The introduction of the implantable cardioverter defibrillator (ICD) as a therapeutic option for patients with life-threatening ventricular arrhythmias has focused interest in factors affecting defibrillation energy requirements. Clinical (1-5) and experimental animal studies (6-10) have demonstrated that antiarrhythmic drugs may modulate the energy required for successful defibrillation. Apart from N-acetyl procainamide (NAPA), clofilium, and d-sotalol, which were found to decrease defibrillation energy requirements, the majority of antiarrhythmic agents, including lidocaine, flecainide, propafenone, and verapamil, have been demonstrated adversely to affect defibrillation energy requirements (11). Some antiarrhythmic drugs (including procainamide, quinidine, and amiodarone) have shown a mixed effect on successful defibrillation (11). The majority of patients with ICDs receive concomitant antiarrhythmic drug therapy (12). The observation that the majority of antiarrhythmic agents adversely influence defibrillation success has important clinical implications. Because of the limited amount of energy delivered by ICDs, a drug-induced increase in defibrillation energy requirements may result in device failure and sudden cardiac death. The aim of this study was to investigate the effect of lidocaine, ajmaline, and diltiazem on defibrillation energy requirements.
Twenty-four New Zealand white rabbits of either sex weighing between 3 and 4.3 kg were used in this study. All animal handling and operation procedures were performed according to the guidelines of the American Physiological Society and approved by the Animal Investigations Committee of the University of Limburg, The Netherlands. After anesthesia with 0.2 ml/kg etomidate and 0.2 μg/kg fentanyl and heparinization (2,000 IU, i.v.), the animals were killed by cervical dislocation. Through a midsternal thoracotomy, the heart was rapidly removed and placed in cold perfusion solution (10°C). The aorta was cannulated, and the heart was attached to a Langendorff perfusion system.
The perfusate was a Tyrode's solution with the following millimolar composition: NaCl, 130; KCl, 5.6; NaHCO3, 24.2; CaCl2, 2.2; MgCl2, 0.6; NaH2PO4, 1.2; and glucose, 12. The perfusate had a pH of 7.35 when equilibrated with a 95% O2/5% CO2 gas mixture and was maintained at 37.5°C. The perfusion of the heart was controlled by a fixed perfusion pressure at 60 mm Hg resulting in a coronary flow of 48 ± 8 ml/min. To avoid heat loss and evaporation, the heart was immersed in a double-walled vessel with thermoconstant perfusion, filled with perfusate, and kept at 37.5°C.
Bipolar left ventricular electrograms were recorded through an olive electrode with 14-mm diameter containing six electrodes with 7.3-mm interelectrode distance placed in the left ventricular cavity. Surface electrocardiogram (ECG) was recorded through two plate electrodes, made of stainless steel, fixed inside the double-walled vessel (Fig. 1). ECG and left ventricular electrograms were displayed on multichannel oscilloscopes (Tektronix, amplifier type 2A61 and 5A18N, Beaverton, OR, U.S.A.).
Ventricular fibrillation was induced with a burst of rectangular stimuli of 2-ms duration at 2 times diastolic threshold and 50- to 70-ms cycle length delivered through the olive electrode. After 10 s of ventricular fibrillation, a defibrillation shock was delivered. For defibrillation, two paddle electrodes with a surface area of 2.2 cm2 were placed in the tissue bath in front of each ventricle in a loose contact to the heart. The biphasic truncated exponential pulses with 60% tilt at variable pulse duration were delivered by an external defibrillator unit (external cardioverter/defibrillator model 2394; Medtronic, Minneapolis, MN, U.S.A.).
Defibrillation threshold (DFT) testing protocol. A first determination of defibrillation requirements was obtained by progressively 20% decreasing the voltage starting at 140 V until failure occurred. Six voltages in 20% steps then were selected immediately surrounding this value (two greater and three less), which were expected to cover the range between failure and success in defibrillation. These six voltage levels were tested 6 times each in random order. An unsuccessful shock was followed immediately by a high-voltage rescue shock. The following equations were used to calculate the delivered energy: Equation (1) where the energy stored (ES) in a capacitance is half the product of the capacitance (C) and the square of the voltage across the capacitor (Vc). Therefore the energy loss due to the shock at a given tilt equals Equation (2)
Only a part of this energy is delivered externally, namely Equation (3) where Ri and Rl are the internal and the lead resistances, respectively. The value of Rl was calculated from the measured pulse width Δt, by using the Equation (4)
ECG and electrogram measurements
Measurements included RR interval, QRS duration, PR interval and rate (R), corrected QT (i.e., QT/√R−R). The ventricular effective refractory period (ERP) was defined as the longest S1-S2 interval, with S2 decrements of 5 ms, not resulting in a propagated premature impulse during pacing with an S1-S1 interval of 300 ms. The ventricular fibrillation cycle length (VFCL) was determined by averaging 10 ventricular activation intervals 2 s before defibrillation. The mean values were calculated from all fibrillation episodes during each experimental phase.
Four study groups each consisting of six isolated hearts were studied. Each study comprised four sequential phases referred to as phases 1 to 4, with a baseline predrug testing performed in phase 1. Each heart served as its own control, with a stabilization period of 1 h after attaching the heart to the Langendorff perfusion system. In each phase, the DFT testing was performed as described. Drugs and Tyrode's solution were infused with an infusion pump (Becton Dickinson, Program 1; VIAL Medical, Brezins, France) to the mainstream through a sidehole of the aortic connecting tube. The perfusion of the heart, as described previously, was controlled by a fixed perfusion pressure determined by the height of the water column and not a fixed perfusion flow set by the pump. The coronary flow, therefore, depended on perfusion pressure and coronary resistance. The coronary flows were measured in each experiment to calculate the drug concentration by using following calculation: Equation (5) where Qd is flow (ml/min) of solution containing the drug, Qh is flow (ml/min) through the heart, Cd is concentration (ng/ml) of drug in syringe solution, and Ch is concentration (ng/ml) of drug in heart perfusate. The difference in flows Qh and Qd causes a dilution of the drug.
Control studies (six hearts). To determine the stability of the preparation over time during baseline condition (phase 1) and phases 2 to 4, defibrillation energy requirements and other measurements were obtained as described previously. During phases 2 to 4, Tyrode's solution was administered to simulate the drug infusion.
Drug studies (six hearts per drug). In these experiments, in each heart measurements were performed at baseline predrug condition (phase 1) followed by drug infusion in phases 2 to 4. Drugs were diluted in Tyrode's solution at low concentration and infused at three adjusted flows corresponding to low, medium, and high concentrations equivalent to therapeutic plasma levels as reported for plasma protein binding (13-15). Drugs were administered as a loading infusion for 20 min and followed by maintenance infusion during measurement periods. Measurements were begun 10 min after the beginning of the maintenance infusion. Higher drug concentrations were achieved by administering further loading and maintenance sequences without a washout period between the sequential experimental phases 2 through 4. These concentrations were administered: lidocaine (Lidocorit; Gebro Broschek, Fieberbrunn, Austria) 1.5, 2, and 3 μg/ml; ajmaline (Gilurytmal; Giulini Pharma GmbH, Hanover, Germany) 0.1, 0.3, and 0.5 μg/ml; and diltiazem (Dilzem; Gödecke AG, Berlin, Germany) 0.025, 0.04, and 0.05 μg/ml.
Analysis of data was performed by using one-way and two-way analysis of variance (ANOVA) and applying an a posteriori test to locate the actual differences (Duncan's multiple-range test). The 50 and 80% effective defibrillation energies were calculated by using nonlinear logistic regression (16): y = ex/(1 + e), where y is the proportion of success at voltage level ED, x is [In4 (ED − ED50)]/(ED80 − ED50), and ED is the energy dose. Therefore estimated ED50 and ED80 were obtained directly rather than by interpolation. We calculated the maximum-likelihood estimates of the regression parameters (slope, intercept) and response rate for our assay response data by using the probit model of statistical analysis system (SAS). The distribution function was chosen to be normal (probit model) and logistic (logistic model), respectively. Because both distributions are symmetric about zero, there was, as for most problems, little difference between the two models. Consequently, only the results of the probit model are listed in our study. The confidence limits were computed by using a critical value of 1.96, which corresponds to an approximate 95% confidence interval.
The analysis represents the data from 24 hearts with a total of 3,456 episodes of ventricular fibrillation and defibrillation. Considerable interindividual variations with regard to heart weight (12 ± 2.1 g; range, 10.10-17 g) and defibrillation energy requirements at baseline (ED50, 0.58 ± 0.2 J; range, 0.19-0.82 J) were present. However, there was no significant correlation between defibrillation energy requirements and heart weight (p = 0.06; r = 0.31).
As summarized in Tables 1 and 2, ED50 and ED80, as well as RR interval during sinus rhythm, QRS and PR duration, QTc interval, and VFCL did not change significantly during all four phases. Figure 2 A shows defibrillation dose-response curves for one experiment.
Effects of lidocaine
The effects of lidocaine in six hearts are summarized in Tables 1 and 2. Lidocaine significantly increased defibrillation energy requirements at all study phases. As compared with baseline, low, medium, and high concentrations of lidocaine caused an increase in ED50 and ED80 to 146, 223, and 312% and 139, 207, and 285%, respectively (Table 2). Defibrillation energy requirements were also significantly increased compared with control and ajmaline series (p = 0.01; two-way ANOVA). However, the slope of the dose-response curves remained stable (6.94 in phase 1 to 5.86 in phase 4; see Table 2). Defibrillation dose-response curves for one experiment are depicted in Fig. 2B. As detailed in Table 1, the RR interval during sinus rhythm increased significantly in phase 2 and remained without significant changes from phases 2-4. The ERP significantly increased at all concentrations, whereas VFCL significantly lengthened only at medium and high lidocaine concentrations. The QTc shortened in phase 3 and 4 significantly. The increase in VFCL was linearly correlated with ED50 (r = 0.74; p = 0.0001) and ED80 (r = 0.67; p = 0.0004). Other changes in ECG measurements did not correlate with changes in ED50 and ED80.
Effects of ajmaline
The effects of ajmaline in six hearts are summarized in Tables 1 and 2. Ajmaline administration was associated with a significant increase in ED50 and ED80 at all study phases. As compared with baseline, low, medium, and high ajmaline concentrations increased ED50 and ED80 to 133, 175, and 251% and 135, 208, and 285%, respectively (Table 2). However, compared with control series, this increase in defibrillation energy requirements was not significantly different (two-way ANOVA). There was not only a shift of the dose-response curves to the right but also a flattening of the slope of the sigmoidal curves (7.04 in phase 1 to 4.24 in phase 4; see Table 2). Defibrillation dose-response curves for one experiment are depicted in Fig. 2C. As listed in Table 1, RR interval during sinus rhythm, QRS duration, and VFCL increased significantly from phase 1 to 4. The ERP and QTc were prolonged significantly during administration of medium and high ajmaline concentrations. ED50 and ED80 correlated linearly with the mean VFCL (r = 0.69; p = 0.0002; and r = 0.71; p = 0.0001) as well as with the mean duration of the QRS complex during sinus rhythm (r = 0.60; p = 0.002; and r = 0.69; p = 0.0002). Furthermore, a linear relation was observed between VFCL and QRS duration (r = 0.60; p < 0.005).
Effects of diltiazem
The effects of diltiazem in six hearts are summarized in Tables 1 and 2. Diltiazem administration resulted in a significant increase in defibrillation energy requirements at all concentrations. As compared with baseline, administration of low, medium, and high diltiazem concentrations resulted in an increase in ED50 and ED80 to 175, 236, and 334% and 158, 212, and 286%, respectively (Table 2). Defibrillation energy requirements were also significantly increased compared with control and ajmaline series (p = 0.01; two-way ANOVA). There was a parallel shift of the dose-response curves to the right without alteration of the slope (8.64 in phase 1 to 7.57 in phase 4; see Table 2). Defibrillation dose-response curves for one experiment are depicted in Fig. 2D. As detailed in Table 1, no change in VFCL was found. The RR interval during sinus rhythm increased significantly from phase 2 to phase 4. The PR interval was prolonged significantly in phases 3 and 4 compared with all previous phases. The ERP was prolonged significantly in phase 4.
The results of this study demonstrate that lidocaine, ajmaline, and diltiazem significantly increase ventricular defibrillation energy requirements in a dose-dependent manner while typical electrophysiologic effects of the drugs are present.
The results of the control studies demonstrated no changes in electrophysiologic variables, including defibrillation properties, over time, proving the suitability and stability of the Langendorff isolated rabbit heart model for assessing ventricular defibrillation energy requirements.
Lidocaine significantly increased defibrillation energy requirements in a dose-dependent manner. This finding is consistent with prior canine experiments that reported a linear dose-dependent relation between lidocaine and defibrillation energy (8-10). These data have recently been confirmed in patients by Echt et al. (17), indicating that results from experimental animal studies may be transferred to human beings. In contrast to our study, adrenergic stimulation was mandatory for lidocaine to increase defibrillation energy requirements in a previously reported isolated rabbit heart preparation (18). The underlying mechanism of this finding has not been defined, and we might argue a dose-related cause because the dosage of lidocaine used in this study was not reported. In our isolated rabbit heart preparations, lidocaine not only significantly increased the mean 50 and 80% effective defibrillation energy requirements without catecholamine administration but also was associated with typical alteration of electrophysiologic parameters as determined in whole animal studies (10,19). Although lidocaine administration resulted in an increase in both ERP and VFCL, we could not detect any relation between these two parameters. In our series, lidocaine administration had no measurable effect on the QRS width during sinus rhythm. This is consistent with findings by Morady et al. (20), who demonstrated a rate-dependent depression of intraventricular impulse conduction by lidocaine with measurable effects on the QRS width only at fast pacing rates. During ventricular fibrillation, drugs with rate-dependent effects would demonstrate the highest ion channel block. Thus it is conceivable that the prolonged VFCL during lidocaine administration resulted from depression of intraventricular conduction.
To the best of our knowledge, this is the first study demonstrating that ajmaline significantly increases defibrillation energy requirements. As with lidocaine, the increase in ED50 and ED80 was dose dependent and significantly correlated with the increase in VFCL. In contrast to lidocaine, depression of ventricular impulse conduction, as evidenced by widening of the QRS complexes, occurred during sinus rhythm, indicating that recovery time from sodium channel block is longer with ajmaline than with lidocaine (20,21). The correlation between ajmaline-induced widening of the QRS complex, VFCL, and defibrillation success was significant. As with lidocaine, the precise mechanism by which ajmaline increases defibrillation requirements remains unknown. Babbs (22) hypothesized that the reduction of sodium current, per se, could explain the elevation of the defibrillation threshold. Our data are consistent with Babb's hypothesis, although a correlation does not substantiate the causality between sodium channel blockade and elevation of defibrillation energy requirements. Conflicting findings result from the fact that the increase of the defibrillation energy requirements was not significantly different compared with control experiments and was significantly less compared with lidocaine and diltiazem. Although we might argue that ajmaline treatment would cause fewer adverse effects on the defibrillation threshold, it should be emphasized that considerable interindividual variations make this conclusion doubtful. Furthermore, we observed a remarkable downward shift in the slope of the dose-response curves during ajmaline perfusion. This phenomenon, which was present in all hearts and was dose dependent, would indicate that ajmaline not only leads to higher defibrillation energy requirements but also reduces the defibrillation success rate regardless of higher defibrillation energies delivered.
Diltiazem perfusion was associated with a significant dose-dependent increase in ED50 and ED80. Schräder et al. (23) reported a significant increase in defibrillation energy requirements during verapamil administration in dogs.
In this study there was a significant correlation between the percentage shortening of VFCL and the percentage increase in defibrillation energy requirements, and the authors speculated that shortening of VFCL caused by shortening of the refractoriness resulted in decreased defibrillation efficacy. The effects of diltiazem on defibrillation energy requirements in our study are in accordance with the results of Schräder et al. (23). However, in contrast, we observed neither a significant alteration in VFCL nor any relation between defibrillation energy requirements and ventricular ERP.
Controversial data on a cause-effect relation between VFCL and defibrillation success have been published. In lidocaine studies, a prolongation of VFCL was found (10); in verapamil studies, a shortening of VFCL could be demonstrated (23); and as shown in our study, no changes in VFCL were detectable during diltiazem perfusion. However, all three antiarrhythmic drugs have a common effect: a significant, dose-dependent increase in defibrillation energy requirements. Thus it seems reasonable to conclude that other mechanisms rather than a simple decrease or increase in VFCL may explain the underlying process responsible for defibrillation threshold elevation.
Binding of drugs by plasma proteins may result in lower free concentration in patients. However, the hearts were treated with increasing concentrations of drugs with regard to their plasma protein binding capacity that resulted in alterations of electrophysiologic parameters as observed at therapeutic plasma concentrations in the clinical setting (24,25). It is questionable whether results from experimental animal studies or isolated heart preparations can directly be transferred to humans. Moreover, the presence of cardiac disorders such as ischemia or scars (26) and increased myocardial mass (27) may affect absolute defibrillation energy requirements. However, precisely to determine the effect of a given drug on defibrillation energy requirements, controlled studies with numerous episodes of fibrillation and defibrillation are necessary; for ethical and clinical reasons, these cannot be performed in humans. Nevertheless, in a recent study Echt et al. (17) provided some clinical evidence that the modulation of defibrillation energy requirements in patients with antiarrhythmic drug therapy paralleled the results obtained from extensive defibrillation testing in experimental studies. Thus it is reasonable to suggest that the relative increase in DFT would be comparable in patients regardless of higher baseline thresholds than in isolated rabbit hearts. In this study we did not determine the mechanisms by which drugs modulate defibrillation energy needs. However, according to the ion-current hypothesis, lidocaine and ajmaline increase defibrillation energy requirements with sodium blocking activity. Because of widely differing structural types of calcium channel blockers and therefore their various ancillary effects, a simple extrapolation of the interaction of diltiazem with defibrillation energy needs to other calcium blockers may not be possible. The precise mechanisms involved in ventricular defibrillation still remain unknown; further studies, preferably at the cellular electrophysiologic level, are needed.
The majority of the patients with ICDs require introduction of additional antiarrhythmic drug therapy after device implantation (12). Currently available devices deliver a maximum amount of shock energy of 34-40 J. Clinical studies (28) have shown that without antiarrhythmic drug therapy, the average DFT is ≤12 ± 5 J, giving a sufficient safety margin of up to 3 times this value. In our study, we demonstrated that lidocaine, ajmaline, and diltiazem double and even triple defibrillation energy requirements in a dose-dependent manner. If these data can be transferred to patients, administration of these drugs after device implantation may result in a critical shrinkage of the safety margin with subsequent device failure and sudden cardiac death. Therefore after introduction of additional antiarrhythmic drug therapy in patients with ICDs, retesting is mandatory to ensure ICD efficacy.
Acknowledgment: This investigation was supported in part by Medtronic, Inc., Vienna, Austria. We thank Jens Kalender, M.S., Fred Lindemans, Ph.D., and Karel Smits for their valuable technical support.
1. Fogoros R. Amiodarone-induced refractoriness to cardioversion. Ann Intern Med
2. Marinchak RA, Friehling TD, Kline RA, Stohler J, Kowey PR. Effect of antiarrhythmic drugs
on defibrillation threshold: case report of an adverse effect of mexiletin and review of the literature. PACE Pacing Clin Electrocardiogr
3. Troup PJ, Chapman PD, Olinger GN, Kleinman LH. The implanted defibrillator: relation of defibrillating lead configuration and clinical variables to defibrillation threshold. J Am Coll Cardiol
4. Dorian P, Feindel C, Lipton I. Usefulness of sotalol in conjunction with automatic implanted in man [Abstract]. PACE Pacing Clin Electrocardiogr
5. Reiffel JA, Coromilas J, Zimmerman JM, Spotnitz HM. Drug-device interactions: clinical consideration. PACE Pacing Clin Electrocardiogr
6. Fain ES, Dorian P, Davy JM, Kates RE, Winkle RA. Effects of encainide and its metabolites on energy requirements for defibrillation. Circulation
7. Dorian P, Fain ES, Davy JM, Winkle RA. Effect of quinidine and bretylium on defibrillation energy requirements. Am Heart J
8. Dorian P, Fain E, Davy JM, Winkle RA. Lidocaine causes a reversible concentration-dependent increase in defibrillation energy. J Am Coll Cardiol
9. Babbs CF, Yim GKW, Whistler SJ, Tacker WA, Geddes LA. Elevation of ventricular defibrillation threshold in dogs by antiarrhythmic drugs
. Am Heart J
10. Echt DS, Black JN, Barbey JT, Coxe DR, Cato E. Evaluation of antiarrhythmic drugs
on defibrillation energy requirements in dogs: sodium channel block and action potential prolongation. Circulation
11. Alt E, Klein H, Griffin JC, eds. The implantable cardioverterldefibrillator.
Berlin: Springer Verlag, 1992:147-56.
12. Winkle RA, Mead RH, Ruder MA, et al. Long term outcome with the automatic implantable cardioverter defibrillator. J Am Coll Cardiol
13. Goodman-Gilman A, Rall TW, Niea AS, Taylor P, eds. The pharmacological basis of therapeutics.
New York: Macmillan Pergamon, 1990:1676-88.
14. Vaughan-Williams EM, ed. Antiarrhythmic Drugs: Handbuch der experimentellen Pharmakologie.
Vol 89. Berlin: Springer Verlag, 1989:201-33, 413-21.
15. Köppel C, Wagemann A, Martens F. Pharmacokinetics and antiarrhythmic efficacy of intravenous ajmaline in ventricular arrhythmia of acute onset. Eur J Drug Metab Pharmacokinet
16. Motulsky HJ, Ransan LA. Fitting curves to data using nonlinear regression: a practical and nonmathematical review. FASEB J
17. Echt DS, Gremillion ST, Lee JT, et al. Effects of procainamide and lidocaine on defibrillation energy requirements in patients receiving implantable cardioverter defibrillator devices. J Cardiovasc Electrophysiol
18. Gremillon ST, Larson TD, Brooks MJ, Echt D. Lidocaine's effect on ventricular defibrillation is modulated by adrenergic agents in the isolated rabbit heart [Abstract]. PACE Pacing Clin Electrophysiol
19. Carmeliet E, Saikawa T. Shortening of the action potential and reduction of pacemaker activity by lidocaine, quinidine, and procainamide in sheep cardiac Purkinje fibers: an effect on Na or K currents? Circ Res
20. Morady F, Dicarlo, LA, Baerman JM, Krol RB. Rate-dependent effects of intravenous lidocaine, procainamide and amiodarone on intraventricular conduction. J Am Coll Cardiol
21. Forth W, Henschler D, Rummel W, Starke K, eds. Allgemein und spezielle Pharmakologie und Toxikologie.
Mannheim: Wirtschaftsverlag, 1992:343-65.
22. Babbs CF. Alteration of defibrillation threshold by antiarrhythmic drugs
: a theoretical framework. Crit Care Med
23. Schräder R, Brooks M, Echt DS. Effects of verapamil and Bay K 8644 on defibrillation energy requirements in dogs. J Cardiovasc Pharmacol
24. Saetre H, Ahlmark G, Ahlberg G. Haemodynamic effects of ajmaline in man. Eur J Clin Pharmacol
25. Members of the Sicilian Gambit. Antiarrhythmic therapy: a pathophysiologic approach.
New York: Futura, 1994:85-101.
26. Ouyang P, Brinker JA, Bulkley BH, Jugdutt BI, Varghese PJ. Ischemic ventricular fibrillation: the importance of being spontaneous. Am J Cardiol
27. Geddes LA, Tacker WA, Rosborough JP, Moore AG, Cabler PS. Electrical dose for ventricular defibrillation of large and small animals using precordial electrodes. J Clin Invest
28. Anvari A, Stix G. Grabenwöger M, Schneider B, Türel Z, Schmidinger H. Comparison of three cardioverter-defibrillator implantation techniques: initial results with transvenous pectoral implantation. PACE Pacing Clin Electrophysiol