Introduction

Top Abstract Introduction Materials and Methods Results Discussion Conclusion References

QT interval prolongation is a risk factor in a number of cardiovascular as well as noncardiovascular diseases. In the congenital long QT syndrome, prolongation of the QT interval is associated with recurrent syncope and sudden cardiac death. Both result from potentially life-threatening polymorphic tachycardia of the torsade de pointes (TdP) type. TdP are not only observed in long QT syndrome but also in clinical conditions such as bradycardia (Kurita et al., 1994) or hypokalemia (Shimizu et al., 1991), especially if occurring in the presence of various drugs, which prolong repolarization (Haverkamp et al., 2000).

The most commonly known cause of TdP is the administration of antiarrhythmic drugs (Eckardt et al., 1998a; Haverkamp et al., 2000). These drugs have in common that they prolong repolarization via block of the rapid component of the delayed rectifier potassium current, I_Kr (Haverkamp et al., 2000). In drug-induced TdP, antiarrhythmic agents still play an important role, but the number of noncardiovascular drugs that is associated with QT prolongation and may have a possible proarrhythmic potential has been rising continuously. Estimation of the true incidence of TdP during treatment with these drugs is difficult. For several noncardiovascular drugs, which have been involved in the generation of TdP, only a few case reports are available (Haverkamp et al., 2000).

Among noncardiovascular drugs that prolong repolarization, macrolide antibiotics are widely prescribed. Apart from their antibiotic effects, macrolide antibiotics were found to prolong action potential duration (Ohtani et al., 2000). It was demonstrated that erythromycin prolongs repolarization by a block of I_Kr (Daleau et al., 1995). In several case reports, TdP was reported after oral (Freedman et al., 1987) and in particular after intravenous administration (Nattel et al., 1990). In the early nineties, erythromycin was the most commonly used macrolide antibiotic in the United States. Since 1998, azithromycin has taken the place of erythromycin with more than 30 million prescriptions in the year 2000 (Shaffer and Singer, 2001). Given the widespread use of erythromycin, it should be noted that erythromycin-related arrhythmias are rare in spite of its QT-prolonging potential (Katapadi et al., 1997; Eckardt et al., 1998b). Clarithromycin has a macrolide structure similar to erythromycin and may thus share similar electrophysiological properties and proarrhythmic potential. It is, therefore, not surprising that several cases of clarithromycin-related TdP were reported (Lee et al., 1998; Wasmer et al., 1999). However, for azithromycin case reports have been extremely rare (Samarendra et al., 2001). Apart from a different prescription behavior this may also reflect different electrophysiological properties of the various macrolide antibiotics. We therefore investigated the electrophysiological effects of erythromycin, clarithromycin, and azithromycin in a previously developed model of TdP (Eckardt et al., 1998b, 2002).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Conclusion References

Preparation of Hearts for Perfusion. The method has been described previously (Eckardt et al., 1998b, 2002). In summary, male New Zealand White rabbits (n = 37) weighing 2.5 to 3.0 kg were anesthetized with sodium thiopental (200-300 mg i.v.). After midsternal incision and opening of the pericardium, the hearts were removed and immediately placed in an ice-cold Krebs-Henseleit solution (1.80 mM CaCl₂, 4.70 mM KCl, 1.18 mM KH₂PO₄, 0.83 mM MgSO₄, 118 mM NaCl, 24.88 mM NaHCO₃, 2.0 mM Na-pyruvate, and 5.55 mM D-glucose). The aorta was cannulated, the pulmonary artery was incised, and the spontaneously beating hearts were retrogradely perfused at constant flow (52 ml/min) with warm (36.8-37.2°C) Krebs-Henseleit solution. Perfusion pressure was kept stable at 100 mm Hg. The hearts were placed in a heated, solution-filled tissue bath. After cannulation the hearts were given 10 min to stabilize. The perfusate was equilibrated with 95% O₂ and 5% CO₂ (pH 7.35; 37°C). The cannulated and perfused hearts were attached to a vertical Langendorff apparatus (Hugo Sachs Elektronik, Medical Research Instrumentation, March-Hugstetten, Germany). A deflated latex balloon was inserted into the left ventricle and connected to a pressure transducer to control hemodynamic stability. The atrioventricular (AV) node was ablated by a surgical tweezers under ECG control to slow the intrinsic heart rate. This resulted in complete AV dissociation with a ventricular escape below 60 beats/min.

Electrocardiographic and Electrophysiological Measurements. A volume-conducted ECG was recorded by complete immersion of the heart into a bath of Krebs-Henseleit solution that had been thermally equilibrated with the myocardial perfusate. Signals from a simulated "Einthoven" configuration were amplified by a standard ECG amplifier (filter settings 0.1-300 Hz). Monophasic action potential (MAP) recording and stimulation were accomplished simultaneously using contact MAP-pacing catheters (EP Technologies, Mountain View, CA). The MAP electrograms were amplified and filtered (low pass, 0.1 Hz; high pass, 300 Hz). MAPs were analyzed using software specifically designed by Franz et al. (1995) permitting precise definition of the amplitude and duration of the digitized signals. The recordings were considered reproducible and, therefore, acceptable for analysis only if they had a stable baseline, stable amplitude with a variation of less than 20%, and a stable duration [MAP duration at 90% repolarization (MAP₉₀) was reproducible within 4 ms]. MAPs were recorded simultaneously. Seven MAPs were evenly spread in a circular pattern around both ventricles, and one MAP was recorded from the left endocardium. One of the right-ventricular catheters was used to pace the heart.

Experimental Protocol. After placing the MAP catheters and achieving complete AV block, cycle length dependence was first investigated under baseline conditions. Thereafter, erythromycin, clarithromycin, or azithromycin (150, 200, and 300 µM) were infused. The concentrations for all three macrolides were several multiples higher than the expected free plasma concentrations in patients to create a maximal proarrhythmic milieu and to better study the underlying mechanisms of proarrhythmia. The experimental setup was designed to reproduce conditions and circumstances that are clinically known to be associated with an increased propensity to the development of TdP (Zabel et al., 1997; Eckardt et al., 1998a). Pacing, MAP recording, and measurement of ECG parameters were repeated after drug infusion. Thereafter, the potassium concentration was lowered to 1.5 mM/l to provoke EAD and TdP. Low potassium concentration has been demonstrated to exert additional block of I_Kr, even in the presence of maximal drug-related I_Kr block (Yang and Roden, 1996). Five minutes later, the potassium concentration was increased to 5.8 mM/l, the drug concentration of the macrolide was thereafter increased to the next dosage, and pacing was repeated. Again, this was followed by lowering the potassium concentration for 5 min. The latter two steps were repeated for each drug concentration.

Data Acquisition and Statistical Analysis. ECG, pressure, volume, and MAPs were recorded on a multichannel recorder. Data were digitized on line at a rate of 1 kHz with 12-bit resolution and stored on a disk. All data are presented as mean ± S.D. The influence of each drug on ECG parameters, and MAP duration, as well as dispersion of repolarization was assessed using Friedmann test. Wilcoxon test was used to investigate cycle length dependence. To compare the three drugs we used the Kruskal-Wallis test and the nonparametric Mann-Whitney U test. To compare the incidence of EAD and TdP, the ² test was used.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Conclusion References

Dose-Dependent Effects of Macrolide Antibiotics on QT Interval and Action Potential Duration. All electrocardiographic parameters reached equilibrium within 10 min. MAP recordings and pacing thresholds (mean threshold 1.6 ± 1.4 mA) remained highly reproducible throughout the experimental protocol. After an initial stabilization period of approximately 5 to 10 min, the MAP amplitude did not change by more than 20% for the subsequent investigation period. The macrolide antibiotics led to a dose-dependent prolongation of QT interval and MAP duration (p < 0.001) (Table 1). When azithromycin was administered in the presence of erythromycin, an additional increase in QT interval was observed. Figures 1 and 2 illustrate the dose-dependent effects of macrolide antibiotics on MAP₅₀ and MAP₉₀. In the presence of 300 µM erythromycin, the increase in MAP₉₀ ranged between 14% at a cycle length of 300 ms and 46% at a cycle length of 900 ms. This marked reverse use dependence was also observed with clarithromycin and azithromycin. In the presence of 300 µM clarithromycin, the increase in MAP prolongation ranged between 28% at 300 ms and 45% at 900 ms. With azithromycin, it measured 21% at 300 ms and 46% at 900 ms. For the three drugs, the increase in MAP₉₀ was paralleled by a dose-dependent increase in MAP₅₀ and QT interval. In accordance to MAP₉₀, the increase in MAP₅₀ and QT interval was also cycle length-dependent. When azithromycin was administered in the presence of erythromycin, the prolongation of MAP₉₀ ranged between 37% at 300-ms cycle length and 77% at 900-ms cycle length.

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Cycle length-dependent effects of 150 to 300 µM clarithromycin (n = 13) and 150 to 300 µM erythromycin (n = 13) on mean MAP duration compared with control in isolated Langendorff-perfused rabbit hearts. , baseline; , 150 µM; , 200 µM; , 300 µM macrolide; p < 0.05.

View larger version (41K): [in this window] [in a new window]   Fig. 2.   Cycle length-dependent effects of 150 to 300 µM azithromycin (n = 11) and the combination of 300 µM erythromycin and 150 µM azithromycin (n = 13) on mean MAP duration compared with control in isolated Langendorff-perfused rabbit hearts. , baseline; , 150 µM; , 200 µM; , 300 µM macrolide; p < 0.05.

Dispersion of Repolarization and Early Afterdepolarizations. All antibiotics led to an increase in MAP₅₀ and MAP₉₀ dispersion (p < 0.001) with increasing drug concentration (Table 1). In the presence of 300 µM erythromycin, there was a significant 70% increase in interventricular dispersion of MAP₉₀. For 300 µM clarithromycin, an increase in dispersion of 145% was observed, whereas 300 µM azithromycin led to an increase of 161% (p < 0.05). In the presence of clarithromycin and especially in erythromycin-treated hearts, EADs and triggered activity were a frequent finding. With erythromycin, all hearts showed MAP recordings with EADs after lowering potassium at a concentration of 300 µM. In the presence of azithromycin, no EAD occurred. In the presence of clarithromycin and erythromycin, TdP was always associated with EADs.

Induction of TdP. After complete AV block and increasing the drug concentration to 300 µM as well as lowering potassium concentration from 5.88 to 1.5 mM/l, TdP occurred in 10 of 13 erythromycin- and clarithromycin-treated rabbit hearts, respectively (Fig. 3). Erythromycin and clarithromycin were found to have a similar proarrhythmic potential (Fig. 3). However, with regard to the number of events of TdP episodes, more episodes were observed in the presence of erythromycin (270) compared with clarithromycin (192) (Fig. 4). Noteworthy, no TdP occurred in the presence of azithromycin despite showing the largest increase in QT interval. Furthermore, azithromycin was found to have an antiarrhythmic potential. When it was administered to 13 hearts that were already treated with erythromycin and had demonstrated TdP, azithromycin suppressed TdP in 7 of 10 hearts.

View larger version (67K): [in this window] [in a new window]   Fig. 3.   Dose-dependent occurrence of EADs and TdP. In 10 of 13 hearts, 300 µM erythromycin and clarithromycin led to repetitive episodes of TdP in the presence of low potassium. No EAD or TdP occurred in the presence of azithromycin despite marked QT prolongation. Azithromycin infused after administration of erythromycin reduced the incidence of EAD and TdP significantly. , TdP; , EAD.

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Early afterdepolarizations and torsade de pointes in the presence of erythromycin (300 µM), AV block, and hypokalemia in an isolated Langendorff-perfused rabbit heart. EADs and triggered activity are associated with the occurrence of TdP.

Triangular versus Rectangular MAP Configuration. Azithromycin had a similar effect on MAP₅₀ (55-ms mean maximal prolongation) and MAP₉₀ (67-ms mean maximal prolongation), which resulted in a MAP₉₀/MAP₅₀ ratio of 1.22. In contrast, erythromycin showed an MAP₅₀ of 29 ms (mean maximal prolongation) and MAP₉₀ of 44 ms, which resulted in a MAP₉₀/MAP₅₀ ratio of 1.52. In the presence of clarithromycin, we observed an MAP₅₀ of 33 ms and an MAP₉₀ of 73 ms, which resulted in a MAP₉₀/MAP₅₀ ratio of 2.21 (Table 1). Thus, in erythromycin and clarithromycin, MAP₉₀ is markedly lengthened, whereas MAP₅₀ is prolonged only moderately, rendering the action potential prolongation triangular. In contrast, azithromycin led to similar prolongation of MAP₅₀ and MAP₉₀, resulting in a rectangular action potential (Figs. 5 and 6). This is illustrated by significantly different MAP₉₀/MAP₅₀ ratios (Fig. 7).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Different MAP configurations with clarithromycin, azithromycin, and erythromycin. In comparison with baseline, erythromycin and clarithromycin show a triangular pattern, whereas azithromycin shows a rectangular pattern.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Phase 3 prolongation after administration of erythromycin and clarithromycin in comparison to phase 2 prolongation after administration of azithromycin.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Relationship between APD90 and APD50. Azithromycin (150-300 µM; n = 11) and baseline show significantly lower values compared with erythromycin (150-300 µM; n = 13) and clarithromycin (150-300 µM; n = 13) (p < 0.05).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Conclusion References

The main finding in the present study is that the three macrolide antibiotics erythromycin, clarithromycin, and azithromycin have a different proarrhythmic potential despite similar QT prolongation. The action potential configuration in the presence of the three drugs but not the extent of QT prolongation, nor the increase in dispersion of repolarization, nor different use dependence explained the occurrence of EADs and the different torsadogenic potential. A triangular MAP shape was related to proarrhythmia, whereas a rectangular MAP configuration was not associated with TdP. Moreover, adding azithromycin, which caused a rectangular pattern, to erythromycin, which caused a triangular MAP shape, inhibited previously induced TdP.

Prolongation of action potential duration is considered a major antiarrhythmic mechanism but paradoxically, it frequently is also proarrhythmic and may induce TdP (Haverkamp et al., 2000). This represents the dilemma in the use of class III antiarrhythmic agents but also a great number of noncardiovascular drugs. The macrolide antibiotics erythromycin, clarithromycin, and azithromycin prolong MAP duration and QT interval. Nevertheless, they have a different torsadogenic potential. In a postmarketing analysis on macrolide antibiotics and TdP, the Food and Drug Administration reported a difference in proarrhythmic potential of macrolide antibiotics in a total number of 156 reported patients. Fifty-three percent of reported TdP occurred in the presence of erythromycin, 36% with clarithromycin-, and only 11% in azithromycin-treated patients (Shaffer and Singer, 2001). Our experimental findings are in agreement with this report. We used a setup that does reproducibly induce TdP in isolated rabbit hearts if I_Kr-blocking drugs such as sotalol are administered (Eckardt et al., 1998b). Using this setup, only two of three macrolides induced TdP. Erythromycin and clarithromycin had a comparable proarrhythmic potential, whereas azithromycin showed no proarrhythmic effects although the QT interval, MAP duration, and dispersion of repolarization were markedly prolonged. Our findings are also in agreement with a study by Ohtani et al. (2000) on the in vivo effects of macrolide antibiotics in rats. They also reported that the arrhythmogenic risk of macrolide antibiotics should be ranked as follows: erythromycin greater than clarithromycin greater than azithromycin.

Our data present evidence that in the presence of noncardiovascular drugs that prolong QT interval, the extent of QT prolongation does not necessarily increase the risk for TdP. Moreover, we demonstrated for the first time that azithromycin suppressed TdP induced by erythromycin. Therefore, QT prolongation alone may not serve as a surrogate marker of cardiotoxicity. Although erythromycin resulted in the smallest increase in MAP duration, it was associated with the highest incidence of TdP.

No TdP occurred in the absence of EADs, which have earlier been acknowledged as the most important mechanism underlying TdP in experimental models of TdP (Eckardt et al., 1998a). In the present setting, there was a high incidence of EADs in erythromycin- and clarithromycin-treated rabbit hearts at low levels of extracellular potassium and at slow heart rates, but no EADs occurred in hearts after administration of azithromycin, which resulted in the largest increase in QT and MAP durations. Thus, the occurrence of EADs was not directly related to the degree of QT prolongation. However, EADs were directly linked to the occurrence of proarrhythmia and the lack of EADs with azithromycin corresponded to the lack of TdP with this macrolide antibiotic. EADs are likely to provide the trigger (i.e., premature ectopic beats) that induces proarrhythmia in the presence of the appropriate substrate (i.e., increased dispersion of repolarization, resulting in electrical heterogeneity with nonuniform repolarization and refractoriness) for the initiation and perpetuation of TdP. Triggered activity seems to be the most probable cause for the appearance of ectopic beats preceding TdP, at least for the first beat in a run of TdP, when EADs reach the critical threshold for activation of a depolarizing current. The subsequent beats may then result from circus movement reentry due to dispersion of repolarization (Habbab and el-Sherif, 1990). The manifestation of EADs is usually associated with a critical prolongation of the repolarization phase due to a reduction in net outward current (Antzelevitch and Sicouri, 1994). We demonstrated that different use dependence, or a different increase in dispersion of repolarization could not explain the observed different torsadogenic potential. Erythromycin was reported to cause prominent dispersion of repolarization in the canine ventricular wall (Antzelevitch et al., 1996; Fazekas et al., 1998). In addition, Verduyn et al. (1997) found an increased bradycardia-dependent interventricular dispersion in dogs with chronic AV block after adding the I_Kr blocker sotalol, and they proposed that dispersion of repolarization should be added to the relevant factors for the initiation of TdP. In the present study, the increase in dispersion may be associated with TdP but was not sufficient enough to explain the occurrence of TdP.

Possible Mechanisms for Different Proarrhythmic Potential. Our study points out for the first time that the difference in MAP configuration may be the reason for the difference in the proarrhythmic potential of macrolide antibiotics. Erythromycin and clarithromycin mainly prolonged phase 3 of the action potential.

	Conclusion

Top Abstract Introduction Materials and Methods Results Discussion Conclusion References

The macrolide antibiotics erythromycin, clarithromycin, and azithromycin prolong myocardial repolarization. Compared with erythromycin and clarithromycin, the torsadogenic potential of azithromycin seems to be remarkably low. In the Langendorff-perfused rabbit heart model of TdP, azithromycin did not display the proarrhythmic profile typical for blockers of I_Kr such as erythromycin or sotalol (Eckardt et al., 1998b). The mechanisms responsible for this behavior of azithromycin are probably multifactorial. Although we demonstrated that the three drugs have similar electrophysiological characteristics such as reverse use dependence and dose-dependent increase in dispersion of repolarization, they presented with a significant different potential to induce EADs and TdP. It is possible that a different mode of interaction between azithromycin and the channel and/or additional pharmacological effects of azithromycin may play a role. Triangulation of the action potential observed in the presence of erythromycin and clarithromycin corresponded to the occurrence to TdP, whereas rectangulation of the action potential due to azithromycin had no proarrhythmic effects and suppressed TdP induced by erythromycin. Thus, azithromycin showed some of the characteristics of what may be considered an ideal antiarrhythmic agent with lengthening of action potential duration but with low risk of proarrhythmia. The present study clearly demonstrated that prolongation of the action potential and QT interval may not necessarily be proarrhythmic. In the absence of triangulation of the action potential it may be safe and not result in proarrhythmia. Further investigation in particular intracellular action potential recordings will be necessary to clarify whether drugs that lead to a rectangle action potential prolongation represent the antiarrhythmic agents of the future.