Classification of the Long QT Interval Syndromes and Torsade de Pointes

As seen in Figure 1, the QT interval recorded on the surface electrocardiogram corresponds to the relatively isoelectric plateau phase of the action potential. The broad T wave is inscribed as a result of the rapid repolarization occurring nonsimultaneously throughout the ventricles. The QT interval is prolonged by any agent or condition that delays repolarization in the ventricular cells. Various agents and conditions cause these long QT interval syndromes Table 1 and initiate the characteristic torsade de pointes tachyarrhythmia (Figure 2).

View larger version (19K): [in this window] [in a new window]   Figure 1. Relation between phases of the cardiac action potential and the surface electrocardiogram. The QT interval roughly corresponds to the plateau phase of the action potential. The broad T wave is inscribed as a result of the rapid repolarization occurring nonsimultaneously throughout the ventricles. The QT interval is prolonged by agents or conditions that delay repolarization in the ventricular cells. ECG = electrocardiographic.

View larger version (52K): [in this window] [in a new window]   Figure 2. Electrocardiogram showing the typical short-long-short RR interval initiation sequence triggering an episode of torsade de pointes. A sinus beat (A) is followed by a premature ventricular depolarization (B) after a short coupling interval, which is followed, after a long pause after extrasystole, by a sinus beat (C) and another short-coupled premature ventricular depolarization (D), which is the first beat of the tachycardia and which is characterized by an apparent twisting of the electrical axis around the isoelectric line.

The long QT interval syndromes have been classified as acquired or hereditary on the basis of the conditions that appear to trigger the polymorphic ventricular tachycardia associated with the syndromes. The acquired long QT interval syndromes are also typically classified as pause-dependent because the torsade de pointes associated with them generally occurs at slow heart rates or in response to short-long-short RR interval sequences. The hereditary long QT interval syndromes are typically considered adrenergic-dependent because the torsade de pointes associated with them is generally triggered by adrenergic activation or enhancement of sympathetic nervous system tone [5]. However, the acquired and hereditary syndromes overlap somewhat; adrenergic-dependent forms are more likely to occur in response to pauses or short-long-short RR interval sequences or during slow heart rates. As seen in Table 1, causes of the acquired long QT interval syndromes include antiarrhythmic drugs, severe bradycardia, electrolyte disturbances (notably hypokalemia), various nonantiarrhythmic drugs, and numerous unrelated clinical disorders. The hereditary long QT interval syndromes include the Jervell and Lange-Nielsen syndrome [8], which is associated with congenital deafness, and the Romano-Ward syndrome [9, 10], in which hearing is normal.

As mentioned above, initiation of the acquired, pause-dependent form of torsade de pointes often (although not invariably [11]) involves a short-long-short RR interval sequence [5]. As seen in Figure 2, a premature ventricular depolarization closely coupled with the last normal QRS complex is followed by a long pause after extrasystole that precedes another closely coupled premature ventricular depolarization, which constitutes the first beat of the tachycardia [5, 12, 13]. The tachycardia has a rate of 150 to 250 beats/min and is usually self-terminating. It is usually nonsustained and produces either no symptoms or only mild symptoms of presyncope, but it can occasionally degenerate into ventricular fibrillation that may be fatal. Although the initiation of torsade de pointes in patients with the hereditary long QT interval syndromes does not require pauses or bradycardia, adrenergic activation or enhanced sympathetic nervous system tone does appear to be necessary to produce the bizarre QT or QTU prolongation characteristic of this form of the disorder.

Cellular Mechanisms of the Long QT Interval Syndromes and Torsade de Pointes

Normal Cardiac Action Potential

The action potential of the ventricular or Purkinje fiber Figure 1 results from the transmembrane movement of ions through channels in the cell membrane. The normal resting potential of most cardiac cells is about –80 to –90 mV. Membrane depolarization results from the net influx of positive charges (inward current) and repolarization occurs secondary to the net efflux of positive charges (outward current). The major ion currents responsible for the action potential in ventricular and Purkinje fibers are listed in Table 2. During diastole (phase 4), the membrane remains polarized near the K⁺ equilibrium potential ( –90 mV) because the inward rectifier K⁺ current (I_K1) is normally the only current of significant magnitude activated at this membrane potential. In Purkinje fibers, diastolic depolarization results from a combination of the decay of the outward delayed rectifier current (I_K) and the activation of the inward pacemaker current (I_f) and the inward Na⁺ background leak current (I_Na-B). The rapid upstroke of the action potential (phase 0) results from the inward fast Na⁺ and L- and T-type Ca²⁺ currents (I_Na, I_Ca-L, and I_Ca-T). Initial repolarization immediately after the overshoot (phase 1) is mediated by activation of the transient outward current (I_TO) carried by K⁺ ions. The plateau (phase 2) results from a balance between inward currents, primarily the slowly decaying I_Ca-L and possibly the electrogenic Na⁺-Ca²⁺ exchange current (I_NaCa) [14] and the outward currents I_K and I_TO.Repolarization (phase 3) develops as the outward currents, especially I (_K), overwhelm the decaying inward currents. An inward "window" current is responsible for the repetitive depolarizations at plateau levels of potential that are characteristic of early afterdepolarizations (see below). This current appears to be generated by the slow inactivation of I_Ca-L [15] and I_Na [16].

Role of Early Afterdepolarizations in the Genesis of Prolongation of the QT Interval and Torsade de Pointes

The study of torsade de pointes is confounded by the transient nature of the condition and the unpredictability of its occurrence and because it generally cannot be induced by programmed electrical stimulation during electrophysiologic study [17, 18]. Thus, animal studies form the basis for the current understanding of this arrhythmia. It is believed that prolongation of the QT interval and torsade de pointes are caused by early afterdepolarizations, defined as single or repetitive depolarizations or oscillations of the transmembrane voltage, that occur at low levels of membrane potential because of a failure of normal, complete membrane repolarization (Figures 3 and 4). Early afterdepolarizations may occur during the plateau phase (phase 2 early afterdepolarizations) or the early rapid repolarization phase (phase 3 early afterdepolarizations) of the action potential (Figures 3 and 4) [19, 20]. Because these afterdepolarizations delay repolarization, they may result in significant prolongation of the QT interval.

View larger version (18K): [in this window] [in a new window]   Figure 3. Schematic drawing illustrating transmembrane action potentials from a Purkinje fiber, a ventricular muscle fiber, and a surface electrocardiogram. It shows the development of early afterdepolarizations and their effect on the surface electrocardiogram, including the QT interval. A. Baseline recordings. B. Phase 2 early afterdepolarization. The action potential is more prolonged in the Purkinje fiber than in the muscle fiber and shows a flattening of the plateau phase. Because of failure of normal repolarization (dotted line), the Purkinje fiber again depolarizes from its plateau phase potential. The depolarizing potential reaches threshold and is transmitted to the muscle fiber. This beat is seen on the surface electrocardiogram as a premature ventricular depolarization arising from the end of a prolonged QT interval (or in the midst of a pronounced U wave) (dotted line). The premature ventricular depolarization coincides with the prolonged action potential duration resulting from the early afterdepolarization. C. Phase 3 early afterdepolarization. The sequence is the same except that the early afterdepolarization arises during phase 3 of the Purkinje fiber action potential. (Reproduced with permission from [32]). PF = Purkinje fiber; MF = ventricular muscle fiber; ECG = surface electrocardiogram.

View larger version (12K): [in this window] [in a new window]   Figure 4. Prolongation of the QT interval resulting from quinidine-induced early afterdepolarizations arising from the plateau phase of the action potential. Transmembrane action potentials recorded from single isolated epicardial, endocardial, and M cells. Action potential recordings from cells electrically stimulated at various cycle lengths (A = 3.5 s; B = 5.0 s; and C = 20 s). Notice the dramatic prolongation of the action potential duration that occurs (note different time base in panel C with the development of repetitive early afterdepolarizations during the plateau phase of the action potential. [Reproduced with permission from Sicouri and Antzelevitch. J Cardiovasc Electrophys. 1993; 4:48-58.] Epi = epicardial; Endo = endocardial.

Although direct evidence that early afterdepolarizations cause torsade de pointes is lacking, considerable indirect evidence supports this hypothesis. Specifically, conditions that induce early afterdepolarizations initiate torsade de pointes [21, 22], especially at slow heart rates [23], and conditions that suppress early afterdepolarizations prevent torsade de pointes [24]. Furthermore, endocardial monophasic action potential signals (which reflect transmembrane electrical activity generated by action potentials) recorded under clinical and experimental conditions have shown that torsade de pointes is associated with voltage deflections resembling early afterdepolarizations [25, 26]. Early afterdepolarizations probably arise from the Purkinje fibers rather than from working myocardial cells because interventions that induce early afterdepolarizations do so far more easily in Purkinje fibers than in ventricular myocardial cells [27-29]. In addition, delay of cardiac repolarization with prolongation of the action potential duration—which appears to be necessary for the development of some types of early afterdepolarizations—is more prominent in Purkinje fibers than in ventricular tissue after exposure to antiarrhythmic drugs that delay repolarization [28-30].

It is important to emphasize that early afterdepolarizations are depolarizing potentials that develop because of a failure of normal repolarization. Because early afterdepolarizations do not develop in the absence of a preceding action potential and are, as it were, triggered by it, they are called triggered activity. Any imbalance between net inward and outward currents leading to failure of membrane repolarization may suffice to initiate phase 2 or phase 3 early afterdepolarizations [27, 31-34]. In general, repolarization failure and prolongation of the action potential duration and the QT interval may be achieved by reducing net outward current (I_K [21, 24, 27-29, 35-55] or I_TO [24]), enhancing net inward current (I_Na [20, 56-59] or I_Ca-L [15, 40, 60, 61]), or both. In all cases, the generation of early afterdepolarizations appears dependent on, at least, a reduction in net outward current that, in the presence of residual inward currents (such as the "window" currents [32-34]), delays or prevents normal repolarization. As a result, the membrane potential can oscillate at the plateau phase level (Figure 4). If these membrane oscillations, or early afterdepolarizations, conduct from their origin to the ventricular myocardium at large, tachyarrhythmia will result [20].

The Acquired Long QT Interval Syndromes and Torsade de Pointes

Torsade de Pointes Caused by Antiarrhythmic Drugs

Antiarrhythmic agents are the best studied and most common causes of the acquired long QT interval syndromes and torsade de pointes (Figure 5). According to the Vaughan-Williams classification of antiarrhythmic drugs Table 3, only class III drugs block K⁺ channels, prolong the action potential duration, and prolong the QT interval. However, in reality, many class IA drugs (primarily Na⁺ channel blockers) also show significant K (⁺) channel blocking behavior. It is best, therefore, to speak in terms of "class III properties" that are found in both class III drugs and many class IA drugs. Consequently, both quinidine (class IA) and sotalol (class III) can induce QT prolongation and torsade de pointes because both show significant K⁺ channel blocking behavior. Drugs with class III properties appear to block components of I_K and I_TO, but I_K1 appears to be affected only slightly. Inhibition of these transmembrane K (⁺) currents may delay repolarization, initiate early afterdepolarizations, and precipitate torsade de pointes.

View larger version (9K): [in this window] [in a new window]   Figure 5. The acquired long QT interval syndrome resulting from treatment with a class IA agent. This 12-lead electrocardiogram was obtained from an 80-year-old woman who was treated with quinidine for atrial tachycardia. A few hours after this electrocardiogram was obtained, the patient had a torsade de pointes-related cardiac arrest. The maximum uncorrected QT interval lasted approximately 600 ms. The patient was successfully resuscitated but the tachyarrhythmia was suppressed only after 10 g of magnesium sulfate was administered intravenously.

Patients who develop torsade de pointes when receiving one class IA drug often develop it when receiving other class IA drugs [6]. It is clear that torsade de pointes occurs with normal or low doses of class IA drugs [4, 46, 62] but generally with only high doses of class III drugs (especially during slow heart rates). Furthermore, clinical evidence indicates that torsade de pointes rarely occurs during treatment with amiodarone. An understanding of the genesis of early afterdepolarizations can help to explain these differences. At high doses of class IA drugs, blockade of I_Na predominates over blockade of I_K [7, 18, 41]. This inhibits the depolarizing current required to initiate early afterdepolarizations in excess of the drug-induced reduction in outward repolarizing K⁺ currents and is in accordance with studies showing inhibition of early afterdepolarizations by Na⁺ channel blocking agents. Because class III agents (except amiodarone) lack significant Na⁺ channel blocking properties, higher doses of these agents increase the likelihood of torsade de pointes. A dose-dependent incidence of torsade de pointes has indeed been reported with sotalol [63]. The rarity of torsade de pointes during treatment with amiodarone [6, 7] may be explained by the drug's additional Ca²⁺ channel and Na⁺ channel blocking effects [7, 64] and its ß-adrenergic receptor blocking properties [8, 64]. In fact, amiodarone rarely [65] causes torsade de pointes, even in patients whose previous treatment with other antiarrhythmic drugs has been discontinued because of torsade de pointes [6, 8, 66]. Finally, a slow heart rate facilitates early afterdepolarizations because it enhances the K⁺ channel blocking effect of class IA and class III drugs, a phenomenon called reverse use-dependence [67-69]. In contrast, Na⁺ channel blockade by class IA drugs is use-dependent; the drugs exert more Na⁺ channel blockade at faster heart rates [64].

These explanations do not account for the variations in susceptibility to drug-induced torsade de pointes seen in some patients. Some patients develop torsade de pointes without a marked prolongation of the QT interval, and the duration of the QT interval in these patients does not predict the occurrence of torsade de pointes [5, 6, 70]. Additionally, some patients have pronounced prolongation of the QT interval at low doses of class IA drugs; other patients show no significant prolongation of the QT interval even at high doses. These findings suggest that prolongation of the action potential alone does not explain the incidence of torsade de pointes and that other modulatory factors may be involved. Stimulation of adrenergic receptors may play a significant role in clinical and experimental settings [51]. Slow heart rate [13] and hypokalemia [13, 71] must also be considered, and ischemia or hypoxia must be ruled out. In addition, genetic differences among patients may play a role: Some phenotypes may possess K (⁺) channels that are more or less susceptible to blockade.

Effect of Adrenergic Stimulation on Drug-Induced Torsade de Pointes

Although adrenergic stimulation is usually associated with the hereditary long QT interval syndromes, it may also facilitate drug-induced torsade de pointes [27]. The effects of ß₁-adrenergic stimulation on early afterdepolarizations are complex because these effects can both enhance and inhibit these afterdepolarizations. The main effect of ß₁-adrenergic activation is to stimulate intracellular adenylyl cyclase activity, which increases cyclic adenosine monophosphate levels [72]. This activates protein kinase A, which phosphorylates Ca²⁺ channels and K⁺ channels and thereby modulates their function. Phosphorylation of the Ca²⁺ channel leads to increased inward current [73-75] that facilitates the development of early afterdepolarizations. Accordingly, early afterdepolarizations can be elicited by administration of isoproterenol to single cells [47, 76]. On the other hand, phosphorylation of K⁺ channels enhances I_K and I_TO [77-79]. This speeds up repolarization, as evidenced by a reduction in the action potential duration [76], and should inhibit the development of early afterdepolarizations.

Other effects of ß₁-adrenergic stimulation are also contradictory. For example, the activity of the electrogenic membrane-bound Na⁺-K⁺ pump is increased [80]. This ion pump extrudes three intracellular Na⁺ ions in exchange for two extracellular K⁺ ions, thus generating a net outward repolarizing current. ß₁-adrenergic activation of electrogenic Na⁺ pumping may enhance repolarization and thereby inhibit the development of early afterdepolarizations. However, it also hyperpolarizes the resting membrane potential, thereby counteracting automaticity and decreasing heart rate, actions that could facilitate the development of early afterdepolarizations. To further complicate matters, a ß₁-adrenergic-induced positive shift of the activation voltage range of I_f in pacemaker cells [81] enhances spontaneous diastolic depolarization and, thus, increases automaticity, accelerates heart rate, and suppresses early afterdepolarizations. Finally, ß₁-adrenergic stimulation facilitates conduction of early afterdepolarizations from Purkinje fibers to the surrounding myocardium [27, 36]. The net effect of these ß₁-adrenergic effects is unclear, and the role of ß₁-adrenergic activation in enhancing or inhibiting early afterdepolarizations remains to be resolved.

The effects of ₁-adrenergic stimulation are even less well clarified. Most hypotheses about the effects of ₁-adrenergic activation on ion currents suggest that this activation should facilitate the development of early afterdepolarizations. Enhancement of I_Ca-L secondary to cyclic adenosine monophosphate-independent [73, 82] phosphorylation of the Ca²⁺ channel by protein kinase C [73, 83, 84] increases net inward current [74]. In addition, ₁-adrenergic blockade of I_K and I_TO delays repolarization and prolongs the action potential duration [77, 85]. Finally, ₁-adrenergic stimulation of electrogenic Na⁺-K⁺ pumping [86] hyperpolarizes the membrane and decreases heart rate. The significance of other effects of ₁-adrenergic stimulation is more speculative. Intracellular Ca²⁺ levels increase after enhanced release from cellular stores by the ₁-adrenergic-induced increase in inositol triphosphate [87, 88]. This elevation of cytosolic Ca²⁺ clearly triggers the development of delayed afterdepolarizations, but its role in early afterdepolarizations is controversial. It is clear that although (₁-adrenergic) stimulation alone cannot elicit early afterdepolarizations [35, 76], it does facilitate induction of early afterdepolarizations and torsade de pointes by K⁺ channel blockers [36, 38, 51].

Role of Heart Rate

Extreme bradycardia as seen during complete or high-grade atrioventricular block may lead to early afterdepolarizations and torsade de pointes [89-92] through prolongation of the action potential duration secondary to both bradycardia-dependent depression of electrogenic Na⁺ pumping [93] and more complete inactivation of I_K [93]. Also, slow heart rates are associated with submaximal activation of I_TO, which shifts the action potential plateau to a voltage range in which the availability of the Ca²⁺ "window" current is increased [94]. Finally, as mentioned above, torsade de pointes caused by class IA and class III antiarrhythmic drugs is enhanced at slow heart rates because these drugs show reverse use-dependence [67-69] and K⁺ blockade is increased. Fast heart rates, of course, tend to oppose these actions, preventing early afterdepolarizations and torsade de pointes.

Effect of Hypokalemia

Hypokalemia prolongs the cardiac action potential and may cause early afterdepolarizations and torsade de pointes both clinically [91] and experimentally [58]. Hypokalemia reduces net outward current by depressing I_K, I_K1, and electrogenic Na⁺ pumping [14, 58]. On the other hand, it enhances inward current by increasing I_Ca-L [58]. Drug-induced torsade de pointes occurs often in the setting of hypokalemia [12, 13].

Ischemia and Hypoxia

No direct evidence suggests that ischemia and hypoxia are involved in drug-induced torsade de pointes. In animal models, early afterdepolarizations and torsade de pointes can be produced in the absence of ischemia and hypoxia [76]. Clinically, torsade de pointes is usually not associated with electrocardiographic or clinical signs of myocardial ischemia. On the other hand, it is difficult to refute the possibility of this association, given the high incidence of coronary artery disease in patients treated with class IA and class III drugs [95]. Theoretically, ischemia could facilitate torsade de pointes in several ways. Increased dispersion of refractoriness and slow conduction could favor the perpetuation of early afterdepolarizations by a reentrant mechanism. In addition, ischemia-induced increases in numbers of ₁-adrenergic receptors [96], increased efficiency of effector-receptor coupling [96], and release of norepinephrine from local nerve endings [97] all enhance the net ₁-adrenergic activation, thereby facilitating early afterdepolarizations. Finally, experimental studies have shown that class III drugs prolong the action potential duration and induce early afterdepolarizations more easily in ischemic than in normoxic Purkinje fibers [48].

Other Causes of the Acquired Long QT Interval Syndromes and Torsade de Pointes

Many nonantiarrhythmic drugs and clinical conditions other than those discussed above have been associated with the acquired long QT interval syndromes and torsade de pointes (Table 1). The underlying mechanisms of this association may involve blockade of I_K (bepridil [98, 99], erythromycin [100, 101], terfenadine [102], and cocaine [103, 104]); stimulation of I_Ca-L (prenylamine [105, 106]); or stimulation of I_Na (ketanserin [107]). In most instances, however, the mechanism is unclear [108-133]. Clinically, it is often unclear whether these agents per se cause torsade de pointes [113] or whether related factors such as hypokalemia or drug-drug interactions are responsible. For example, hypokalemia has been associated with bepridil [98], ketanserin [134, 135], hyperparathyroidism [136], hyperaldosteronism [137], and subarachnoid hemorrhage [138]; all of these agents and conditions are associated with the acquired long QT interval syndromes and torsade de pointes. Also, although ketoconazole and itraconazole have not been shown to cause torsade de pointes, concomitant use of these agents with terfenadine has been associated with the arrhythmia. The mechanism responsible appears to involve an increase of terfenadine in the blood to toxic levels; this is caused by the inhibition by itraconazole or ketoconazole of the enzymatic degradation of terfenadine through the cytochrome P-450 system [139, 140]. Although it has been suggested that overdoses of tricyclic or tetracyclic antidepressants may prolong the QT interval and may have a proarrhythmic effect, especially in the presence of ischemia, the induction of torsade de pointes by these agents has not been unequivocally shown [111, 112].

Alternative Hypotheses for the Acquired Long QT Interval Syndromes

The theory that early afterdepolarizations cause the long QT interval syndromes and torsade de pointes is disputed. Various other hypotheses have been proposed, including those involving reentry, intracellular Ca²⁺ overload, and myocardial M cells.

Reentry

Although torsade de pointes is believed to be initiated by early afterdepolarizations, it is unclear whether the subsequent beats of the arrhythmia necessarily result from successive early afterdepolarizations originating from different foci or whether they might result from sustained conduction within a reentrant pathway with different exit points or circuits [56]. That torsade de pointes is not generally inducible by programmed electrical stimulation has been used as an argument against reentry as the underlying mechanism of the tachyarrhythmia [18, 89, 141]. However, the observation that increased dispersion of refractoriness—reflected by an increase in the QT interval dispersion on the surface electrocardiogram—often correlates with the occurrence of drug-induced [8, 25, 52] and hereditary [142] torsade de pointes supports the hypothesis that reentry is the underlying mechanism [143, 144].

The development of early afterdepolarizations results in the prolongation of the action potential duration and the QT interval on the surface electrocardiogram. If early afterdepolarization occurs heterogeneously throughout the heart—for example, only in some regions of the Purkinje network—dispersion of the repolarization (QT interval) increases, which may facilitate reentry [145, 146]. Thus, reentry may be the mechanism by which torsade de pointes, initiated by an early afterdepolarization, may be perpetuated. Of course, the dispersion of repolarization may be an incidental finding that necessarily follows from the spatially nonhomogeneous distribution of early afterdepolarizations but that does not by itself lead to reentry. Preliminary data from experimental mapping studies [39, 42] in which torsade de pointes was induced by administration of cesium suggest that triggered activity from more than one site is involved; these studies have found no evidence for reentry.

Intracellular Ca²⁺ Overload

Another theory proposes that intracellular Ca²⁺ overload underlies early afterdepolarizations [43, 147]. The role of intracellular Ca²⁺ overload in the genesis of delayed afterdepolarizations, a type of rate-dependent triggered activity, is well recognized [148], for example, in digitalis toxicity. Delayed afterdepolarizations result from the activation of the transient inward current (I_TI)_by elevations of cytosolic Ca²⁺ caused by numerous agents or conditions [19, 149]. This current is carried by either a nonselective cation channel or the electrogenic Na⁺-Ca²⁺ exchanger [149] that is stimulated by oscillatory Ca²⁺ release from the sarcoplasmic reticulum. In contrast, the role of Ca²⁺ overload in early afterdepolarizations and torsade de pointes is controversial. In vitro evidence [147] suggests that intracellular Ca²⁺ loading during rapid pre-pause heart rates during a short-long-short sequence may enhance the development of early afterdepolarizations during the beat after the long pause. Also, it has been proposed that the Na⁺-Ca²⁺ exchanger may carry the depolarizing charge for early afterdepolarizations [43]. Indeed, reducing Ca²⁺ overload with flunarizine suppresses experimental sotalol-induced torsade de pointes [49]. However, inhibiting release of Ca²⁺ from the sarcoplasmic reticulum with ryanodine abolished early afterdepolarizations in some studies [50, 76] but not in others [40]. The role of the Na (⁺) -Ca²⁺ exchanger in the generation of early afterdepolarizations is also controversial. An in vitro study [76] showed that benzamil and reduced extracellular Na⁺ concentration abolished early afterdepolarizations induced by isoproterenol, possibly by blocking the Na (⁺) -Ca²⁺ exchanger. However, another study [15] showed that withdrawing extracellular Na⁺ failed to abolish both early afterdepolarizations and the inward current associated with them, suggesting that the Na⁺-Ca²⁺ exchanger does not carry the depolarizing charge for early afterdepolarizations.

M Cells

Some investigators have suggested that a special group of ventricular myocardial cells may play a role in the genesis of early afterdepolarizations Figure 4 [93]. Although early afterdepolarizations are usually not inducible in ventricular endomyocardium, they can be elicited in isolated single ventricular cells, termed M cells. Some investigators have explained this by proposing that these single ventricular cells, M cells, are a distinct type of myocardial cell originating in deep subepicardial or midmyocardial layers. Electrophysiologically, M cells closely resemble Purkinje fibers, except for the absence of spontaneous diastolic depolarization [93]. Of special interest is the finding that the cells show a marked prolongation of the action potential duration at slower stimulation rates [93], making them the possible origin of early afterdepolarizations during bradycardia. Further research on this interesting group of cells is needed.

The Hereditary Long QT Interval Syndromes

The hereditary (Romano-Ward and Jervell and Lange-Nielsen syndromes) and acquired long QT interval syndromes share many features, including similar morphologies on surface electrocardiogram, the typical short-long-short initiating sequence, abnormal U waves, prolongation of the QT or QTU interval, and facilitation by hypokalemia. Clinically, however, hereditary (but not acquired) long QT interval syndromes are clearly related to adrenergic stimulation or enhancement of sympathetic nervous system tone. Typically, bizarre and dynamic long QT or long QTU segments are observed during periods of ß₁ – adrenergic stimulation or excitement associated with intense sympathetic activation (Figure 6).

View larger version (59K): [in this window] [in a new window]   Figure 6. Example of extreme prolongation of the QT interval in a 21-year-old woman with the hereditary long QT interval syndrome (the Romano-Ward syndrome). This 12-lead electrocardiogram was obtained 48 hours after the patient had a cardiac arrest. The patient had a long history of exertion-related syncope, probably due to recurrent episodes of adrenergic-dependent torsade de pointes. The QT segment was bizarrely inverted in all leads and measured 520 ms even though the RR interval was only 630 ms. Because of the dramatic T-wave changes, the patient was initially misdiagnosed with myocardial ischemia, but a coronary arteriogram and a left ventricular angiogram were normal. She was successfully treated with ß-adrenergic blockade and overdrive permanent pacing.

Although the hereditary long QT interval syndromes were considered idiopathic in the past, a recently discovered defect in the Harvey ras-1 gene on chromosome 11 [150] points to a possible single gene defect as the cause of the syndromes. The specificity of this genetic defect is equivocal because it is not present in all families with hereditary long QT interval syndromes [151]. Nevertheless, this gene encodes for a G-protein-type molecule; G proteins are intracellular elements that transduce the signal resulting from agonist binding to a membrane-bound receptor to intracellular effector systems [73]. Evidence suggests that the gene product of the Harvey ras-1 gene may help to modulate K⁺ channel function, especially in response to ß₁-adrenergic activation [152]. The genetic defect may cause malfunction of some K⁺ channels, reducing outward repolarizing K⁺ current and prolonging the action potential duration [153]. This defect may also lead to an inappropriate adaptive response of the action potential duration to changes in heart rate [154].

It has been shown that patients who develop torsade de pointes while receiving a class IA drug may have a longer QT interval at baseline [89, 95] and may respond to drug treatment with a more marked prolongation of the QT interval [89] than patients who do not have drug-induced proarrhythmia. Also, a reduction in pacing rates has been associated with greater prolongation of the QT interval in patients who develop bradycardia-induced torsade de pointes than in patients who do not [89]. It is tempting to speculate that some patients with the acquired long QT syndromes may have a subclinical genetic defect that allows agents that prolong the action potential duration to have an effect on the QT interval that is more pronounced than that in patients who lack this defect [95]. Thus, although it is by no means clear that patients with the acquired long QT interval syndromes have a genetic defect, some investigators have suggested that the acquired and hereditary long QT interval syndromes may be different expressions of a shared genetic basis [5]. Given the various ionic mechanisms that can cause early afterdepolarizations and torsade de pointes, some patients may have a genetic defect involving other ion channels. Furthermore, studies show that female sex is an independent risk factor for the development of torsade de pointes and sudden cardiac death in patients with the hereditary [155] and acquired [156] long QT interval syndromes, suggesting that genetic predisposition is important to this syndrome. Clearly, more research is needed.

Short- and Long-Term Treatments of the Long QT Interval Syndromes and Torsade de Pointes

The cornerstone of effective treatment of the acquired and hereditary long QT interval syndromes and torsade de pointes is the identification and withdrawal of the offending agent or reversal and correction of any underlying predisposing condition. After this is done, a rational treatment plan can usually be formulated based on an understanding of the ionic basis for the specific syndrome. In the future, specific tailored therapy may target the specific ionic currents responsible for the induction of the early afterdepolarizations. For example, if the dominant charge carrier for the early afterdepolarizations is the Na⁺ "window" current, Na⁺ channel blockade may be more effective [58], but Ca²⁺ channel blockers may be more effective when the Ca²⁺ "window" current is critical [58]. It has been shown experimentally that early afterdepolarizations induced with Anemonia sulcata toxin II, or other drugs that delay inactivation of the Na⁺ channel, cannot be suppressed with Ca²⁺ channel blockers [20]. In contrast, early afterdepolarizations induced by K⁺ channel blockers such as clofilium, cesium, and quinidine can be suppressed by K⁺ channel openers [28, 29, 41, 44, 45, 157], Ca (²⁺) channel blockers [27, 41, 61], and Na⁺ channel blockers [21, 24, 27, 41, 46, 53].

Direct extrapolation from experimental studies to bedside management is not yet possible. Therefore, treatment is generally directed at suppressing or preventing early afterdepolarization generation and conduction by increasing the magnitude of repolarizing currents by enhancing outward K⁺ currents, decreasing the magnitude of depolarizing currents by enhancing Na⁺ and Ca²⁺ channel blockade, or inhibiting conduction of early afterdepolarizations from their site of origin to the surrounding myocardium. Table 4 summarizes short- and long-term treatment options.

Activation of the K⁺ Channel

Short-term treatment that appears to be effective includes the enhancement of I_K using overdrive pacing, lidocaine [158, 159], or infusion of isoproterenol, although the infusion of isoproterenol is contraindicated in the hereditary long QT interval syndromes. Although no clinical studies have yet confirmed the efficacy of enhancement of I_K, experimental studies have suggested that torsade de pointes and early afterdepolarizations can be suppressed by opening K⁺ channels and thereby shortening the action potential duration. For instance, activating the adenosine triphosphate-sensitive K⁺ current (I_K-ATP) [160] with pinacidil [28, 29, 61], cromakalim [29], or nicorandil [44] or activating the acetylcholine-sensitive K⁺ current (I_K-ACh) [161] with vagal stimulation [41, 45] or administration of acetylcholine [143] may suppress early afterdepolarizations and torsade de pointes. These options are currently considered experimental, and further research is needed to determine whether they will be effective.

Blockade of the Na⁺ Channel

Although torsade de pointes can be induced by class IA drugs that have both Na⁺ and K⁺ blocking properties, it can be suppressed using pure Na⁺ channel blockade by reducing the depolarizing current (especially the Na⁺ "window" current). The class IB drug lidocaine and the puffer fish toxin tetrodotoxin suppress early afterdepolarizations and torsade de pointes induced by K⁺ channel blockers *RF 21,24,27,41,46,53 *, Na⁺ channel agonists [57, 58], Ca²⁺ channel agonists [15], and hypokalemia [58] in experimental preparations. Case reports [90, 162] suggest that lidocaine may also be clinically effective, although its effects are inconsistent [6]. Lidocaine is also advantageous because it may enhance outward K⁺ currents [160, 161]. The efficacy of Na⁺ channel blockade may also partially result from the inhibition of conduction of early afterdepolarizations from the Purkinje network to the myocardium [53]. Because clinical experience with Ca²⁺ channel blockers has been minimal and because of concern about possible negative chronotropic effects Ca²⁺ channel blockers are not considered first-line therapy for either the hereditary or the acquired long QT interval syndromes.

Blockade of the Ca²⁺ Channel

Verapamil suppresses premature ventricular depolarizations and reduces the amplitude of early afterdepolarizations (as assessed by recordings of the monophasic action potential) in patients with the hereditary long QT interval syndromes and bradycardia-dependent torsade de pointes [145, 163, 164]. In experimental preparations, numerous Ca²⁺ channel blockers, including verapamil [15, 27], D-600 [61], nifedipine [27], nitrendipine [15, 61], and cadmium [58], suppress early afterdepolarizations and torsade de pointes, but the effect of diltiazem varies [41, 51]. In general, suppression of torsade de pointes and early afterdepolarizations by Ca²⁺ channel blockade is not associated with shortening of either the QT interval or the action potential duration.

Magnesium

Intravenous administration of magnesium is an effective and safe way to clinically suppress torsade de pointes [165]. Although magnesium may inhibit the ischemia-dependent enhancement of I_K-ATP [166], it abolished early afterdepolarizations and torsade de pointes induced by cesium, quinidine, and 4-aminopyridine in experimental studies [21, 34, 35, 41]. The effect of magnesium on early afterdepolarizations and torsade de pointes may be mediated by blockade of Ca²⁺ or Na⁺ "window" currents [24, 27, 35, 167]. Regardless of the mechanism, the therapeutic effect of magnesium is not associated with a reduction of the action potential duration [90, 150].

Isoproterenol Infusion and Pacing

As mentioned above, adrenergic stimulation is contraindicated in torsade de pointes associated with the hereditary long QT interval syndromes [162]. Adrenergic stimulation may facilitate induction and propagation [37] of early afterdepolarizations in the acquired long QT interval syndromes, but, from a clinical standpoint, infusion of isoproterenol is often effective in the strongly pause-dependent forms of the syndrome. Whereas low doses of isoproterenol may enhance the depolarizing current necessary for early afterdepolarizations, higher doses usually suppress early afterdepolarizations and torsade de pointes by enhancing outward K⁺ currents, accelerating heart rate and repolarization, and shortening the action potential duration [27, 70, 77-81]. Overdrive temporary or permanent electrical pacing is also an effective treatment for both forms of the long QT interval syndromes, possibly because it enhances repolarizing K (⁺) currents or prevents long pauses, either of which may suppress early afterdepolarizations and shorten the QT interval *RF 6,21,59,70,90,91,168-171 *.

ß-adrenergic Blockade and Surgical Sympathectomy

All patients with the hereditary, adrenergic-dependent long QT interval syndromes require treatment with ß₁-adrenergic receptor blocking agents. Additional treatments may be required in some patients, especially those clinically judged to be at high risk (for example, those who have had cardiac arrest). Some authorities [172] have proposed high thoracic left sympathectomy (surgical removal of the first four or five left-sided thoracic stellate ganglia) as either an initial treatment or adjuvant therapy in patients failing pharmacologic ß₁-adrenergic blockade alone. Left sympathectomy is effective because the predominantly left-sided sympathetic innervation of the heart underlies the adrenergic-dependent long QT interval syndromes [172]. The fact that left sympathectomy is effective in patients in whom ß₁-blockade has failed suggests that ₁-adrenergic stimulation may play a role in the genesis of the syndromes [172].

In some experimental models of the acquired drug-induced long QT interval syndromes, ß₁-adrenergic blockade suppresses torsade de pointes [41], possibly by suppressing propagation of early afterdepolarizations from Purkinje fibers to the myocardium [37]. Although it is of experimental interest, ß₁-adrenergic inhibition is contraindicated in the acquired forms of the long QT interval syndromes and torsade de pointes because it decreases heart rate.

Implantable Cardioverter-Defibrillators

High-risk patients with a hereditary long QT interval syndrome, or patients with an acquired form of the syndrome that cannot be treated effectively by simply withdrawing the offending agent or condition, may require aggressive management with an implantable cardioverter-defibrillator. Designating a patient as "high-risk" requires a clinical judgment, but patients with recurrent episodes of syncope, those who have survived a cardiac arrest, or those that have failed to respond to conventional medical therapy clearly should be viewed as high risk [172].