to increase cardiaccalcium currents.

Functional Definition ofDrug ActionAlthough characterization of a drug's action on its

target channel, pump, or receptor is an effective wayto determine its molecular site of action, studies ofdrug effects on function are crucial to an understand-ing of the complex interactions between drugs andthe heart. These studies at the cellular, tissue, orintact organism level can be performed using avariety of experimental and clinical indexes, as re-viewed in this section and Table 2.

Pharnacodynamics. A variety of experimental varia-bles can be used to study pharmacodynamics, as listedin Table 2. Those variables modified by drug actioninclude the ionic currents that flow through channels,the activities (i.e., free concentrations) of ions intracel-lularly and extracellularly, the transmembrane actionpotential characteristics of single cells or intact tissuesmeasured using microelectrodes; surface electrogrammeasurements of electrical activity in tissues; and theproperties of excitability, conduction, refractoriness,and impulse initiation.Drug-channel interactions. A major determinant of

the effects of antiarrhythmic drugs is seen in theiractions at their channel binding sites. It must beunderstood that the drug binding sites in channelsare thought to reside at specific loci, the access towhich is determined by molecules or "gates" thatmay be opened or closed depending on time- and/orvoltage-regulated processes. The channels them-selves may be construed of as residing in restingstates, where ionic current is not being passed; inopen states, where ions can pass; and in inactivatedstates, where ions are no longer being carried and atransition back to the resting state is occurring.Whereas most channel-active agents act to blockionic currents, a few [e.g., activators of IK(ATP)I canmodify electrophysiological behavior by increasingcurrent.47-50Most antiarrhythmic drugs behave as though they

do not have continuous access to their binding sites.As a result, their intermittent blocking action de-


Dow

nloaded from



TABLE 2. Pharmacological Indices Associated With the FunctionalDefinition of Drug Action

Experimental variables used to study pharmacodynamicsIonic currentsIonic activityTransmembrane action potentialExtracellular electrogramExcitabilityConductionRefractorinessImpulse initiation

PharmacodynamicsDrug-channel interactions

Tonic blockPhasic blockUse dependenceVoltage dependence

Recovery from phasic blockVoltage dependence

Competition and interactionIonsDrugs

Receptor-mediated modulation

Nonelectrophysiological propertiesCardiac contractilityVascular toneCardiac disease

PharmacokineticsBioavailabilityAbsorptionDistributionClearanceMetabolites

pends not only on the drug concentrations but alsoon the variables that influence association with anddissociation from their binding sites on the pro-tein.47-52 Local anesthetic antiarrhythmic drugs thathave continuous access to binding sites on the chan-nel will demonstrate tonic block, that is, block thatdevelops in the absence of excitation.47,49 This can beassessed by measuring the degree of reduction inionic current that is present with infrequent stimula-tion. In general, the degree of tonic block correlateswith the degree of lipid solubility.49'52When heart rate or rate of stimulation of isolated

tissues is increased, most local anesthetic antiarrhyth-mic drugs induce an additional reduction of sodiumcurrent, that is, phasic block.47-52 Phasic block isgenerally attributed to intermittent access to or vari-able affinity for channel binding sites that depends on

the gating state of the channel.53'54 Most antiarrhyth-mic drugs have relatively low affinity for sodiumchannels in the resting state, and the affinity increases

as the channel activates.49-55 Furthermore, drug affin-ity for the different activated or inactivated states mayvary, so that association of the drug with the channelcan occur preferentially during channel activation tothe open state (i.e., just before or during the upstrokeof the action potential) or during the prolongedplateau period when the channel is inactivated.49-55 Inthis way, use dependence of phasic block may dependon the duration of the action potential as well as thenumber of activations occurring.The magnitude of use dependence is also voltage

dependent.56 This reflects the fraction of sodiumchannels that activate in response to a depolarizingpulse and become accessible to the blocking action ofthe drug as well as the direct effects of voltage on theionized form of the drug. Voltage dependence as-sumes particular importance in pathophysiologicalsettings such as ischemia, which reduce the mem-brane potential of cardiac cells. Drugs such aslidocaine show voltage dependence, having a fargreater effect on conduction in depolarized than innormally polarized tissues.

Because'phasic block represents the balance be-tween drug association with and dissociation from itsbinding site, measurement of recovery from blockalso has been helpful in elucidating drug-channelinteractions, as recovery is believed primarily toreflect dissociation.47,4952 Recovery from block caneither be directly or inversely dependent on voltageor independent of voltage.Many antiarrhythmic drugs prolong repolarization,

an effect that is usually mediated by their actions onpotassium channels.57'58 It is of interest that mostantiarrhythmic drugs that alter repolarization do soin such a way that prolongation of repolarization isgreatest at the slowest heart or stimulation rates andprogressively decreases as the rate of stimulationincreases. This phenomenon is described as "reverseuse dependence." This has been attributed to areduction in potassium channel blockade during theplateau or at positive potentials such that blockdecreases as use increases.Because refractoriness is determined by the avail-

ability of sodium channels to respond to excitation aswell as by other determinants of excitability, changesin repolarization as well as residual sodium channelblock present during phases 3 and 4 of the transmem-brane action potential will determine a drug's effectson refractoriness. Given the complicated interplaybetween stimulus rate and ion channel blockade,drug effects on refractoriness will be rate dependent.The magnitude of this rate dependence will vary fordifferent drugs.49,'5758Drug competition and interaction. Some drugs inter-

fere with each other by directly competing for thesame receptor or by producing offsetting electrophys-iological effects.59'60 Others can enhance each other'sefficacy either by summating physiological effects,thereby allowing the administration of lower doses ofboth drugs and reducing the toxicity of both, or byachieving the kinetic mix of on-rate and off-rate that


Dow

nloaded from



best achieves "therapeutic" use dependence. In gen-eral, antiarrhythmic drug combinations have notbeen studied systematically, although such a studyrepresents an opportunity to improve patient manage-ment. Even less is known about interactions betweenthe parent drugs and their metabolites, although ex-amples of antagonism have been demonstrated.

Modulation. Little is known at present about theeffects of receptor-mediated modulation of channelbehavior or drug interaction, but any alteration inmembrane potential or temporal distribution ofchannel state (as described under "Receptors")would be expected to affect drug action.

Effects of nonelectrophysiological properties. Drugsthat block the calcium or sodium channel can alsohave substantial negative inotropic effects, curtailingtheir usefulness in clinical conditions associated withdepressed cardiac contractile function. Adrenergicreceptor modulation may also affect contraction.Similarly, changes in vascular tone consequent todirect drug action on the vasculature have potentiallyimportant effects on heart rate and metabolic state.Finally, disease states themselves may alter channelproperties as well as the actions of receptors orpumps, further modifying drug action.

Pharnacokinetics. A complete drug evaluation re-quires definition not only of its pharmacodynamicsbut its pharmacokinetics as well. Important consid-erations are kinetics of absorption, distribution, andclearance (elimination and inactivation); formationof metabolites and their actions; relative potency ofstereoisomers when the drug is a racemic mixture;and pharmacogenetic defects in metabolism. Often,metabolites have major therapeutic or toxic effects.Stereoisomers may have different rates of eliminationand differing, and sometimes opposing, actions. Dis-tribution of drugs into the brain (via the blood-brainbarrier) or other tissues can contribute to toxicity.Direct or indirect interaction with other drugs mayinfluence a drug's effective blood level or therapeuticusefulness. Target specificity plays an important rolehere, as well. As discussed earlier, many antiarrhyth-mic drugs have more than one target, rendering theseactions and interactions exceedingly complex.

Arrhythmogenic Mechanisms in the Celland the Heart

Arrhythmogenic Mechanisms and the Concept of the"Vulnerable Parameter"Thus far, we have reviewed the potential targets of

drug action and placed them in a framework whereinthe mechanisms of drug action might begin to beconsidered at the cellular level. In this section, weconsider the arrhythmia mechanisms themselves anddevelop the framework for classification of drugs. Thearrhythmogenic mechanisms selected to provide thebasis for classification are those most commonly asso-ciated with clinically important tachyarrhythmias.We assume that for each arrhythmogenic mecha-

nism an alteration in one or more of several electro-

physiological properties will be sufficient to terminatethe arrhythmia or to prevent its initiation. In addition,we assume that among the several possible effectivechanges in electrophysiological properties, usually oneis most susceptible to alterations while manifesting aminimum of undesirable effects on the heart. It is thisproperty that we characterize by the term "vulnerableparameter." In the following section, we briefly de-scribe the various arrhythmogenic mechanisms andidentify their most likely vulnerable parameters. Themechanisms and vulnerable parameters are summa-rized at the cellular electrophysiological level in Table3 and for the intact heart in Table 4.Enhanced normtal automaticity. Automaticity is the

property of spontaneous impulse initiation by cardiacfibers. This results from depolarization during phase4, secondary to an inward current carried by Na+ anddesignated 1f9 (Figure 2). The slow depolarizationduring phase 4 can be enhanced by various factors,such as adrenergic stimulation of either the normalsinus node pacemaker or subsidiary pacemakers inspecialized tissues, which may then usurp control ofcardiac rhythm. Modification of the rate of an en-hanced normally automatic focus can result frominterventions that change, singly or in concert, themaximum diastolic potential, the rate of phase 4depolarization, the voltage of the threshold potential,or action potential duration.60The complexity of the relations among these vari-

ables is seen in the fact that a change in one is likelyto change some of the others. For example, anincrease in K+ currents will hyperpolarize the mem-brane (which in itself will slow the automatic rate).However, increasing K+ currents also will shorten theaction potential duration and activate If sooner, andboth changes will increase the rate. Further, increas-ing outward K+ currents may counteract the inwardpacemaker current, thereby decreasing rate. Hence,the ultimate expression of pacemaker rate reflects abalance achieved among a number of interactingfactors, no one of which is of paramount impor-tance.6' Arrhythmias caused by enhanced normalautomaticity that may need treatment are inappro-priate sinus and atrial tachycardias and some accel-erated idioventricular rhythms. For such arrhythmiascaused by enhanced normal automaticity, the vulner-able parameter is phase 4 depolarization.Abnormal automaticity. Abnormal automaticity is

the mechanism by which spontaneous impulses aregenerated in fibers that are partially depolarizedbecause of some pathological process. Atrial andventricular muscle can become automatic by thismechanism, which also may change the basis forautomatic firing in specialized fibers other than thoseof the sinus node. The characteristics of abnormalautomaticity, judged from experimental studies, are afunction of the magnitude of the membrane depolar-ization.62 As described above, at high levels of mem-brane potential in the Purkinje system, the pace-maker current is If, an inward Na' current. Ifmaximum diastolic potential is reduced to -75 to


Dow

nloaded from



TABLE 3. Classification of Arrhythmogenic Mechanisms at the Cellular Level in Terms of Vulnerable Parameters

Ionic currents most likelyVulnerable parameter (antiarrhythmic to modulate vulnerable

,f-chanism of arrhythmia effect) parameter

AutomaticityEnhanced normal automaticity Phase 4 depolarization (decrease) If; 1Ca-T (block)

IK(ACh) (activate)Abnormal automaticity Maximum diastolic potential IK; IK(ACh) (activate)

(hyperpolarize) orPhase 4 depolarization (decrease) 'Ca-L; INa (block)

Triggered activity based onEarly afterdepolarizations (EAD) Action potential duration (shorten) or IK (activate)

EAD (suppress) ICa-L; INa (block)Delayed afterdepolarizations (DAD) Calcium overload (unload) or 1Ca-L (block)

DAD (suppress) ICa-L; INa (blti'k)Reentry dependent on sodium channels

Primary impaired conduction (long Excitability and conduction (decrease) INa (block)excitable gap)Conduction encroaching on Effective refractory period (prolong) IK (block)refractoriness (short excitable gap)

Reentry dependent on calcium channels Excitability and conduction (decrease) 'Ca-L (block)

Other mechanismsReflection Excitability (decrease) INa; 1Ca-L (block)Parasystole Phase 4 depolarization (decrease) If (block) (if maximum

diastolic potential high)

-70 mV, as might occur in the setting of mildischemia, If could still provide some of the phase 4depolarizing current, and the upstrokes of the actionpotentials generated by the abnormally automaticcells would depend largely on sodium entry. If, on theother hand, maximum diastolic potential is reducedto -50 to -55 mV (as the result of a greater degreeof ischemia), the depolarization during phase 4would not be caused by If but rather by the decay ofrepolarizing K' currents, and the action potentialsgenerated, like those of the sinus node, would de-pend primarily on 1Ca.63'64 Examples of arrhythmiasprobably caused by abnormal automaticity are cer-tain ectopic atrial tachycardias65-68 and acceleratedidioventricular rhythms and ventricular tachycardias24-72 hours after experimental myocardial infarc-tion69-71 or during the first few days after clinicalmyocardial infarction.72,73 In these cases, the mostattractive vulnerable parameter is the primary abnor-mality, that is, the reduced maximum diastolic poten-tial or resting potential, which may respond to K'channel activation [e.g., IK(ACh) in atrial cells] or toNa+-K+ pump stimulation. However, from a practicalpoint of view, an additional vulnerable parameter isphase 4 depolarization itself.

Triggered activity. Triggered activity refers to afamily of arrhythmias that arise as a consequence ofafterdepolarizations.74 The occurrence of an afterde-polarization depends on the prior impulse (or seriesof impulses). The same relation holds for triggeredrhythms induced by afterdepolarizations, in whichthe triggered action potentials depend on preceding

events for their initiation. Afterdepolarizations caneither interrupt the process of repolarization oroccur after completion of repolarization and aredesignated respectively as early (EAD) or delayed(DAD) afterdepolarizations (Figure 3).EARLY AFIERDEPOLARIZATIONS. There are several

mechanisms for EADs. All result in a change in netmembrane inward current that delays or interruptsrepolarization. The primary mechanism may be areduction in the normal repolarizing current, IK, anabnormal prolongation of inward current carried bysodium or calcium channels, or the simultaneousoperation of enhanced inward and reduced outwardcurrent (Figure 2).74-76Depending on the level of the transmembrane po-

tential at which a response is triggered, the upstroke ofthat response results primarily from inward sodium orcalcium current. Parenthetically, if an EAD triggers aseries of repetitive responses at a sufficiently positivemembrane potential, the mechanisms may be indistin-guishable from abnormal automaticity.77 EAD mostoften arise at slow heart or stimulation rates or afterlong pauses in the presence of interventions that pro-long the action potential such as decreased extracellu-lar K+ concentrations or drugs that block IK or IK1.Studies in experimental animals indicate that EAD caninduce torsade de pointes,78 and some clinical datasupport this view.79 EAD may also play a role inreperfusion arrhythmias.80The vulnerable parameter for EAD-related trig-

gered activity is the prolonged action potential dura-tion, and the most rational approach is to shorten the


Dow

nloaded from



TABLE 4. Classification of Drug Actions on Arrhythmias Based on Modification of Vulnerable Parameter

Arrhythmia Mechanisms Vulnerable parameter Representative drugsAutomaticity

Enhanced normalInappropriate sinus tachycardia Phase 4 depolarization (decrease) l3-Adrenergic blocking agentsSome idiopathic ventricular Sodium channel blocking agents

tachycardiasAbnormal

Ectopic atrial tachycardia Maximum diastolic potential M2 agonists(hyperpolarization) or

Phase 4 depolarization (decrease) Calcium or sodium channelblocking agents

M2 agonistsAccelerated idioventricular Phase 4 depolarization (decrease) Calcium or sodium channel

rhythms blocking agentsTriggered Activity

EADTorsade de pointes Action potential duration (shorten) P-Agonists; vagolytic agents

or EAD (suppress) (increase rate)Calcium channel blocking agents;

Mg2+; P-adrenergic blockersDAD

Digitalis-induced arrhythmias Calcium overload (unload) or Calcium channel blocking agentsDAD (suppress) Sodium channel blocking agents

Certain autonomically Calcium overload (unload) or P-Adrenergic blocking agentsmediated ventricular DAD (suppress) Calcium channel blocking agents,tachycardias adenosine

action potential (accomplished by increasing heartrate, increasing extracellular K' concentration, useof action potential shortening drugs, ,3-adrenergicactivation, and/or withdrawal of the offending drugs)or to use maneuvers that suppress the triggeringinward currents. Suppression of triggering inwardcurrents may be effected with drugs that block cal-cium and sodium currents, drugs that block 13- ora-adrenergic receptors, and Mg2+.DELAYED AFTERDEPOLARIZATIONS. DAD are

caused by intracellular calcium overload that resultsin repetitive release.of calcium from the sarcoplasmicreticulum.74 The resulting oscillatory changes in in-tracellular calcium activity cause an inward, depolar-izing current that underlies the DAD.74,81 It is uncer-tain whether the inward current is carried by thecalcium-activated nonspecific cation channel (INS)and/or by the sarcolemmal Na/Ca exchanger.82,83

Arrhythmias likely to be DAD-mediated includeextrasystoles and tachycardias related to digitalis ex-cess,'4-86 certain catecholamine-dependent atrial andventricular tachycardias,87'88 and perhaps some arrhyth-mias caused by ischemia and reperfusion.89 The vulner-able parameter is calcium overload, and therefore,DAD-mediated arrhythmias can be countered by anycif various means to reduce intracellular calcium (e.g.,calcium channel blockade). In addition, drugs maysuppress DAD by blocking inward Na+ current carriedby INS or by increasing potassium conductance andthereby increasing outward current.90'91

Reentry. Reentrant excitation can result from cir-cus movement or reflection. The former is the morecommon mechanism. Circus movement is impulse

propagation that traverses a reentrant loop. A circusmovement tachycardia is said to have an excitablegap if at some point during the tachycardia cycle, astimulus can result in premature excitation of thereentrant loop by gaining access to the loop (at thesite of the gap). Our understanding of circus move-ment, and more particularly, the types of paths overwhich circus movement can occur, is incomplete.Therefore, explanations or descriptions must incor-porate simplifying assumptions.92'93 The first is thatthe path for the circus movement that underliesreentrant arrhythmias other than fibrillation ordi-narily must have adequate central and lateral bound-aries to prevent short-circuiting. Second, the lengthof the path must exceed the wavelength determinedby effective refractoriness. Finally, to initiate thecircus movement, there must be unidirectional blockof the initiating impulse.These basic requirements can be satisfied in vari-

ous ways. The central barrier may be anatomical (asin orthodromic reciprocating tachycardia in theWolff-Parkinson-White syndrome), functional (as insome sustained ventricular tachycardias followingmyocardial infarction), or combined. Much evidenceindicates that many electrophysiological parameters(e.g., conduction) are anisotropic, that is, dependenton direction.94 Anisotropy is ordinanrly determinedby fiber orientation such that conduction proceedsmost rapidly along the long axis of fibers rather thantransversely. Anisotropy may become important inthe genesis of some reentrant arrhythmias (Figure 4).The fit between wavelength and path length can beachieved by uniform slowing of conduction along the


Dow

nloaded from



TABLE 4. Continued

Arrhythmia Mechanisms Vulnerable parameter Representative drugs

Reentry(sodium channel dependent)

Long excitable gapAtrial flutter type I Conduction and excitability Atrium: sodium channel blocking

(depress) agents (except lidocaine,mexiletine, tocainide)

Circus movement tachycardia Conduction and excitability Sodium channel blocking agentsin WPW (depress) (except lidocaine, mexiletine,

tocainide)Sustained monomorphic Conduction and excitability Ventricle: sodium channel

ventricular tachycardia (depress) blocking agentsShort excitable gap

Atrial flutter type II Refractory period (prolong) Potassium channel blockersAtrial fibrillation Refractory period (prolong) Potassium channel blockersCircus movement tachycardia Refractory period (prolong) Amiodarone, sotalol

in WPWPolymorphic and sustained Refractory period (prolong) Quinidine, procainamide,monomorphic ventricular disopyramidetachycardia

Bundle branch reentry Refractory period (prolong) Quinidine, procainamidedisopyramide

Ventricular fibrillation Refractory period (prolong) BretyliumReentry

(calcium channel dependent)AV nodal reentrant Conduction and excitability Calcium channel blocking agents

tachycardia (depress)Circus movement tachycardia Conduction and excitability Calcium channel blocking agents

in WPW (depress)Verapamil-sensitive ventricular Conduction and excitability Calcium channel blocking agents

tachycardia (depress)EAD, early afterdepolarization; DAD, delayed afterdepolarization; WPW, Wolff-Parkinson-White syndrome.

entire path, by local slowing of conduction, or byshortening of the effective refractory period in someor all portions of the reentrant pathway. The slowingof conduction may be a result of primary depressionof the excitatory current (INa or Ica) or of delinea-tion(s) arising when tissues in the pathway have onlyincompletely recovered excitability and responsive-ness. Slowing might also result from changes inpassive properties (membrane resistance or gap junc-tional resistance) that adversely influence the spreadof excitatory current. All these mechanisms mightoperate in concert in the same or in different parts ofthe circuit.Depending on the local speed of propagation of the

impulse and the local duration of the effective refrac-tory period, it may be possible to initiate a newimpulse in advance of the circulating wavefront. If thisis possible, the circuit is said to have an excitablegap.95,96 If speed of conduction varies as the impulsetraverses the circuit, there may be an excitable gap atsome sites but not at others. Usually, during theexcitable gap there is not full recovery of excitabilityand thus a prematurely induced impulse will propa-gate more slowly than the primary circulating impulse.

Reentrant excitation due to circus movement canbe terminated in various ways, including prematureactivation, overdrive pacing (impulses entering the

circuit via the excitable gap block both in the ante-grade and retrograde direction during the samecycle), and drugs. Most drugs now used are effectivebecause they block conduction by directly reducing

50

0

mV

-50

-1000 500 1000 1500

mSFIGURF/, 3. Schematic representation of a transmembraneaction potential. Repolarization time course is interrupted byan early afterdepolarization (EAD) and diastole is interruptedby a delayed afterdepolarization (DAD). ---, Repetitive FADS;* -, an action potential upstroke arising from a DAD.


Dow

nloaded from



ANISOTROPIC REENTRY

impulse enteringexcitable gap

FIGURE 4. Schematic representation ofan anisotropic reen-

trant circuit. The line of block is parallel to fiber orientation.The part ofthe circuit that is absolutely refractory is black; thepart that is relatively refractory is stippled, and the part that isfully excitable is white. Conduction parallel to the long fiberaxis (longitudinal conduction) is fast, and conduction perpen-dicular to it (transverse conduction) is slow. Only the twolongitudinal limbs of the circuit have an excitable gap, as

shown by the schematic drawings of the action potentials inboth longitudinal limbs (a and c) and at the two pivotingpoints of transverse conduction (b and d). In b and d, there isno diastolic interval separating repolarization phase and up-stroke of the next action potential, whereas this is present in a

and c. Therefore, an impulse originating outside the circuit canpenetrate the excitable gap only in the longitudinal limbs. Itwill conduct retrogradely and collide with the oncomingreentrant wavefront. Depending on the state of recovery of thetissue in the antegrade direction, the penetrating impulse maycontinue in the circuit (resetting or entraining) or block.

the excitatory current or they delay recovery ofresponsiveness (prolong the refractory period) (seeFigure 5). However, reduction of excitatory currentmay be a two-edged sword (see Figure 6); in areas

with mildly depressed excitability, further impair-ment of excitability may create unidirectional blockand thus facilitate the initiation of reentry. In addi-tion, reduction in conduction velocity in a circuitwithout block will shorten the wavelength and facil-itate reentry. Enhancement of conduction and-under certain circumstances- shortening of refracto-riness also may be antiarrhythmic, but theseproperties do not characterize most currently avail-able drugs.

Fibrillation is characterized by the presence ofmultiple wavefronts of excitation. These multiplewavefronts engage either partially or absolutely re-

fractory tissue, and the changing course of the wave-fronts is dictated by the available corridors of respon-sive tissue. Persistence of fibrillation depends on thepresence of a critical number of simultaneous wave-fronts.97 Prolongation of the effective refractory pe-riod is a beneficial countermeasure (other than elec-trical countershock) because it will reduce thenumber of simultaneously present wavefronts (seeFigure 5).The suppression of reentry generally is accom-

plished by one of the many agents that predictablyand consistently alter conduction and/or refractori-ness, the two vulnerable parameters. Recently, therole of intercellular connections (gap junctions) con-tributing to the slowing and/or block of conduction inreentrant circuits has been emphasized. A relativesparsity of intercellular connections promotes con-duction slowing and/or block in some experimentalmodels of reentry.98 This factor has not been incor-porated explicitly in the classification because nopractical therapeutic intervention has been identi-fied; however, a category for agents that influencegap junctions could be added later.

A Prinmary impaired conduction(!ong excitable gap)

drug acton: depress conduction

C Fibrillation(very short excitable gap)

drug acton: prolong refractoriness

B Conducton encroaching onrefkactoriness (short excitable gap)

drug action: prolong relractoriness

Diagrams Illustrating reentry

Abslute refractoy period

Relative refratory period

FIGURE 5. Schematic representation of re-

entry and the ways to terminate reentiy. Inpanel A, there are impaired conduction andexcitability in a segment of the reentrant cir-cuit (delineated by the two lines). Furtherdepression of conduction and excitability willresult in conduction block in that segment. Asa precondition in panel B, conduction duringthe reentrant tachycardia encroaches on therelative refractory period. Prolongation of re-

fractoriness will cause block of the reentrantwavefront in its own refractory tail (or theimpulse may merely slow down). In panel C,fibrillation is maintained by the presence ofmany independent wavefronts. Prolongationofthe wavelength ofrefractoriness reduces thenumber of wavefronts in a given chamberbelow a critical number; block and collisionterminate the arrhythmia.


Dow

nloaded from



and excitability and conduction fu~rer(unidirecional block) -m bidirectional blo

zone of mildly depressed conduction depress excitabilityand excitability and conduction further

(bidirectoal conduction) - unidirectional block

FIGURE 6. Diagram illustrating how depression of excitabil-ity and conduction may eitherprevent orpromote the initiationofreentry. In the upper left panel, excitability and conductionin a segment are depressed so that unidirectional block sets thestage for induction of reentry by a premature impulse enteringthe circuit. In the upper right panel, excitability and conduc-tion arefurther impaired, creating a zone ofbidirectional blockso that reentry can no longer be initiated. In the lower leftpanel, the segment is only mildly depressed. Bidirectionalconduction is responsible for collision of wavefronts (causingbidirectional block). Further depression creates the setting forunidirectional block (lower right panel).

Other mechanisms. We have included these mech-anisms for the sake of completeness but will notconsider them extensively in terms of vulnerableparameters and drug actions.

REFLECTION. In a nonbranching bundle, an im-pulse returning over the same bundle may result fromreentry based on longitudinal dissociation within thebundle or from reflection.99 Reflection may becaused by electrotonic transmission across an inex-citable segment in a linear bundle. The delay inimpulse transmission, if long enough, allows fibersproximal to the inexcitable segment to recover theirexcitability and be available for excitation by currentflowing retrogradely across the inexcitable zone. Re-flection has been identified in isolated cardiac tissuesstudied using interventions such as the sucrose gap. Itis uncertain whether it occurs in the intact heart.

PARASYSTOLE. Parasystole is caused by a mechanismfor abnormal impulse initiation, usually thought to beautomatic, whose expression is determined by the pres-ence and extent of entrance and exit block.100 Forexample, an automatic focus in a strand of Purkinjefibers may be surrounded by a small rim of inexcitablefibers that are only capable of transmitting electrotonic

currents. Normal sinus impulses are not able to directlyexcite the focus (the focus is "protected") but canproduce subthreshold depolarizations that, dependingon their timing with respect to the intrinsic cycle of theautomatic focus, delay or accelerate automatic firing.Automatic impulses may excite the tissue proximal tothe depressed segment and be a cause for extrasysto-les.100 Whereas parasystole is generally thought to bebenign, it could conceivably provide a trigger for sus-tained reentrant rhythms.

Heart Rate DependencyHeart rate is a critical determinant of the electro-

physiological behavior of normal and abnormal car-diac tissues and their response to drugs. Interven-tions that can be antiarrhythmic at one heart ratemay be ineffective or even provoke arrhythmias at adifferent rate.Whether a premature beat is effective as a trigger

for a reentrant arrhythmia may depend on bothunderlying heart rate and the degree of prematurity.It has been suggested that whether parasystolic im-pulses become manifest depends on the relation ofunderlying heart rate and the intrinsic rate of theectopic focus. Electrotonic transmission and reflec-tion are critically dependent on heart rate; drugs thatimprove transmission across an inexcitable gap mayshift the reflection zone to higher rates and preventreflection at slower rates. On the other hand, drugsthat impair transmission across an inexcitable gapabolish reflected extrasystoles at fast heart or stimu-lation rates and promote them at slow rates.101'102The combined effects of sodium channel blocking

agents and heart rate may explain why in the settingof experimental acute ischemia, lidocaine has beenreported both to be antifibrillatory'03 and to increasethe incidence of ventricular fibrillation.104 In zones ofmildly depressed conduction and excitability, sodiumchannel blockers may at rapid rates create zones ofunidirectional block and thus promote the initiationof reentrant rhythms. In areas of severely depressedexcitability and conduction, such a combination maycreate zones of complete inexcitability that mayprevent reentry or facilitate circus movement byproviding an inexcitable central barrier (see Figure6). Since binding of some drugs to and dissociationfrom their targeted ion channels are well recognizedto be rate- (and voltage-) dependent, the effects ofsodium channel blockers can be expected to be morepronounced at fast rates, especially in partially depo-larized fibers such as those present in acute ischemia.An exception is open-channel binding, which de-creases with increasing inactivation. Drugs that rap-idly increase their binding to (and block of) sodiumchannels with an increase in rate ("fast on") may intheory be more effective in preventing tachycardiaonset, while "slow on" drugs may terminate tachy-cardias more effectively, after many beats.Drugs that prolong action potential duration usu-

ally have their greatest effects at slow rates56 (wherethey also may be toxic, by generating triggered activ-


Dow

nloaded from



ity from early afterdepolarizations). In theory, suchdrugs may become less effective in preventing orterminating reentrant rhythms as the rates of thetachycardias increase; however, experimental datasuggest a prominent antifibrillatory role for drugsthat prolong the action potential.

Arrhythmias and Antiarrhythmic Drugs inthe Clinical Setting

Arrhythmias in the Human Heart: Relation to BasicExperimental Information

For many common clinical arrhythmias (e.g., atrialfibrillation and flutter, AV nodal reentry and circusmovement tachycardia in the Wolff-Parkinson-Whitesyndrome, many ventricular tachycardias and ventric-ular fibrillation) the mechanisms have been generallywell characterized95,96,105 or are sufficiently supportedby data to justify selection of a therapeutic drugregimen. However, in individual cases, especially themajority of ventricular tachycardias, it is rarely if everpossible to be able to define the anatomic andfunctional properties of a reentrant circuit suffi-ciently to permit specific drug selection. Moreover,interpatient variability of drug effects, the effects ofmetabolites, and the effects of disease on drug dis-tribution all can modify drug action. The result is thatthe best that can be expected of a classification is thatit provide a relatively small set of reasonable optionsfor therapy rather than a single optimal choice andbring order and logic to therapy.

Certain clinical arrhythmias may result from morethan one mechanism because different mechanismsmay operate at onset and during maintenance (e.g.,EAD triggering nonsustained ventricular tachycardiaor inducing reentrant ventricular tachycardia or fi-brillation) or because different variants of a mecha-nism may operate in the same arrhythmia (e.g.,reentry with sodium-dependent conduction in onepart of the circuit and calcium-dependent conductionin another part, as seen in circus movement tachy-cardia in the Wolff-Parkinson-White syndrome). Forany given ventricular tachycardia it may not be clearwhether conduction is primarily impaired in differentparts of the circuit, with a long excitable gap, orimpaired because of encroachment on refractoriness,or both. There is evidence that some ventriculartachycardias depend on calcium current for reentrantexcitation, others on EAD or DAD. Thus, in Table 4ventricular tachycardia appears in various categories.We recognize that the information contained in

Table 4 is complex and not readily memorized.However, the purpose of our deliberations was not toprovide an overly simplified system for the classifica-tion of drug actions, but rather to provide a systemthat permits rational understanding to promote re-search and systematic analyses. Nonetheless, theactual mechanisms whereby drugs may act -in lightof a vulnerable parameter - can be considered insimpler fashion than that which is demonstrated inTables 3 and 4, because the mechanisms for arrhyth-

TABLE 5. Classification of Arrhythmogenic Mechanisms

AutomaticityEnhanced normalAbnormal

TriggeringEADDAD

Na+-Dependent reentryPrimary impaired conduction (long excitable gap)Conduction impaired by refractoriness (short excitable gap)

Ca2'-Dependent reentryReflectionParasystole

EAD, early afterdepolarization; DAD, delayed afterdepolariza-tion.

mias are rather few. Hence, the classification of drugactions in Tables 3 and 4 can be considered in light ofthe mechanisms as they are summarized in Table 5.In considering Table 5, it must be understood thatthe use of a term in relation to an antiarrhythmicaction would imply an action countering the mecha-nism, not an action promoting it. For convenient use,each drug action could be described by a short termor expression, for example, drugs countering "auto-maticity," "triggering," "Na' reentry," and "Ca2+reentry." The separation of the sodium-dependentreentry class into subdivisions, depending on whetherexcitatory current or refractoriness is the critical (andvulnerable) parameter, is based on the long-appreci-ated critical roles of conduction and refractoriness inthe mechanism of reentry, the ready availability ofagents that affect predominantly each of these prop-erties by fundamental actions on sodium or potas-sium channels, and the separation of these classes ofagents in the Vaughan Williams classification. It isrecognized that often both conduction and refracto-riness may be "critical" in the same circuit; often it isnot possible to determine the critical parameter in anindividual case, especially in ventricular tachycardia;and even when encroachment on refractoriness iscritical, agents depressing excitatory currents may beeffective. Thus, selection for the group "sodium-reentry-long gap" (sodium channel blockers) versus"sodium-reentry-short gap" (potassium channelblockers) will often be arbitrary. However, thesecategories, so designated, may encourage clinicalinvestigations of the responses of the circuit to de-termine the critical parameter, for example, by ana-lyzing the responses of appropriate sites to pro-grammed stimulation.

Clinical Approach to the Use ofAntiarrhythmic DrugsTo a large extent, the management of clinical car-

diac arrhythmias requires the use of antiarrhythmicdrugs. Nonetheless, concerns about the long-termefficacy, safety, and cost-effectiveness of pharmacolog-ical therapy106-108 have resulted in a movement towardthe use of nonpharmacological methods of treat-


Dow

nloaded from



ment,109,110 such as those based on devices or ablationmanagement strategies for selected arrhythmias.However, drugs are likely to remain in a mainstay oftherapy for the majority of patients with cardiacarrhythmias.Today the physician may choose from among 30 to

80 agents being tested or marketed for the control ofcardiac arrhythmias, and in the near future many newdrugs will become available. Each of these drugs hasa unique profile of action, and none is uniformlyeffective and free of adverse effects. In clinical prac-tice, drugs are chosen in order to achieve efficacywithout encountering unacceptable side effects. Thebasis for this choice may be past experience, fashion,protocol, or antiarrhythmic drug classification.111,112At times the physician may know the mechanism of

the tachycardia and can design specific therapy. Forexample, the regular, narrow QRS complex tachycar-dia usually seen in association with the Wolff-Park-inson-White syndrome is due to reentry over a largecircuit with two potentially vulnerable segments (AVnode and accessory atrioventricular connection),which may be assessed by clinical electrophysiologi-cal techniques. The mechanism that triggers initia-tion of the tachycardia (atrial or ventricular ectopicactivity, etc.) can sometimes be assessed by long-termECG monitoring. Armed with this information, theclinician can prescribe therapy specifically to abolishthe trigger mechanism or, preferably, adjust thesubstrate to prevent initiation or continuation ofreentry. The latter may be achieved by choosingdrugs that impair slowAV nodal conduction (calciumchannel dependent), such as digitalis glycosides,,3-adrenergic blockers or calcium channel blockers,or drugs that retard fast conduction through theaccessory atrioventricular connection (sodium chan-nel dependent) such as flecainide or procainamide.The mechanisms of many other arrhythmias are

less clear.112 For example, atrial tachycardia may becaused by reentry, abnormal automaticity, or trig-gered activity. Clinical clues may suggest a specificmechanism toward which therapy may be targeted,but if the mechanism remains in doubt, the physicianmay opt to offer treatment speculatively, on the basisof the relative likelihood of the operation of aparticular mechanism.To choose a particular drug or drug combination to

treat specific arrhythmias, the physicians must notonly understand the putative mechanism of the ar-rhythmia (see for example Table 3) but must alsoappreciate the wide range of antiarrhythmic drugactions that are available as shown in Table 4. Thetabulated presentation of the drug actions in Table 4has been enlarged on in Figure 7. This listing incor-porates several important principles:

1) The listing is flexible-no numbers or lettersare used as shorthand descriptors of drug actions.

2) The drugs are not placed within one or morecells of a catalogue. Rather, they are listed in the firstcolumn of a row and their important actions areindicated in the other cells in that row.

3) The vertical order of the antiarrhythmic drugsmay be changed to emphasize different drug clusters.For example, drugs that act at the ,-receptor couldbe placed together, and the different actions of thesedrugs on other receptors, channels, and pumps wouldthen become more readily apparent. Hence, one cancollate specific major drug actions in light of aparticular vulnerable parameter and then considerthe associated actions of the various drugs and therisks or benefits these may entail.

Concluding RemarksIn summary, the collation of clinical and electro-

cardiographic information allows the physician toreach a descriptive diagnosis of a cardiac arrhythmia.This diagnosis may imply, with variable certainty, anarrhythmia mechanism that suggests a particulartreatment. Effective pharmacological therapy re-quires that the physician attempt to identify a drugwith the most appropriate profile to attack the mostvulnerable parameter of the mechanisms of the car-diac arrhythmia. As information about the mecha-nisms of clinical cardiac arrhythmias becomes morecomplete, the actions of antiarrhythmic drugs aremore fully understood, and the number and variety ofantiarrhythmic drugs increases, specific pharmaco-logical management of cardiac arrhythmias will pro-gressively improve. In the meantime, the series oftables and figures provided here, incorporating thebeginnings of a classification based on arrhyth-mogenic mechanism and the identification of vulner-able parameters, provides a basis not only for ratio-nal consideration of arrhythmias and their therapy,but for communication among basic and clinicalinvestigators and practicing physicians. Hence, agroundwork for continued growth and incorporationof knowledge is provided.

AcknowledgmentsThe authors express their gratitude to Eileen

Franey and Pinuccia De Tomasi for their carefulattention to the preparation of the manuscript.

AppendixThis paper summarizes the outcome of a workshop

held in Taormina, Sicily, December 1-4, 1990. Theneed for the workshop was proposed by Peter J.Schwartz and Michiel J. Janse. It was organized bythe Working Group on Arrhythmias of the EuropeanSociety of Cardiology, with joint sponsorship by theBasic Science Council of the American Heart Asso-ciation and the American College of Cardiology. Itsfunding was administered by the European Society ofCardiology. The workshop was chaired by Michael R.Rosen. Participating in the workshop and coauthor-ing the paper were J. Thomas Bigger Jr., GunterBreithardt, Arthur M. Brown, A. John Camm, Ed-ward Carmeliet, Harry A. Fozzard, Brian F. Hoff-man, Michiel J. Janse, Ralph Lazzara, AlessandroMugelli, Robert J. Myerburg, Dan M. Roden,Michael R. Rosen, Peter J. Schwartz, Harold C.


Dow

nloaded from



0: AqstfA=Ag~t/Antag.

A : Activated stat blocker1:ahtvated state blocker

FIGURE 7. Summary of the potentially most important actions ofdrugs on membrane channels, receptors, and ionicpumps in theheart. Included are examples ofdrugs used to modify cardiac rhythm. Most are already marketed as antiarrhythmic agents, but someare not yet approved for this purpose. The drugs (rows) are ordered in a fashion similar to the columns so that generally the entriesfor theirpredominant action(s) forn a diagonaL Drugs with multiple actions (e.g., amiodarone) depart strikingly from the diagonaltrend. The actions of drugs on the sodium, calcium, potassium (IKJ, and If channels are indicated. Sodium channel blockade issubdivided into three groups of actions characterized by fast (<300 msec), medium (Med) (200-1,500 msec) and slow (>1,500msec) time constants for recoveryfrom block. Thisparameter is a measure ofuse dependence andpredicts the likelihood that a drugwill decrease conduction velocity ofnormal sodium-dependent tissues in the heart andperhaps the propensity ofa drugfor causingbundle-branch block orproarrhythmia. The rate constantfor onset ofblock might be even more clinically relevant. Blockade in theinactivated (I) or activated (A) state is indicated. Information on the state dependency of the block caused by moricizine,propefenone, encainide, and flecainide is especially limited and may be altered with additional research. Drug interaction withreceptors (a; A muscarinic subtype 2 [M2], andA]purinergic [P]) and drug effects on the sodium/potassiumpump (Na/K, A TPase)are indicated. Filled circles indicate antagonist or inhibitory actions; unfilled circles indicate direct or indirect acting agonists or

stimulators. The darkness ofthe symbol increases with the intensity ofthe action. Half-filled circlesfor bretylium indicate its biphasicaction to initially stimulate a- and [-receptors by release ofnorepinephrine followed by subsequent block ofnorepinephrine releaseand indirect antagonism of these receptors.

Strauss, Raymond L. Woosley, and Antonio Zaza.Not attending but acting as advisors were RonaldW.F. Campbell and Albert L. Waldo. The finalpreparation and organization of the manuscript werethe responsibility of A. John Camm, Harry A. Foz-zard, Michiel J. Janse, Ralph Lazzara, Michael R.Rosen, and Peter J. Schwartz.

References1. Vaughan Williams EM: Classification of antiarrhythmic drugs,

in Sandoe E, Flensted-Jensen E, Olesen K (eds): CardiacArrhythmias. Sweden, Astra, Sodertaije, 1981, pp 449-472

2. Harrison DA, Winkle RA, Sami M, Mason J: Encainide: a

new and potent antiarrhythmic agent, in Harrison DC (ed):Cardiac Arrhythmias: A Decade ofProgress. Boston, Mass, GKHall, 1981, pp 315-330

CHANELS RECEPTORS PUMPSDRUG NCaK a| MaIPLdcie @_Mexhle 0Tocakide 0

Procamde *Disoyrarde 0 0,

Propfenne* e

Bretylu ;

Sotaio

AEkc*ine 0

Nadolol:

AtropinID

Adenosem 0

Dlqoxin _ 0 0


Dow

nloaded from



3. Vaughan William EM: A classification of antiarrhythmicactions reassessed after a decade of new drugs. J ClinPharmacol 1984;24:129-147

4. Catterall WA: Structure and function of voltage-sensitive ionchannels. Science 1988;242:50-61

5. Cribbs LL, Stain J, Fozzard HA, Rogart RB: Functionalexpression of the rat heart I Na+ channel isoform. FEBS Lett1990;275:195-200

6. Brown HF: Electrophysiology of the sinoatrial node. PhysiolRev 1982;62:505-530

7. Hagiwara N, Irisawa H, Kasanuki H, Hosoda S: Backgroundinward current sensitive to sodium ion in the sinoatrial nodeof rabbit heart. J Physiol (Lond) (in press)

8. Pelzer D, Peizer S, MacDonald TF: Properties and regula-tion of calcium channels in muscle cells. Rev Physiol BiochemPharnacol 1990;114:107-207

9. DiFrancesco D, Ferroni A, Mazzanti M, Tromba C: Proper-ties of the hyperpolarizing-activated current (if) in cellsisolated from the rabbit sinoatrial node. J Physiol (Lond)1986;377:61-88

10. Ehara T, Noma A, Ono K: Calcium activated non-selectivecation channel in ventricular cells isolated from adult guineapig heart. J Physiol (Lond) 1988;403:117-133

11. Giles WR, Imaizumi Y: Comparison of potassium currents inrabbit atrial and ventricular cells. J Physiol (Lond) 1988;405:123-145

12. Sanguinetti MC, Jurkiewicz NK: Two components of cardiacdelayed rectifier K current: Differential sensitivity to block byclass III antiarrhythmic agents. J Gen Physiol 1990;96:195-215

13. Walsh KB, Kass RB: Regulation of a heart potassium chan-nel by protein kinase A and C. Science 1988;242:67-69

14. Coraboeuf E, Carmeliet E: Existence of two transient out-ward currents in sheep cardiac Purkinje fibers. Pflugers Arch1982;392:352-359

15. Sakmann B, Noma A, Trautwein W: Acetylcholine activationof single muscarinic K channels in isolated pacemaker cells ofthe mammalian heart. Nature 1983;303:250-253

16. Carmeliet E, Storms L, Vereecke J: The ATP-dependent Kchannel and metabolic inhibition, in Zipes DP, Jalife J (eds):Cardiac Electrophysiology: From Cell to Bedside. Philadelphia,WB Saunders Co, 1990, pp 103-108

17. Harvey RD, Clark CD, Hume JR: Chloride current inmammalian cardiac myocytes. J Gen Physiol 1990;95:1077-1102

18. Page E, Manjunath CK: Communicating junctions betweencardiac cells, in Fozzard HA, Haber E, Jennings RB, KatzAM, Morgan HE (eds): Heart and the Cardiovascular System.New York, Raven Press, Publishers, 1986, pp 573-600

19. Nakai J, Imagawa, Hakamata Y, Shigekawa M, Takeshima H,Numa S: Primary structure and functional expression fromcDNA of the cardiac ryanodine receptor/calcium releasechannel. FEBS Lett 1990;271:169-177

20. Gadsby DC: The Na/K pump of cardiac cells. Ann RevBiophys Bioeng 1984;13:373-378

21. Sheu S-S, Blaustein MP: Sodium/calcium exchange and reg-ulation of cell calcium and contractility in cardiac muscle,with a note about vascular smooth muscle, in Fozzard HA,Haber E, Jennings RB, Katz AM, Morgan HE (eds): TheHeart and Cardiovascular System. New York, Raven Press,Publishers, 1986, pp 509-535

22. Gilman AG: G proteins: Transducers of receptor-generatedsignals. Ann Rev Biochem 1987;56:615 -649

23. Jones CR, Molenaar P, Summers RJ: New views of humancardiac beta-adrenoceptors. J Mol Cell Cardiol 1989;21:519-535

24. Hescheler J, Trautwein W: Modulation of calcium current ofventricular cells, in Piper HM, Isenberg G (eds): IsolatedAdult Cardiomyocytes. Boca Raton, Fla, CRC Press Inc, 1989,pp 128-154

25. DiFrancesco D: The cardiac hyperpolarizing-activated cur-rent, if: Origins and developments. Prog Biophys Molec Biol1985;46:163-183

26. Yatani A, Okabe K, Codina J, et al: Heart rate regulation byG proteins acting on the cardiac pacemaker channel. Science1990;249:1163-1166

27. Yazawa K, Kameyama M: Mechanism of receptor-mediatedmodulation of the delayed outward potassium current inguinea pig ventricular myocytes. Am J Physiol 1990;428:135-150

28. Harvey RD, Hume JR: Isoproterenol activates a chloridecurrent, not the transient outward current, in rabbit ventric-ular myocytes. Am J Physiol 1989;257:C1177-C1181

29. del Balzo U, Rosen MR, Malfatto G, Kaplan LM, SteinbergSF: Specific a1-adrenergic receptor subtypes modulate cate-cholamine induced increases and decreases in ventricularautomaticity. Circ Res 1990;67:1535-1551

30. Shah A, Cohen IS, Rosen M: Stimulation of cardiac alphareceptors increases Na/K pump current and decreases gk viaa pertussis toxin sensitive pathway. Biophys J 1988;54:219-225

31. Apkon M, Nerbonne JM: Alpha1-adrenergic agonists selec-tively suppress voltage dependent K' currents in rat ventric-ular myocytes. Proc Natl Acad Sci USA 1988;85:8756-8760

32. Fedida DY, Shimoni S, Giles WR: A novel effect of norepi-nephrine on cardiac cells is mediated by a, adrenoceptors.Am J Physiol 1989;256:H1500-H1504

33. Hamra M, Rosen MR: Alpha-adrenergic receptor stimula-tion during simulated ischemia and reperfusion in caninecardiac Purkinje fibers. Circulation 1988;78:1495-1502

34. Kimura S, Cameron JS, Kozlovskis PL, Bassett AL, MyerburgRJ: Delayed afterdepolarizations and triggered activityinduced in feline Purkinje fibers by alpha-adrenergic stimu-lation in the presence of elevated calcium levels. Circulation1984;70:1074-1082

35. Sheridan DJ, Penkoske PA, Sobel BE, Corr PB: Alphaadrenergic contributions to dysrhythmia during myocardialischemia and reperfusion in cats. J Clin Invest 1980;65:161-171

36. Schwartz PJ, Locati EH, Moss AJ, Crampton RS, Trazzi R,Ruberti U: Left cardiac sympathetic denervation in thetherapy of congenital long QT syndrome. Circulation (inpress)

37. Peralta EG, Ashkenazi A, Winslow JW, Smith DH, Ra-machandran J, Capon DJ: Distinct primary structures, li-gand-binding properties and tissue-specific expression of fourhuman muscarinic acetylcholine receptors. EMBO J 1987;6:3923-3929

38. Fields JZ, Roeske WR, Morkin E, Yamamura HI: Cardiacmuscarinic cholinergic receptors: Biochemical identificationand characterization. J Biol Chem 1978;253:3251-3258

39. Yatani A, Codina J, Brown AM, Birnbaumer L: Directactivation of mammalian atrial muscarinic potassium chan-nels by GTP regulatory protein GK. Science 1987;235:207-211

40. Lindemann JP, Watanabe AM: Mechanisms of adrenergicand cholinergic regulation of myocardial contractility, inSperelakis N (ed): Physiology and Pathophysiology of theHeart, ed 2. Boston, Mass, Kluwer Academic Publishers,1989, pp 423-452

41. DiFrancesco D, Ducouret P, Robinson RB: Muscarinic mod-ulation of cardiac rate at low acetylcholine concentrations.Science 1989;243:669-671

42. Vanoli E, DeFerrari GM, Stramba-Badiale M, Hull SS Jr,Foreman RD, Schwartz PJ: Vagal stimulation and preventionof sudden death in conscious dogs with a healed myocardialinfarction. Circ Res 1991;68:1471-1481

43. Kleiger RE, Miller JP, Bigger JT Jr, Moss AJ: The Multi-center Post-Infarction Research Group: Decreased heartrate variability and its association with increased mortalityafter acute myocardial infarction. Am J Cardiol 1987;59:256-262

44. LaRovere MT, Specchia G, Mortara A, Schwartz PJ: Barore-flex sensitivity, clinical correlates and cardiovascular mortal-ity among patients with a first myocardial infarction: Aprospective study. Circulation 1988;78:816-824


Dow

nloaded from



45. Kirsch GE, Codina J, Birnbaumer L, Brown AM: Coupling ofATP-sensitive K' channels to A1 receptors by G proteins inrat ventricular myocytes. Am J Physiol 1990;259:H820-H826

46. Belhassen B, Pelleg A, Shoshani D, Geva B, Laniado S:Electrophysiologic effects of adenosine-5'-trisphosphate onatrioventricular reentrant tachycardia. Circulation 1983;68:827-833

47. Strauss HC: Mechanisms of local anesthetic interaction with thesodium channel, in Rosen MR, Janse MJ, Wit AL (eds):Cardiac Electrophysiology: A Textbook in Honor of Bran F.Hoffman. Mt Kisco, NY, Futura Publishing Co, 1990, pp995-1012

48. Brown AM: Modulation of Ca2' channels, in Rosen MR,Janse MJ, Wit Al (eds): Cardiac Electrophysiology:A Textbookin Honor of Brian F. Hoffiman. Mt Kisco, NY, Futura Pub-lishing Co, 1990, pp 1031-1042

49. Grant AO, Starmer CF, Strauss HC: Antiarrhythmic drugaction: Blockade of the inward sodium current. Circ Res1984;55:427-439

50. Hondeghem LM, Katzung BG: Antiarrhythmic agents: Themodulated receptor mechanism of action of sodium andcalcium channel blocking drugs. Ann Rev Pharmacol Toxicol1984;24:387-423

51. Starmer CF, Grant AO: Phasic ion channel blockade. Akinetic model and parameter estimation procedure. MolPharmacol 1985;28:348-356

52. Courtney KR: Review: Quantitative structure/activity rela-tions based on use-dependent block and repriming kinetics inmyocardium. J Mol Cell Cardiol 1987;19:319 -330

53. Hondeghem LM, Katzung BG: Time- and voltage-dependentinteractions of antiarrhythmic drugs with cardiac sodiumchannels. Biochim Biophys Acta 1977;472:373-398

54. Starmer CF, Grant AO, Strauss HC: Mechanisms of use-dependent block of sodium channels in excitable membranesby local anesthetics. Biophys J 1984;46:15-27

55. Gilliam FR III, Starmer CF, Grant AO: Blockade of rabbitatrial sodium channels by lidocaine: Characterization ofcontinuous and frequency-dependent blocking. Circ Res 1989;65:723-739

56. Hondeghem LM, Snyders DJ: Class III antiarrhythmic agentshave a lot of potential but a long way to go: Reducedeffectiveness and dangers of reverse-use dependence. Circu-lation 1990;81:686-690

57. Gintant GA, Hoffman BF: Use-dependent block of cardiacsodium channels by quaternary derivatives of lidocaine.Pflugers Arch 1984;400:121-129

58. Colatsky TJ, Follmer CH, Starmer CF: Channel specificity inantiarrhythmic drug action: Mechanism of potassium channelblock and its role in suppressing and aggravating cardiacarrhythmias. Circulation 1990;82:2235-2242

59. Clarkson CW, Hondeghem LM: Evidence for a specificreceptor site for lidocaine, quinidine, and bupivicaine asso-ciated with cardiac sodium channels in guinea pig myocar-dium. Circ Res 1985;56:496-506

60. Whitcomb DC, Gilliam FR III, Starmer CF, Grant AO:Marked QRS complex abnormalities and sodium channelblockade by propoxyphene reversed with lidocaine. J ClinInvest 1989;84:1629-1636

61. Hoffman BF, Cranefield PF: Electrophysiology of the Heart.New York, McGraw-Hill, 1960, pp 104-174

62. Dangman KH, Hoffman BF: Studies on overdrive stimulationof canine cardiac Purkinje fibers: Maximal diastolic potentialas a determinant of the response. JAm Coll Cardiol 1983;6:1183-1190

63. Grant AO, Katzung BG: The effects of quinidine and vera-pamil on electrically-induced automaticity in the ventricularmyocardium of guinea pig. J Pharn Exp Ther 1976;196:407-419

64. Imanishi S: Calcium-sensitive discharge in canine Purkinjefibers. Jpn J Physiol 1971;21:443-463

65. Gillette PC, Garson A Jr: Electrophysiologic and pharmaco-logic characteristics of automatic ectopic atrial tachycardia.Circulation 1977;56:571-575

66. Goldreyer BN, Gallagher JJ, Damato AN: The electrophys-iologic demonstration of atrial ectopic tachycardia in man.Am Heart J 1973;85:205-215

67. Scheinman MM, Basu D, Hollenberg M: Electrophysiologicstudies in patients with persistent atrial tachycardia. Circula-tion 1974;50:266-273

68. Rossi L, Matturri L: ClinicopathologicalApproach to CardiacArrhythmias. Torino, Italy, Centro Scientifico, Torinese Edi-tore, 1990, p 176

69. Scherlag BJ, El-Sherif N, Hope RR, Lazzara R: Character-ization and localization of ventricular arrhythmias resultingfrom myocardial ischemia and infarction. Circ Res 1974;35:372-383

70. Friedman DL, Steward JR, Wit AL: Spontaneous andinduced cardiac arrhythmias in subendocardial Purkinjefibers surviving extensive myocardial infarction. Circ Res1973;32:612-626

71. Lazzara R, El-Sherif N, Scherlag BJ: Electrophysiologicalproperties of canine Purkinje cells in one day old myocardialinfarction. Circ Res 1973;33:722-734

72. Norris RM, Mercer CJ: Significance of idioventricularrhythms in acute myocardial infarction. Prog Cardiovasc Dis1974;16:455-468

73. Sclarovsky S, Straberg B, Fuchs J, Lewin RF, Arditis A,Klainman E, Kracoff OH, Achmon J: Multiform acceleratedidioventricular rhythm in acute myocardial infarction: Elec-trocardiographic characteristics and response to verapamil.Am J Cardiol 1983;52:43-47

74. Cranefield PF, Aronson RS: in Cardiac Arrhythmias. The Roleof Triggered Activity and Other Mechanisms. Mt Kisco, NY,Futura Publishing Co, 1988

75. Peper K, Trautwein W: Ueber den Mechanismus der Extra-systolen des Aconitin-vergifteten Herzmuskels. Pflugers Arch1966;291:R16

76. Isenberg G: Cardiac Purkinje fibers: Cesium as a tool to blockinward rectifying potassium currents. Pflugers Arch 1976;365:99-106

77. Damiano BP, Rosen MR: Effects of pacing on triggeredactivity induced by early afterdepolarizations. Circulation1984;69:1013-1025

78. Brachman J, Scherlag BJ, Rosenshtraukh LV, Lazzara R:Bradycardia-dependent triggered activity: Relevance to drug-induced multiform ventricular tachycardia. Circulation 1983;68:846-856

79. Krikler DM, Curry PVL: Torsade de pointes, an atypicalventricular tachycardia. Br Heart J 1976;38:117-120

80. Priori SG, Mantica M, Napolitano C, Schwartz PJ: Earlyafterdepolarizations induced in vivo by reperfusion of isch-emic myocardium: A possible mechanism for reperfusionarrhythmias. Circulation 1990;81:1911-1920

81. Vassalle M, Mugelli A: An oscillating current in sheepcardiac Purkinje fibers. Circ Res 1981;48:618-631

82. Kass RS, Lederer WJ, Tsien RW, Weingart R: Role ofcalcium ions in transient inward currents and aftercontrac-tions induced by strophantidin in cardiac Purkinje fibers.J Physiol 1987;281:187-208

83. Eisner DA, Lederer WJ: Na-Ca exchange: Stoichiometry andelectrogenicity. Am J Physiol 1985;248:C189-C202

84. Rosen MR, Gelband HB, Hoffman BF: Correlation betweeneffects of ouabain on the canine electrocardiogram andtransmembrane potentials of isolated Purkinje fibers. Circu-lation 1973;47:65-72

85. Ferrier GR, Saunders JH, Mendez C: A cellular mechanismfor the generation of ventricular arrhythmias by acetylstro-phanthidin. Circ Res 1973;32:600-609

86. Gorgels APM, de Wit B, Beekman HDM, Dassen WRM,Wellens HJJ: Triggered activity induced by pacing duringdigitalis intoxication. PACE 1987;10:1309-1321

87. Wit AL, Cranefield PF: Triggered and automatic activity inthe canine coronary sinus. Circ Res 1977;41:435-445

88. Malfatto G, Rosen TS, Rosen MR: The response to over-drive pacing of triggered atrial and ventricular arrhythmias inthe canine heart. Circulation 1988;77:1139-1148


Dow

nloaded from



89. Ferrier GR, Moffat MP, Lukas A: Possible mechanisms ofventricular arrhythmias elicited by ischemia followed byischemia followed by reperfusion: Studies on isolated canineventricular tissues. Circ Res 1985;56:184-194

90. Aronson RS, Cranefield PF, Wit AL: The effects of caffeineand ryanodine on the electrical activity of the coronary sinus.JPhysiol 1985;368:593-610

91. Rosen MR, Danilo P Jr: Effects of tetrodotoxin, lidocaine,verapamil and AHR-2666 on ouabain-induced delayed after-depolarizations in canine Purkinje fibers. Circ Res 1980;46:117-124

92. Wit AL, Cranefield PF: Reentrant excitation as a cause ofcardiac arrhythmias.Am JPhysiol 1978;235(Heart Circ Physiol4):H1-H17

93. Janse MJ, Wit AL: Electrophysiological mechanisms of ven-

tricular arrhythmias resulting from myocardial ischemia andinfarction. Physiol Rev 1989;69:1049-1169

94. Spach MS, Miller WT, Geselowitz DB, Barr RC, KootseyJM, Johnson EA: The discontinuous nature of propagation innormal canine cardiac muscle: Evidence for recurrent discon-tinuities of intracellular resistance that effect membranecurrents. Circ Res 1981;48:39-54

95. Waldo AL, Ohlshansky B, Okumura K, Henthorn RW: Currentperspective on entrainment of tachycardias, in Brugada P,Wellens HJJ (eds): Cardiac Arrhythmias: Where to Go FromHere? Mt Kisco, NY, Futura Publishing Co, 1987, pp 171-189

96. Waldo AL, Henthorn RW, Plumb VJ, MacLean WAH:Demonstration of the mechanism of transient entrainmentand interruption of ventricular tachycardia with rapid atrialpacing. JAm Coll Cardiol 1984;3:422-430

97. Allessie MA, Lammers WJEP, Bonke FIM, Hollen J: Exper-imental evaluation of Moe's multiple wavelet hypothesis ofatrial fibrillation, in Zipes DP, Jalife J (eds): CardiacArrhyth-mias. Orlando, Fla, Grune and Stratton, 1985, pp 265-276

98. Dillon S, Allessie WA, Ursell PC, Wit AL: Influence ofanisotropic tissue structure on reentrant circuits in thesubepicardial border zone of subacute canine infarcts. CircRes 1988;63:182-206

99. Antzelevitch C: Electrotonus and reflection, in Rosen MR,Janse MJ, Wit AL (eds): Cardiac Electrophysiology: A Text-book in Honor of Brian F Hoffman. Mt Kisco, NY, FuturaPublishing Co, 1990, pp 491-516

100. Jalife J, Moe GK: Effect of electrotonic potentials on pace-

maker activity of canine Purkinje fibers in relation to para-

systole. Circ Res 1976;39:801-808101. Davidenko JM, Antzelevitch C: The effect of milrinone on

conduction, reflection and automaticity in canine Purkinjefibers. Circulation 1984;69:1026-1035

102. Shen X, Antzelevitch C: Mechanism underlying the antiar-rhythmic and arrhythmogenic action of quinidine in a Pur-kinje fiber-Ischemic gap preparation of reflected reentry.Circulation 1986;73:1342-1352

103. Cardinal R, Janse MJ, van Eede I, Werner G, Naumann d'Alnoncourt C, Durrer D: The effect of lidocaine on intracel-lular and extracellular potentials, activation and ventriculararrhythmias during acute regional ischemia in the isolatedporcine heart. Circ Res 1981;49:792-806

104. Carson DL, Cardinal R, Savard P, Vasseur C, Nattel S,Lambert C, Nadeau R: Relationship between an arrhyth-mogenic action of lidocaine and its effects on excitationpatterns in acutely ischemic porcine myocardium. J Cardio-vasc Pharmacol 1986;8:126-136

105. Brugada P, Wellens HJJ: The role of triggered activity inclinical ventricular arrhythmias. PACE 1984;7:260-271

106. The Cardiac Arrhythmia Suppression Trial (CAST) investi-gators: Preliminary report: Effect of encainide and flecainideon mortality in a randomized trial of arrhythmia suppressionafter myocardial infarction. N Engi J Med 1989;21:406-412

107. Coples SE, Antman EM, Berlin JA, Hewitt P, Chalmers TC:Efficacy and safety of quinidine therapy for maintenance ofsinus rhythm after cardioversion: A meta-analysis of random-ized controlled trials. Circulation 1990;82:1106-1116

108. Surawicz B: Prognosis of ventricular arrhythmias in relationto sudden death: Therapeutic implications. JAm Coll Cardiol1987;10:435-447

109. Saksena S, Goldschlager N (eds): Electrical Therapy for CardiacArrhythmias. Philadelphia, WB Saunders Co, 1990

110. Fisher JD, Brodman RF, Kim SG, Ferrick KJ, Roth JA:VT/VF: 60/60 protection. PACE 1990;13:218-222

111. Wellens HJJ, Brugada P: Treatment of cardiac arrhythmias:When, how, and where?JAm Coll Cardiol 1989;14:1417-1428

112. Zipes DP: Cardiac electrophysiology: Promises and contribu-tions. JAm Coll Cardiol 1989;13:1329-1341

KEY WoRDs * Point of View * antiarrhythmic drug classification* vulnerable parameter


Dow

nloaded from


Arrhythmias of the European Society of Cardiology.on their actions on arrhythmogenic mechanisms. Task Force of the Working Group on

The Sicilian gambit. A new approach to the classification of antiarrhythmic drugs based

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1991 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.84.4.1831

1991;84:1831-1851Circulation.

http://circ.ahajournals.org/content/84/4/1831the World Wide Web at:

The online version of this article, along with updated information and services, is located on

http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer information about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. FurtherEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculationpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:


Dow

nloaded from

http://circ.ahajournals.org/content/84/4/1831

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/


the sicilian gambit - circulationcirc.ahajournals.org/content/circulationaha/84/4/1831.full.pdfthis...

Documents