Arrhythmogenic actions of antiarrhythmic drugs

ArrhythmogenicActions of Antiarrhythmic Drugs MICHAEL R. ROSEN, MD, Cardiac arrhythmias may result from abnormalities of impulse propagation or abnormalities of impulse initiation. When arrhythmias are initiated by antiarrhythmic drugs, the most common mechanisms appear to be (1) conduction block or reentry, and (2) abnormal impulse initiation, which may be triggered by afterdepolarizations. The effects of drugs on conduction may result from their actions on the fast sodium channel, the slow calcium channel or their ability to prolong repolarization. The extent to which a drug that depresses fast sodium or slow calcium entry will exert its toxic effects depends in large part on its binding characteristics to its channel receptor site. Such toxicity represents a continuum for the therapeutic effects of these drugs. The factors that control drug access to binding sites, including and ANDREW L. WIT, PhD lipid solubility, molecular size and extent of ionization, are reviewed, as are the contributions to conduction abnormalities of drug-induced changes in repolarization. The mechanisms whereby drugs induce abnormalities of impulse initiation are still a matter of conjecture. Apparently, drugs that increase inward plateau currents or decrease repolarizing potassium ion currents carry increased risk. Moreover, there is evidence for the role of early after depolarizations occurring secondary to prolonged repolarization as a possible cause of arrhythmias, including torsades de pointes. The mechanisms whereby antiarrhythmic drugs may contribute to this type of tachyarrhythmia are reviewed. - C (Am J Cardiol 1987;59:10E-18E) Although it is conceivable that arrhythmias induced by antiarrhythmic drugs might result from any or all of these mechanisms, there is only limited information concerning the cellular electrophysiologic basis for such arrhythmias. Therefore, in this article we concentrate only on those areas in which the combination of cellular electrophysiologic information and clinical evidence of arrhythmias enables us to posit a realistic, rather than a conjectural, basis for discussing arrhythmogenic mechanisms. ardiac arrhythmias result from abnormalities of impulse conduction or impulse initiation.lx2 Although we tend to think of arrhythmias as occurring secondary to disease processes or autonomic abnormalities, a spectrum of the mechanisms responsible for arrhythmias can be induced by the same antiarrhythmic drugs usually used for treatment. Most reviews of the mechanisms responsible for arrhthymias stress that each major category of arrhythmogenic mechanism (e.g., abnormal impulse propagation and initiation] consists of multiple subcategories. Hence, conduction abnormalities subsume unidirectional and bidirectional block of propagation, reentry (both anatomic and functional) and reflection. Abnormal impulse initiation includes automaticity (impulse initiation occurring de novo) and triggered rhythms (induced by early or delayed afterdepolarizations). AbnormalImpulseConductionInduced by CardiacDisease Two forms of abnormal impulse conduction will be discussed: the occurrence of reentry, resulting from unidirectional conduction block and retrograde propagation, and simple conduction block, uncomplicated by reentry. The former may be expected to induce various patterns of extrasystoles, ranging from single premature depolarizations to tachycardias; the latter may be expected to induce single or multiple dropped beats. The model for reentry to which we shall refer is that described by Schmitt and Erlanger.3 The model requires an anatomic site of unidirectional conduction block through which anterograde activation cannot proceed (Fig. 1). Propagation continues normally through other limbs of the conducting system, there- From the Departments of Pharmacology and Pediatrics, and the Division of Developmental Pharmacology, Columbia University College of Physicians and Surgeons, New York, New York. Some of the studies referred to were supported by grants HL 28223 and HL 30557 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland. Address for reprints: Michael R. Rosen, MD, Department of Pharmacology, Columbia University College of Physicians and Surgeons, 630 West 168 Street, New York, New York 10032. 10E April 30, 1987 THE AMERICAN JOURNAL OF CARDIOLOGY Volume 59 IlE tion tends to increase with further loss of membrane potential, and when membrane potential attains approximately -60 mV, the fast inward Na current may fail entirely. As the fast inward Na current diminishes, there are decreases in the amplitude and upstroke velocity of the action potential, and a resultant depression of conduction. AbnormalImpulseConductionInduced by AntiarrhythmicDrugs FIGURE 1. A reentrant circuit in the ventricular conducting system. An impulse propagating in the anterograde direction, A, comes to a bifurcation in the Purkinje system and continues to conduct through limbs B and B’. The impulse in limb B’ arrives at a depolarized segment of myocardium (shaded area) in which it is blocked. The impulse in limb B continues to propagate and activates the myocardium, C. The latter impulse then enters the distal limb of the depressed segment, C, and propagates retrogradely through the depressed segment itself. Assuming the refractory period of the proximal fibers here has terminated, the impulse then can reenter the proximal system, D. Adapted from Schmitt and Erlanger.3 fore bypassing the site of unidirectional block, and the impulse enters the depressed segment retrogradely via its distal portion. Retrograde activation proceeds through the depressed section, and if sufficiently slow, the impulse arrives at the proximal interface with normal tissue after termination of the refractory period. The resultant impulse then can continue its circuit within the specialized conducting system, inducing a single reentrant beat, couplets or salvos of reentrant impulses. In the presence of cardiac disease, depolarization of the cell membrane is one cause of the aforementioned unidirectional block. Such depolarization may result from failure of the sodium-potassium (Na-K) pump [in normal fibers, Na-K pumping drives Na out of, and K into, the cell in an electrogenic fashion, such that there is a high, intracellular K concentration and a relatively negative transmembrane potential).4 This process depends on Na+-K+ adenosine triphosphatase for its maintenance. Pathologic events such as ischemia also depolarize the cell: this is caused by K+ loss secondary to increased K+ permeability of the ce11.5 The Na+-K+ exchange compensates only in part for this K+ loss. It is hypothesized that net K+ loss results from the effects of lactate and phosphate generated by the cell6 When there is a decrease in resting membrane potential, the fast inward Na current responsible for the phase 0 upstroke of the normal cardiac action potential is inactivated (Fig. 2).‘js7Such inactiva- Like disease, drugs can depress conduction and induce reentry. However, the mechanisms whereby antiarrhythmic drugs tend to exert these effects differ from those usually associated with disease states. As described previously, in the presence of ischemia, there is a metabolic failure in the heart, resulting in a decrease in resting membrane potential5 The decrease in amplitude and upstroke velocity of the action potential and the failure of propagation tend to be secondary to the pathologically induced decrease in membrane potential. When ischemia is relieved, membrane potential is restored and conduction improves. Also, disease processes can irreversibly damage the cell membrane, not only altering cell metabolism, but also the channels responsible for passage of ions across the membrane. .2 set C 700 il max ( Vjsec 1 0 LL 50 90 MP (-mV) FIGURE 2. Panel A shows the transmembrane action potential of a normal Purkinje fiber, having a resting membrane potential of -95 mV, and a large phase 0 overshoot. Panel B shows the same fiber depolarized to a resting potential of -70 mV by elevating [K+]o. Note the lower phase 0 overshoot. In both panels, the line of 0 potential and the phases of the action potential are indicated. Panel C is the plot of the maximum rate of rise of phase 0 (V,,,) as a function of the membrane potential of a fiber that was sequentially depolarized by adding Kt to the superfusate. Note that as membrane potential decreases, so does V,,,, reflecting a decrease in rapid Na+ entry. 12E A SYMPOSIUM: PERSPECTIVES ON PROARRHYTHMIA A: Channel States B: Local Anesthetic Access FIGURE 3. Panel A demonstrates the states of the fast Naf channel during dlfferent phases of the action potential. The resting state Is that seen during phase 4 of the transmembrane potential; the open state, phase 0; the lnactlvated state, phase 2 and part of phase 3. L = lipid bllayer of membrane; m. h = channel gates; P = channel proteln; SF = selectivity filter. (See text.) Panel B shows the access of local anesthetics to their blndlng site, shown here as a cross-hatched rectangle lnslde the selectlvlty fllter. D represents un-lonlzed molecules, which gain equivalent access to the blndlng site through the restlng, /eff, open, right, and inactivated (not shown) channels. & Is an lonlzed molecule that traverses the lipid bllayer poorly, but, on gaining access to the aqueous cytosol, can readily enter the channel when It Is open, rlghf, but not when It Is closed, left. In contrast, both the therapeutic and toxic depressant actions of antiarrhythmic drugs on conduction result from the binding of such drugs to specific sites in ionic channels that traverse the cell membrane. Let us use the fast Na+ channel as an example. This channel is the conduit for the rapid Na entry that determines phase 0 of the normal action potential of cells in the specialized conduction system and myocardium. Only cells in the sinoatrial and atrioventricular nodes and on the atria1surfaces of the mitral and tricuspid valves are not uniquely dependent on this channel. Results of studies performed primarily in nerve cells have enabled us to form an idea of the structure of the fast Na+ channel and have approximated the site within the channel to which most local anesthetic antiarrhythmic drugs bind (Fig. 3).‘-11The channel structure includes the following essential components? (1) the channel protein, which is the boundary for an aqueous pathway through the lipid bilayer of the cell membrane; and (2) a “selectivity filter,” which is near the outer surface of the channel and determines the ionic species that can pass through the channel and enter the cell. in the fast channel, the diameter is approximately 4A. The channel is highly specific for Na+, although other ions, such as H+, can also enter.12 We have thus far described the physical characteristics of a channel that will permit a specific ion to enter the interior of a cell. It must be stressedthat the channels are far more than passive, ion-specific portals in the membrane. Rather, they function in a complex fashion that results in their maintaining (1) a resting state, during which ions cannot enter, (2) an open state, during which ions pass freely from outside to inside and (3)an inactivated state,during which transit of ions does not occur.7-12The channel state (resting, open or inactivated) is determined by so-called gating proteins that are located on the interior of the channel (Fig. 3). For the fast sodium channel, 2 such proteins have been described and designated as m and h.13In the resting state, m is closed and h is open. When the cell is depolarized to its threshold potential, m moves to the open position and Na+ ions enter, depolarizing the cell. As phase 0 depolarization proceeds, the h gate closes, inactivating the channel. Subsequently, the m gate assumes the closed position and h reopens, returning the cell to the resting state. Whether the channel is in an “open” or “closed” position depends on the voltage acrossthe membrane; moreover, the processis time dependent. The binding site for local anesthetic-antiarrhythmic drugs is located within the channel, between the selectivity filter and the channel gates (Fig. 3B). Drugs cannot attain their binding sites via the outer-channel pore, becausethe diameter of the selectivity filter is too April 30.1997 HH’ FIGURE 4. The modulated receptor hypothesis for antiarrhythmic drug action. Channels are depicted as existing in the resting, R, open or activated, A, and inactivated, I, states. When combined with drug, channels are in modified resting, R’, activated, A’, and inactlvated, f, states. Different constants, K, are used to describe the binding and unbinding of drug to the channel in each state. The kinetics for changes among channel states are those described by Hodgkin and Huxley, HH, and are modified in the presence of drug, Hkf. Reproduced with permission from Biochem Biophys Acta. small to permit passage of the drug molecule. Rather, the drug must traverse the lipid bilayer of the membrane, a process that is facilitated if the molecule is lipophilic.ll As for ionized forms of the molecule, these cross the lipid membrane relatively slowly. Once they enter the cytosol, however, ionized molecules travel readily through this aqueous medium and gain access to their binding site at times when the gates are in the open configuration. Hence, local anesthetic antiarrhythmics have 2 means whereby they can reach their binding sites.7~8~11.14-17 (1) Th e access of lipophilic forms appears to depend on the degree of lipid solubility of the molecule and is not limited by cardiac rate. The type of rate-independent channel block induced by such lipid-soluble molecules is referred to as “tonic.” Channel block is dependent on membrane potential, however, which increases in magnitude as the membrane is depolarized. (2) In contrast to lipophilic, un-ionized drugs, ionized molecules have increasing accessto the channel as heart rate increases [i.e., as there are more gate openings per unit time). Many of these show voltage-dependent effects as well. The rate-dependent block produced by ionized drugs is also referred to as use dependent. As shown in Figure 4, specific constants have been identified that describe the rates at which individual drug molecules gain accessto and leave their channelbinding sites. These vary among antiarrhythmic local THE AMERICAN JOURNAL OF CARDIOLOGY 13E Volume 59 anesthetics; some, like lidocaine, bind rapidly, but they also unbind rapidly. Whatever block is produced as a result of their binding to the channel site in its open and inactivated states also dissipates rapidly during diastole. Such drugs have limited cardiac toxicity. Other drugs, like disopyramide, bind rapidly (although less rapidly than lidocaine) but have much slower’ rates of unbinding. l8 When heart rate is rapid, more drug will accumulate at the binding site, as less leaves during diastole, and the likelihood of drug-induced conduction block is much greater. Whether molecules unbind rapidly or slowly from a channel is determined by their molecular weight. Drugs having high molecular weights tend to dissociate from their binding sites more slowly than those of low molecular weight. As stated previously, the binding of a number of antiarrhythmic drugs is voltage dependent. This property is demonstrated in Figure 5, using lidocaine as an example. Note that at a high level of membrane potential, lidocaine reduces the maximum rate of depolarization (ir,,,] of phase 0 of the action potential (which is an indirect measure of the fast inward current and conduction velocity) to only a modest extent. However, as the membrane is depolarized, the blocking effect of \ -I Control .--. L ido, 6)glg/ml \ \ \ 0' 90 I I I 80 70 60 MP (-mV) FIGURE 5. Effects of itdocaine on the i,,,,. of canlne Purkinje fibers. The vertical axis is i maxexpressed as percent of control (control = 100%). The horizontal axis is membrane potential of a series of Purkinje fibers that were progressively depolarized by elevating K+. Note that in the presence of lidocaine 6 pg/ml, the curve is displaced downward and to the left, with greater depression occurring at the less negatlve levels of membrane potential. This greater depression of i,,,,, at more positive potentials demonstrates the voltage dependency of lidocaine’s actions. 14E A SYMPOSIUM: PERSPECTIVES ON PROARRHYTHMIA 100 I Wsec L idocoine Smg/L I 50 mV J 200 msec FIGURE 6. Effects of fldocaine on the action potential of a canine subendocardiaf PurkinJe fiber in a region of myocardiaf infarction. Left, the infarcted anterior septum, from which the action potentials on the rrghf were recorded. Right, top trace is 0 potential and the action potential, and bottom is the V,,, . Panel A, control; Panel Bshows the effect of fidocaine 5 mg/fiter. Thls concentration of drug decreases the amplitude of the action potential and the V,,,. Reproduced with permission from Am Heart J 1975;99:526-536. lidocaine on V maxbecomes more marked. Extrapolating this information to the intact heart in which a cardiac arrhythmia is occurring, we first envision that lidocaine will tend to have rather little effect on conduction in normal fibers, since these have high membrane potentials. This has been verified both in isolated tissue studies, where lidocaine has been shown to have little effect on normal Vmax14J5 and conduction,lg and in the in situ heart, where lidocaine has little effect on conduction or QRS morphology and duration. However, lidocaine has a highly specific effect on the abnormal action potentials that occur in depressed tissues (Fig. 6)14; this may induce reentry of the type shown in Figure 1. Such slowly propagating action potentials have low Vmax,and in the presence of lidoCaine, V,,, is further diminished, with the possibility that conduction may fail altogether. This would convert unidirectional to bidirectional conduction block and terminate the arrhythmia. Although these voltage-dependent effects of drugs are generally thought of as antiarrhythmic [Fig. 7A,7B), it is easy to see how these effects might generate arrhythmias as well. For example, let us consider a partially depolarized site that is still showing intact anterograde conduction in the absence of drug treatment [Fig. 7C). Such a site would not be likely to generate an arrhythmia. However, the same concentration of a drug that may induce bidirectional block and cessation of arrhythmias at a more markedly depressed site may induce a sufficient diminution of V,,, to slow conduction in a partially depolarized area, resulting in unidi- rectional block and possible reentry (Fig. 7D). Moreover, such a phenomenon may show rate dependence; i.e., at slow heart rates, insufficient quantities of the drug molecule may gain access to the binding site to induce unidirectional block. At fast heart rates, however, with more gate openings per unit time, there may be further drug accumulation at the receptor site, such that unidirectional block and reentry occur. Thus far we have explained the means whereby local anesthetic drugs attain their binding sites, and we have explored the basis for their contribution to conduction block and to reentry. Other examples of druginduced toxicity include conduction block at the sinoatria1 and atrioventricular nodal level. In the regions of the sinus and atrioventricular nodes, conduction is mediated by so-called “slow response” action potentials that are largely Ca++ dependent. Conduction block here can be induced by drugs that bind to sites in the “slow” Ca++ channel, such as diltiazem and verapamil. These slow-channel blocking drugs can also show use dependence. As is the case with drugs acting on fast channels, the important factor in reducing the rate of propagation of the cardiac impulse and either blocking conduction or inducing reentry is access of drug to its channel binding site. Finally, it is not the effect of drugs on the upstroke of the action potential alone that may depress conduction. Drugs that prolong repolarization, whether in addition to depressing phase 0 of the action potential [e.g., quinidine) or in the absence of effects on phase o (e.g., sotalol), can slow conduction. This will most April 30, 1987 A Control B Drug 0 Drug THE AMERICAN JOURNAL OF CARDIOLOGY Volume 59 15E 0 fc 4 C Control A 0 FIGURE 7. The modffication of conduction by an antiarrhythmic drug. Panel A shows a reentrant circuit under the same conditions depicted in Figure 1. A markedly depressed action potentlal is depicted as occurring in the depressed segment, and there are anterograde conductlon block and reentry. In panel /3, an antiarrhythmic drug has rendered the site inexcitable; there Is bidirectional conduction block, and arrhythmia has ceased. In panel C, there is slight depression of propagation at a slte in the conducting system, but anterograde activation persists. In panel D, this she has been depressed further by drug; there is now anterograde conduction block and reentry. FIGURE 8. Modification of transmembrane potential characteristics by prolonging repolariration. Panel A, control. The numbers 1,2 and 3 designate basic cycle length; the letters A and 8 indicate premature action potentials. In the control, premature action potential A arrives early during repolarization and can be assumed to propagate slowly and abnormally. In contrast, premature actlon potential B arrives after full repolarization and would be expected to conduct normally. In panel f3, an antiarrhythmic drug has markedly prolonged repolarlzation. Premature A now finds the membrane refractory and cannot propagate. Premature ~9excites a markedly depolarized membrane, but one that is still capable of generating an action potential. The result is slow conduction and the occurrence of an arrhythmia different from that elicited by premature B in the upper panel. readily occur when a propagating impulse arrives before repolarization is complete. This can be most readily understood if we consider the following example. Usually, as repolarization is prolonged and the duration of refractoriness increases, we would expect increasing block of the propagation of premature impulses. However, as shown in Figure 8, a premature impulse that originally occurred during diastole might now occur during phase 3 repolarization after prolongation of the action potential by an antiarrhythmic drug. In this event, the impulse, which was premature but propagated normally during the control situation, would propagate slowly and abnormally after drug administration and might thereby predispose the heart to a more complex arrhythmia than was originally seen. Abnormal impulse initiation: It has been known for some time that a variety of antiarrhythmic drugs can induce ectopic impulse initiation and tachycardias. These are primarily drugs that prolong repolarization, such as quinidine .20It has been suggested that arrhythmias induced by quinidine result from the initiation of automaticity .20One problem in considering such automaticity is that the concentrations of quinidine needed to depolarize the membrane markedly and induce automatic activity are far higher than one expects to attain in the clinical setting, even in the presence of drug overdose. We have recently learned that quinidine concentrations well within the clinically relevant range not only prolong repolarization, but also induce oscillations during phases 2 and 3.21Such so-called early afterdepolarizations can then initiate Moreover, the prolongation triggered arrhythmias. 22,23 of repolarization appears crucial to the development of afterdepolarizations. This prolongation of repolarization may result from either an increased inward current during repolarization or a decreased repolarizing current, or both. To be specific, during the action potential plateau, or phase 2 (Fig. 2) net membrane current is outward throughout the range of membrane potentials between 0 and the resting potential. The net repolarizing current during phase 2 is the result of an imbalance between inward and outward membrane currents.24 Inward components include a background Na current,25 Na flowing through incompletely inactivated Na channels,26 and the slow inward (Ca++) current.26*27 One of the outward currents is probably caused by K flowing through a time- and voltage-dependent channel.2s There may also be a small, time-independent K current2g and a current generated by the electrogenic Na+-K+ pump. 30,31Under normal circumstances the net outward membrane current moves membrane potential steadily in a negative direction, resulting in action potential repolarization seen as phase 3. An early afterdepolarization can occur if there is a shift in the current-voltage relation such that f-i I-N-L t I WE A SYMPOSIUM: A CL=2sec PERSPECTIVES ON PROARRHYTHMIA repolarizing K currents. This blocking effect is attributed to a decrease in fully activated K conductance.35 Quinidine prolongs Purkinj e fiber action potential duration, especially when the stimulation rate is slow and K+ is lower than norma121 Prolongation of the action potential may result from quinidine’s blocking effect C CL=6sec on a repolarizing current carried by K+,37and is unrelated to quinidine’s block of the Na channel. In fact, V,,, is not affected by the low concentrations of quinidine that can cause early afterdepolarizations. This failure to block Na channels at low quinidine concentrations may favor the development of early afterdepoD CL=lOsec larizations, because an intact inward Na+ current may be important. One form of drug-induced arrhythmia that may re20 mV sult from early afterdepolarizations and triggered acSTset tivity is torsades de pointes. Abnormalities of ventricuFIGURE 9. Early afterdepolarizations and triggered activity. In panel lar repolarization are often associated with torsades de A the preparation Is driven at a cycle length of 2 seconds. The pointes. Although quinidine can induce torsades de superfusate Is Tyrode’s solution, contalnlng ceslum. In panel IY, at a pointes, the arrhythmia may occur at low plasma drug slower drive rate, there Is a “shoulder” occurring during phase 3 concentrations that do not cause prolongation of the repolarlzatlon (the early afterdepolarlzatlon), as well as a triggered QRS complex.38 Nonetheless, hypokalemia and braaction potential after the next-to-last driven action potentlal. In dycardia, both of which prolong repolarization, may panel C, at still a longer cycle length, there Is now trlggered actlvlty cause a predisposition to the occurrence of quinidineafter each drlven action potentlal, resulting In a blgemlnal rhythm. In induced torsades de pointes.3gTorsades de pointes has panel D, at a longer drlve cycle length, there are now salvos of also been associated with other drugs that prolong the triggered actlvlty. Reproduced wlth permlsslon from Clrculatlon.33 QT interval (procainamide, disopyramide, amiodarone, N-acetylprocainamide40-42 and sotalo143).It must be emphasized that while information linking druginduced torsades de pointes to prolonged repolarizanet inward current occurs during phase 2.32 This would delay or prevent repolarization and might lead tion is well documented, the links among drug toxicity, to an oscillation during the plateau, or phase 3, if a prolonged repolarization, early afterdepolarizations and torsades de pointes are far more tenuous. In isolatregenerative inward current is activated. In contrast to automatic@, which can occur de ed tissue studies, it has been possible to induce pronovo, early afterdepolarizations must depend on ac- longed repolarization and early afterdepolarizations tion potentials for their initiation. Moreover, they are using a combination of quinidine and low superfusate K concentrationszl We are unaware, however, of any markedly influenced by the rate of impulse initiation of the action potentials that trigger them.33 At a range’ success in inducing the same arrhythmia in intact dogs of cycle lengths that encompasses the normal sinus having low plasma K concentrations and therapeutic or toxic quinidine levels. In animal models in which rhythm of the adult human heart, early afterdepolarizations are rarely seen. As cycle length is increased quinidine has induced an arrhythmia resembling torsades de pointes, an additional intervention, myocarand repolarization is prolonged, early afterdepolarizations are generated and their amplitude increases (Fig, dial ischemia, was required.a4 In this study, conduction 9). When their amplitude is sufficiently great, they abnormalities were described, but evidence for early afterdepolarizations was neither sought nor acquired. can attain threshold potential and induce tachyIn other animal models in which torsades de pointes’ arrhythmias. The means whereby drugs might induce early af- was induced and evidence for early afterdepolarizaterdepolarizations and triggered activity include (1) re- tions was acquired using monophasic action-potential tarding repolarization, in which case membrane po- recordings, the intervention that produced the artential remains at the low level at which triggered rhythmia was intravenous cesium rather than an antiactivity induced by early afterdepolarizations can oc- arrhythmic drug. 45Finally, some monophasic actioncur, whether as a result of decreasing the outward, potential recordings have been made in patients repolarizing K current or increasing the steady, back- having torsades de pointesj6 They showed oscillations interpreted by the investigators as early afterdepolariground Na+ current; (21increasing the inward current zations, although argument continues over the possiresponsible for the depolarization phase of triggered bility that these might be artifacts. action potentials; and (3) increasing phase 4 depolarIn summary, we have emphasized in this brief reization of triggered action potentials, either by increasview that arrhythmias induced by antiarrhythmic drug ing inward current or by decreasing repolarizing toxicity are not unlike those induced by disease. Nonecurrent. theless, the underlying mechanisms differ, because Among the antiarrhythmic drugs that prolong action potential duration of Purkinje fibers are sota- they are influenced by the specific binding character-, 101,34J5 N-acetylprocainamides6 and quinidinea21 Sota- istics of drug to its receptor site. In addition, in the presence of a diseased myocardium, administration of 101prolongs the action potential duration by inhibiting B CL =4 set April 30. 1987 an antiarrhythmic drug may result in an arrhythmia, whereas either alone might not be arrhythmogenic. Hence, in considering the possible toxicity of antiarrhythmic drugs, we must take into account both the mechanism of action of the individual drug and myocardial state. Finally, although much circumstantial evidence links drug toxicity, prolonged repolarization and triggered arrhythmias, much work needs to be done before this relation is proven. Acknowledgment: The authors express their gratitude to Linda Morris for her careful attention to the preparation of this manuscript. References 1. Cranefield PF, Wit AL, Hoffman BF. Genesis of cardiac arrhythmias. Circulation 1973;47:190-204. 2. Hoffman BF, Cranefield PF. The physiological basis of cardiac arrhythmias. Am J Med 1964;37:670-684. 3. Schmitt FO, Erlanger J. Directional differences in the conduction of the impulse through heart muscle and their possible relation to extrasystolic and fibrillary contractions. Am J Physiol 1929;87:326-347. 4. Hoffman BF. Cranefield PF. Electrophysiology of the Heort. New York: McGraw-Hill, 1960. 5. Kleber AG. Resting membrane potential, extracellular potassium activity and intracellular sodium activity during acute global ischemia in isolated perfused guinea pig hearts. Circ Res 1983;52:442-450. 6. Jack JJB. Noble D. Tsien RW. Electric Current Flow in Excitable Cells. Oxford: Clarendon Press, 1975. 7. Hondeghem L. Katzung B. Time and voltage dependent interactions of antiarrhythmic drugs with cardiac sodium channels. Biochem Biophys Acta 1977;472:373-398, 8. Strichartz GR. The inhibition of sodium currents in myelinoted nerve by quaternary derivatives of lidocaine. J Gen Physiol 1973;62:37-57. 9. Hille B. Ionic channels in nerve membranes. Prog Biophys Mel Biol 1970:21:1-32. 10. Hille B. The pH-dependent rate of action of local anesthetics on the node of Ranvier. J Cen Physiol 1977;69:475-496. 11. Hille B. Local anesthetics: hydrophilic and hydrophobic pathways for the drug-receptor reaction. J Gen Physiol 1977;69:497-515. 12. Hille B. Ionic Channels of Excitable Membranes. Sunderland, Massachusetts: Sinauer Associates, 1984. 13. Hodgkin AL, Huxley AF. A quantitative description of membrane current and its application to conduction excitation in nerve. J Physiol 1952;117: 500-544. 14. Chen CM, Gettes LS. Combined effects of rate, membrane potential and drugs on maximum rate of rise (6,,,] of action potentials upstroke of guinea pig papillary muscle. Circ Res 1976;38:464-469. 15. Chen CM, Gettes LS, Katzung RG. Effects of lidocaine and quinidine on steady-state characteristics and recovery kinetics of [dv/dt)max in guinea pig ventricular myocardium. Circ Res 1975;37:20-29. 16. Courtney KR. Mechanism of frequency-dependent inhibition of sodium currents in frog myelinated nerve by the lidocaine derivative CEA-968. r Pharmacol Exp Ther 1975;195:225-236. 17. Courtney KR. Interval-dependent effects of small ontiarrhythmic drugs on excitability of guinea-pig myocardium. [ Mol Cell Cardiol 1980;12: 1273-1286. 18. Vaughan Williams EM. Subdivision of class I drugs. In: Reiser H. Horowitz L. eds. Mechanisms and Treatment of Cardiac Arrhythmias: Relevance of Basic Studies to Clinical Management. Baltimore: Urban B Schwarzenbert, 1985:165-172. 19. Rosen M, Merker C, Pippenger C. The effects of lidocoine on the canine THE AMERICAN JOURNAL OF CARDIOLOGY Volume 59 17E ECG and electrophysiologic properties of Purkinje fibers. Am Heart r 1976:91:191-202. 20. Bellet S. Clinical Disorders of the Heart Beat. Philadelphia: Lea 6 Febiger, 1971. 21. Roden DM, Hoffman BF. Action potential prolongation and induction of abnormal automaticity by low quinidine concentrations in canine Purkinje fibers. Relationship to potassium ond cycle length. Circ Res 1985;56:857-867. 22. Cranefield PF. The Conduction of the Cardiac Impulse: The Slow Response and Cardiac Arrhythmias. Mt. Kisco: Futura Press, 1975. 23. Cranefield PF. Action potentials, afterpotentials and arrhythmias. Circ Res 1977:41:415-423. 24. Vassalle M. Electrogenesis of the plateau and pacemaker potential. Ann Rev Physiol 1979;41:425-440. 25. Gadsby DC, Cranefield PF. Two levels of resting potential in cardiac Purkinje fibers. J Gen Physiol 1977;70:725-746. 26. Coraboeuf E. Voltage clamp studies of the slow inward current. In: Zipes DF, Bailey IC, Elharrar V. eds. The Slow Inward Current and Cardiac Arrhythmias. The Hague: Martinus Nijhoff, 1980:25-96. 27. Carmeliet E. The slow inward current: non-voltage-clamp studies. In: Ref 2f3:97-110. 28. McAllister RE, Noble D. The time and voltage dependence of the slow outward current in cardiac Purkinje fibers. J Physiol (Land) 1966;186:632-662. 29. Dude1 J, Peper K, Rude1 R, Trautwein W. The potassium component of membrane current in Purkinje fibers. Pfluegers Arch 1967;296:308-327. 30. Isenberg G, Trautwein W. The effects of dihydro-ouabain and lithium ions on the outward current in cardiac Purkinje fibers. Pfluegers Arch 1974: 350:41-54. 31. Gadsby DC, Cranefield PF. Electrogenic sodium extrusion in cardiac Purkinje fibers. [ Gen Physiol 1977;70:725-746. 32. Trautwein W. Mechanisms of tachyarrhythmias and extrasystoles. In: Sandoe E, Flenstad-Jensen E. Oleson K, eds. Symposium on Cardiac Arrhythmias. Sodertalje. Sweden: AB Astra. 1970:53-66. 33. Damiano BP, Rosen M. Effects of pacing on triggered activity induced by early afterdepolarizations. Circulation 1984;69:1013-1025. 34. Strauss HC, Bigger JR, Hoffman BF. Electrophysiological and beta-blocking effects of MJ 1999 on dog and rabbit cardiac tissue. Circ Res 1970; 26:661-678. 35. Carmeliet E. Electrophysiologic and voltage clamp analyses of the effects of sotalol on isolated cardiac muscle and Purkinje fibers. \ Pharmacol Exp Ther 1985;232:817-825. 36. Dangman KH, Hoffman BF. In viva, in vitro. antiarrhythmic and (ITrhythmogenic effects of N-acetyl procainamide. r Pharmacol Exp Ther 1981; 217:851-862. 37. Colatsky TJ. Mechanism of action of lidocaine and quinidine on action potential duration in rabbit cardiac Purkinje fibers: an effect on steady-state sodium currents? Circ Res 1982;50:17-27. 38. Seizer A, Wray HW. Quinidine syncope. Paroxysmal ventricular fibrillation occurring during treatment of chronic atria1 arrhythmias. Circulation 1964;30:17-26. 39. Smith WM, Gallagher JJ.“Les torsades de pointes”: an unusual ventricular arrhythmia. Ann Intern Med 1980;93:578-584. 40. Olshansky B, Martins J, Hunt S. N-acetylprocainamide causing torsades de pointes. Am J Cardiol 1982;50:1439-1445. 41. Stratmann HG, Walter KE. Kennedy HL. Torsades de pointes associated with elevated N-acetylprocainamide. Am Heart [ 1985;109:375-377. 42. Chow MJ, Piergies AA, Bowsker DJ, Murphy JJ,Kushner W. Ruo T, Asada A, Talano J, Atkinson AJ. Torsades de pointes induced by N-acetylprocainamide. [ACC 1984;4:621-624. 43. Kuck KH, Kunze KP. Roewer N, Bleifeld W. Sotalol-induced torsades de pointes. Am Heart J 1984;107:179-180. 44. Bardy G, Ungerleider R, Smith W. Ideker R. A mechanism of torsades de pointes in a canine model. Circulation 1983;67:52-59. 45. Brachmann J, Scherlag BJ, Rosenstraukh LV, Lazzara R. Bradycardiadependent triggered activity: relevance to drug-induced multiform ventricular tachycardia. Circulation 1983;68:846-856. 46. Bonatti V, Rolli A, Botti G. Recordings of monophasic action potentials of the right ventricle in the long QT syndrome complicated by severe ventricular arrhythmias. Eur Heart r 1983;4:168-172. Discussion Dr. Woosley (Nashville, Tennessee]: There are probably multiple mechanisms by which repolarization is prolonged. Could it be possible that drugs such as quinidine have one type of action to prolong repolarization and drugs like bepridil or lidoflazine prolong repolarization by a totally different mechanism? Dr. Rosen [New York, New York): Yes, as drugs may modify repolarization by decreasing outward current and/or by increasing inward current. The ionic channels that are involved carry Na, Ca and K. Incidentally, we were particularly intrigued by the study by Hanley in which a group of patients were treated with propranolol and lidoflazine. They 18E A SYMPOSIUM: PERSPECTIVES ON PROARRHYTHMIA had very slow sinus rates, and a very high percentage of these patients developed torsades de pointes. We tried lidoflazine in a canine model. As we expected, they responded with long QTs and no arrhythmia. We then gave the dogs a combination of propranolol and lidoflazine and saw similar QT prolongation, slower sinus rates and no torsades. We assumed that the rate was not slow enough and then induced complete heart block in some animals. Those with idioventricular rates of 40 beats/min had long QTs, and we finally induced torsades in 1 animal. We have now been working more with quinidine than lidoflazine and have found that an additional event is needed. That is, we can induce torsades by prolonging repolarization with quinidine [or lidoflazine) and then adding aconitine, applied locally to 2 sites on the heart. This produces a classic twisting of the pointes. If you put aconitine on just 1 site, the tachycardia is monomorphic. Even with application to 2 sites, torsades usually will not occur if the QT is not prolonged. Dr. Zipes (Indianapolis, Indiana): That may be because cycle lengths are faster when the refractory periods are shorter. Dr. Rosen: We have covered a range of heart rates from 60 to about 200 beats/min. It would be helpful to know how often this group sees torsades occurring in patients with supposedly healthy hearts except for the presence of a life-threatening arrhythmia. Can you fill us in? Dr. Horowitz (Philadelphia, Pennsylvania): Almost never. However, our experience is primarily with patients with heart disease, because this is who gets treated. If 1,000 healthy people took a drug that produced torsades, perhaps it would occur in an equal number of them as well. The only finding that suggeststhat this not the case is that the worse the organic heart disease is, the more likely the patient is to develop a proarrhythmic effect. By extrapola- tion, then, if your heart is normal and you have no arrhythmia that requires an antiarrhythmic agent, the odds of a proarrhythmic effect would seem very low if you were to receive the drug. Dr. Woosley: We see torsades frequently in patients with atria1 fibrillation. These patients usually have normal ventricles without large scars. Dr. Horowitz: But they have heart disease as a reason for atria1 fibrillation. Dr. Podrid (Boston, Massachusetts): Not necessarily. We have had patients who are “lone fibrillators” in that they have no heart disease. Some have developed torsades while receiving quinidine, but it is a rare event. Dr. Zipes: A patient can have long QTs without arrhythmia or normal QTs with malignant ventricular arrhythmia. To my knowledge, there is no relationship between a critical degree of QT prolongation and the development of an arrhythmia. Dr. Rosen: I was trying to relate QT specifically to torsades because our animal studies suggest that the long QT facilitates not the overall incidence of arrhythmias, but the torsades pattern in particular. Dr. Podrid: You mentioned altering 2 separate areas on the myocardium, and that may be important. Our perception is that the incidence of arrhythmia aggravation, whether it be torsades or other forms, is much higher in patients with organic heart disease. In patients with heart disease we are dealing with a myocardium that not only has an appropriate substrate but which also has areas that are at least partially damaged. Consequently, you have not only QT prolongation but disparate electrophysiologic properties within different parts of the myocardium that act in concert. The addition of another factor, such as isoproterenol or hypokalemia, which may enhance this disparity, can further increase the potential for arrhythmia when the appropriate preconditions are present.