Capsaicin (trans-8-methyl-N-vanillyl-6-nonenamide), the active (pungent) ingredient of hot peppers, has been used as a topical analgesic in clinical medicine. Previous studies demonstrated that capsaicin can have a number of effects on polymodal nociceptive sensory neurons, including acute release and depletion of neuropeptide stores, cell excitation followed by desensitization, and cell necrosis at high concentration (1,2). The capsaicin-induced activation of sensory neurons is likely to be produced by interaction with a specific membrane receptor (3).
The effects of capsaicin on the myocardium have not been extensively studied. Likewise, the effects of capsaicin on the electrophysiologic properties of the heart have received little attention. In studies of isolated rat ventricular myocytes, Castle (4) found that capsaicin inhibited the potassium currents Ito, IK, and IK1, resulting in prolongation of the action-potential duration (APD). This observation suggests that capsaicin may induce cardiac effects apart from any interaction with sensory neurons that innervate the heart. The results also imply that capsaicin could have significant electrophysiologic effects in ventricular myocardium. The presence of capsaicin receptors on myocardial cells has not been determined.
D'Alonzo et al. (5) studied the effects of capsaicin on both reperfusion and spontaneous ventricular fibrillation and ventricular tachycardia in isolated rat and guinea pig hearts by using a Langendorff preparation. They found a reduction in both reperfusion ventricular fibrillation and tachycardia in rat and an antifibrillatory effect on reperfusion in guinea pigs with capsaicin. Capsaicin also reduced ischemic ventricular tachycardia and fibrillation in isolated rat hearts. However, by using standard microelectrode techniques, they found a significant shortening of the APD at 90% repolarization (APD90) of guinea-pig papillary muscle, contrasting with the results of the studies by Castle in isolated cells. The authors speculated that blockade of calcium channels contributed to these effects. The reasons for these conflicting results of the effects of capsaicin on ventricular muscle are unclear.
Zucapsaicin (cis-8-methyl-N-vanillyl-6-nonenamide), also known as civamide, is the synthetically produced conformational cis-isomer of capsaicin. Conformational changes to a compound can result in alterations of its pharmacologic or toxicologic profile. In this respect, civamide appears to be a more potent antinociceptive agent, but less irritating when compared with capsaicin (6; K. M. Verburg, 1996, personal communication). It also is possible that civamide might have myocardial effects different from those of capsaicin. We therefore elected to study the cardiac electrophysiologic effects of civamide in intact dogs and in isolated canine Purkinje fibers. The hypothesis to be tested was that civamide significantly modulated the cardiac Purkinje AP and had antiarrhythmic effects on inducible ventricular tachycardia in the first 3 h after coronary artery occlusion.
Adult mongrel dogs of either sex weighing 18-24 kg were used for these studies. The protocols were approved by the University of Iowa Animal Use and Care Committee and conformed to the guidelines of the American Physiological Society.
Dogs were anesthetized with sodium thiopental, 500 mg, and α-chloralose, 100 mg/kg, intravenously as a bolus. Anesthesia was maintained with a continuous intravenous infusion of α-chloralose at 8 mg/kg/h. The animals were intubated orotracheally and ventilated on a volume-controlled ventilator (Harvard Apparatus, S. Natick, MA, U.S.A.). Tidal volume was adjusted to achieve an arterial Pco2 in the physiologic range, and positive end-expiratory pressure was applied to maintain PO2 in normal range (80-150 torr) by using oxygen-enriched air if necessary. Sodium bicarbonate was infused as necessary to maintain the pH within a physiologic range. Serum electrolytes (K+, Ca2+, and Mg2+) were measured periodically and were always within normal limits. Arterial pressure was continuously monitored by using a Statham P23D transducer (Hato Rey, Puerto Rico) with a fluid-filled polyethylene catheter placed at the aortic arch, via the femoral artery. The femoral vein was cannulated to infuse drugs and isotonic saline.
The heart was exposed by a median sternotomy, and the pericardium was incised and sutured to the wound edges to support the heart. The left anterior descending coronary artery was dissected proximally, and a snare was placed around the vessel. This site of the dissection was chosen to obtain as large an ischemic zone as possible without involving the first septal perforator artery. The sternal wound was covered by a plastic sheet and an infrared heating lamp directed to the heart to maintain thoracic temperature >37°C. After the experiment, the dog was killed by induction of ventricular fibrillation.
For in vitro studies, dogs were anesthetized as described earlier. Their hearts were excised and placed in ice-cold Tyrode's solution containing (in mM): 125 NaCl, 24 NaHCO3, 4.0 KCl, 2.7 CaCl2, 0.5 MgCl2, 0.25 Na2HPO4, and 5.5 dextrose with a pH of 7.4, and equilibrated with 95% O2, 5% CO2. Free running strands of Purkinje fibers were removed from the right and left ventricles and placed in a tissue bath perfused with oxygenated Tyrode's solution at 37 ± 0.5°C. The bath had a volume of 6 ml and was perfused at a rate of 18 ml/minute, allowing the bath contents to be changed 3 times per minute.
Electrophysiologic measurements in vivo
The right atrium was paced with a bipolar electrode at a cycle length of 300 ms. The sinus node was clamped by a hemostat to control the atrial rate. Surface electrocardiographic leads II and V5R were continuously monitored. Ventricular pacing was performed from the innermost pole of each of the 16 pole needles placed in three locations (apex, septum, and lateral ventricular wall) outside the risk zone of coronary artery occlusion. Cathodal stimuli (2 ms in duration at 4 times diastolic threshold) were applied to the pacing electrode while the anode was located in abdominal subcutaneous tissue. Twenty multipolar plunge needles were inserted into the risk zone of the left anterior descending coronary artery occlusion with spacing ∼10 mm apart (Fig. 1). Each transmural needle recorded three bipolar electrograms from circumferential transmural electrodes, oriented perpendicular to the epicardial surface. The electrograms were sampled at 1 kHz, filtered at 3-300 Hz, and recorded on a system with a 286 computer with a 30-megabyte hard disk, 40 K base memory, and 80287 math processor (Bard, Billerica, MA, U.S.A.). Presampling allowed electrogram acquisition ≤4 s before an event. Three-dimensional maps were constructed from the transmural signals.
The electrogram that was shown to be from the site of earliest activity was labeled electrogram-F (focus of origin). Other electrograms were labeled east, west, north, and south with regard to position of the recording electrode relative to the focus of origin. Thus north indicated that the electrogram was recorded from an electrode toward the base of the heart relative to the focus, east from an electrode lateral to the focus, west from an electrode medial to the focus and south from an electrode located toward the apex of the heart relative to the focus. The letter "O" was used to indicate electrograms overlying the focus of origin. This format of labeling is used for the electrogram figures.
Ventricular effective refractory period was determined by delivering extrastimuli after eight paced complexes with effective refractory period defined as the longest interval between the drive pacing (S1) and the first extrastimulus (S2) that did not capture the ventricle. Programed ventricular stimulation consisted of a drive of eight regular stimuli followed by up to four premature stimuli. The following induction protocol was used in each experiment. After the ventricular effective refractory period was measured, the first premature stimulus (S2) was fixed at 4 ms longer, and a second stimulus (S3) was used at the same coupling interval. The S3 was shortened in 10-ms decrements until either ventricular tachycardia induction or failure to capture occurred. If no ventricular tachycardia was induced, the same procedure was followed for the third (S4) and fourth extrastimuli (S5). There was a pause of 1 s before the next drive started.
Ventricular tachycardia was defined as a wide QRS rhythm with atrioventricular dissociation that lasted longer than three beats. Ventricular tachycardia was considered to be sustained if it lasted longer than 30 s, or either pace termination or cardioversion was required because of hemodynamic collapse.
Ventricular tachycardia was designated to have a focal origin when the earliest electrical activity was recorded on one electrogram, proceeding sequentially to all surrounding electrograms, and when no electrical activity was recorded on all adjacent sites in three dimensions between the latest activation of one QRS complex and the earliest of the next QRS. Moreover, adjacent electrodes could not manifest conduction delay accounting for the majority of cycle length of the ventricular tachycardia. Focal origin thus implies that the ventricular tachycardia has a single site of origin with activation spreading out from that site in all directions.
A mechanism was defined as reentrant when the earliest activation site was located immediately adjacent to the site of the latest activation from the previous complex, and continuous diastolic activation was recorded between complexes. A reentrant mechanism also demonstrated unidirectional and functional block to the subsequently earliest site of activation, which usually occurred during the extrastimuli. Reentry, therefore, differs from a focal origin in that it tends to follow a certain pathway of electrical excitation involving multiple electrodes, which returns to the site of earliest activation.
Ischemia was defined as a reduction in voltage of the electrograms of >55% from baseline in the epicardium and midmyocardium and ≥45% for endocardium (7).
Experimental protocol for in vivo studies
A total of 16 dogs was studied. After instrumentation of the risk zone with 20 multipolar plunge needles and before coronary artery occlusion, induction of ventricular tachycardia was attempted to exclude animals with ventricular tachycardia due to electrode instrumentation alone. The left anterior descending coronary artery was then occluded, and ventricular tachycardia induction attempted by using serial induction protocols from 1 to 3 h after occlusion. No spontaneous ventricular tachycardia or ventricular fibrillation was observed during the period of 1-3 h after coronary artery occlusion. When ventricular tachycardia was induced with pacing from one of the endocardial electrodes, repeated induction was attempted from that same site to ensure reproducibility before giving civamide. If the dogs had morphologically similar inducible ventricular tachycardia on at least two consecutive attempts, civamide at a total dose of 50 μg/kg was given. The initial dose was usually 20 μg/kg, and this was followed by 30 μg/kg, both given as an intravenous infusion over ∼20 min. The total dose of 50 μg/kg was the lowest that had significant effects on the ventricular muscle refractory period. The induction protocol was then repeated from the same site to assess the effect of the drug.
Purkinje fibers were stimulated at a cycle lengths of 700 ms, with bipolar electrode placed directly on the Purkinje fiber, with the use of square-wave pulses of 2-ms duration at 2 times diastolic threshold. Fibers were impaled with 3 M KCl-filled glass microelectrodes with tip resistances of 15-40 MΩ. The microelectrodes were connected to a high-input impedance preamplifier having capacity neutralization (Axoclamp 2A; Axon Instruments, Foster City, CA, U.S.A.), and a tissue bath was grounded by a 3 M KCl-agar bridge. Transmembrane potentials were displayed on an oscilloscope, acquired and stored on an 80386-based computer (Gateway 2000, North Sioux City, SD, U.S.A.) with use of pCLAMP (Axon Instruments) software sampling data at 2 kHz. The data were analyzed by using pCLAMP (Foster City, CA, U.S.A.) software. Ten consecutive beats were digitally averaged. The resting membrane potential (RMP), action-potential amplitude (APA), and the APDs at 50 and 90% of repolarization were measured by using cursors in clampan of pCLAMP. These measurements were accurate to within 0.1 mV or 0.5 ms. Baseline parameters were obtained after ∼60-min equilibration with Tyrode's buffer.
Experimental protocols for in vitro studies
Dose-response curves were obtained for civamide at doses from 10−8 to 10−5M. Purkinje fibers were exposed to various concentrations of civamide for 10 min and ADPs recorded after 5 and 10 min of exposure to each concentration. After exposure to drugs, Purkinje fibers were perfused again with Tyrode's solution for 30-60 min until at least partial reversal of drug effects was seen. Studies also were performed in which Purkinje fibers were preexposed to nifedipine at a dose of 10−6M. This was done to assess whether the effect of civamide could be prevented by blockade of calcium channels. After the Purkinje fibers had been exposed to nifedipine (10−6M) for 10 min, civamide (10−5M) was added to the solution containing nifedipine. Measurements were obtained after superfusion for 10 min. A dose of 10−6M was chosen for nifedipine, as this has been shown to significantly shorten the APD50 and APD90 of isolated Purkinje fibers previously in this laboratory (8).
Civamide, in either ethanol vehicle or 0.1% dimethyl sulfoxide (DMSO), was supplied by GenDerm Corporation (Lincolnshire, IL, U.S.A.). Nifedipine was purchased from Sigma Chemical Co., St. Louis, MO, U.S.A.).
All data are expressed as mean ± SEM. Statistical analysis was performed by analysis of variance with Dunnett's multiple-comparison test. Student's t test was used when appropriate. Differences with a p value of <0.05 were considered significant.
Characteristics and mechanisms of ventricular tachycardias induced
Ventricular tachycardia was not inducible before coronary artery occlusion in any of the 16 dogs. After 1-3 h of coronary artery occlusion, ventricular tachycardia was inducible in 13 of 16 dogs. In all cases, the tachycardia originated from within the ischemic zone, as defined by a decrease in voltage of electrograms as described earlier. Nine of the ventricular tachycardias were mapped to have a focal endocardial origin (Figs. 2 and 3). Four other dogs with inducible ventricular tachycardia had an epicardial reentrant mechanism. The mean cycle length of the tachycardias of focal endocardial origin was 118 ± 11 ms, and of reentrant tachycardias, 138 ± 9 ms.
Effects of civamide in vivo
Civamide did not alter activation times in response to drive pacing (S1) in nonischemic myocardium [25 ± 2 to 26 ± 2 ms (p = NS)] or ischemic zones [37 ± 3 to 38 ± 2 ms (p = NS)]. The transmural activation times in response to the first extrastimuli (S2) were unaltered in both normal zones [32 ± 3 to 29 ± 3 ms (p = NS)] and ischemic zones [48 ± 3 to 46 ± 3 ms (p = NS)]. However, civamide shortened the refractory period in normal myocardium from 138 ± 3 to 132 ± 4 ms (p < 0.05). The mean arterial pressure also was decreased by civamide from 78 ± 7 to 66 ± 7 mm Hg (p < 0.05). Review of each electrode voltage before and after administration of civamide did not reveal consistent changes that suggested antiischemic or ischemic effects of the drug. In the three dogs with focal endocardial ventricular tachycardia not given civamide, the ventricular tachycardia was repeatedly inducible. In these dogs, both mean arterial pressure (91.7 ± 20 to 90.0 ± 22 mm Hg; p = NS) and muscle refractory periods (146.7 ± 6 to 146.3 ± 10 ms; p = NS) were stable throughout the study period.
Civamide, at a total dose of 50 μg/kg, altered previously inducible ventricular tachycardia of focal endocardial origin in five of the six dogs (Table 1). In dogs 2 and 6, there was no inducible ventricular tachycardia after civamide, and in dog 3, there was only nonsustained ventricular tachycardia of focal endocardial origin inducible after civamide, whereas ventricular tachycardia was sustained before the drug. In dog 4, only ventricular tachycardia with an epicardial reentrant mechanism was inducible after civamide; the faster focal endocardial tachycardia could not be induced after drug. In the dog 1, repeated induction revealed a ventricular tachycardia of multifocal origin with a cycle length of 60 ms. In dog 5, civamide slowed the cycle length of the same ventricular tachycardia by 10 ms. In three of four dogs with inducible VT of epicardial reentrant mechanism, civamide neither prevented induction of ventricular tachycardia nor changed the mechanism or cycle length (Table 1). However, reentrant ventricular tachycardia was not inducible in one dog after administration of the drug.
Effects of civamide in vitro
The dose-response relation of civamide on Purkinje APD is shown in Table 2. Civamide (n = 10) at doses of 10−5M shortened the APD50 in isolated Purkinje fibers from 193 ± 13 to 177 ± 12 ms (p < 0.01) and APD90 from 260 ± 15 to 248 ± 13 ms (p < 0.01) (Fig. 4). Civamide had no effect on the APA or RMP (Table 2). Solvent vehicle alone (ethanol or DMSO at 10−5M) had no significant effect on the Purkinje APD, APA, or RMP. Because civamide shortened Purkinje APD but had no effect on APA and RMP, we suspected that these might be mediated through inhibition of calcium channels. Studies were performed examining the effects of civamide on Purkinje action potential after pretreatment with nifedipine at 10−6M (n = 5). Nifedipine alone shortened Purkinje APD50 from 181 ± 4 to 145 ± 5 ms (p < 0.05) and the APD90 from 272 ± 7 to 254 ± 9 ms (p < 0.05). Addition of civamide (10−5M) did not result in any further shortening of the APD50 (to 149 ± 6 ms, p = NS vs. nifedipine), and APD90 (to 259 ± 4 ms, p = NS vs. nifedipine).
The main finding of this study is that civamide alters inducibility of ventricular tachycardia with a focal endocardial origin but not epicardial reentrant mechanism in intact dogs, whereas the inducibility of focal endocardial ventricular tachycardia in time controls not given civamide was unaltered. Additionally, civamide shortens the APD of isolated superfused Purkinje fibers in vitro. The effects of civamide in vitro were not seen in preparations that were pretreated with the L-type calcium channel-blocker nifedipine.
Effects of civamide on inducible ventricular tachycardia
In this model of inducible ventricular tachycardia 1-3 h after coronary artery occlusion, arrhythmias with a focal origin occur in up to three fourths of experiments (9). The underlying mechanism of focal ventricular tachycardia is not clear. The term focal implies that there was no evidence of block or extreme conduction delay in the immediate vicinity to the focus of origin (10).
Previous studies with this model demonstrated that ventricular tachycardia of focal origin (including focal endocardial origin) is reproducibly inducible over 1-3 h and can be prevented by agents with calcium-blocking effects. Doxorubicin, which blocks both calcium entry and calcium release from the sarcoplasmic reticulum (11), prevents inducible ventricular tachycardia with focal origin in this model (12), as does flunarizine, a calcium entry blocker (J.B. Martins, unpublished observations). Others demonstrated an effect of flunarizine on ventricular tachycardia with a triggered mechanism due to digitalis toxicity but not on ventricular tachycardia with a reentrant mechanism (13,14). Thus these previous results are consistent with triggered activity as a mechanism of focal arrhythmias in this model. In addition, dogs with ventricular fibrillation produced by simultaneous anterior descending and circumflex coronary artery ligation had ventricular fibrillation latency more than doubled or fibrillation blocked by either calcium-entry blockade with diltiazem or reduced serum ionized calcium (15). Therefore, the in vivo results with civamide are consistent with effects of other agents that block calcium entry.
Civamide did produce a decrease in mean arterial pressure of a mean of 10 mm Hg. These blood pressure changes are relatively small, and in previous studies with this model, a decrease in mean arterial pressure of the magnitude of 40-50 mm Hg was required to affect transmural activation times and the rate of reentrant ventricular tachycardia in the ischemic and hypertrophied ventricle (16). Inducibility of ventricular tachycardia, on the other hand, was not affected. It is therefore unlikely that the relatively minor changes in blood pressure affected the inducibility of the ventricular tachycardia in this study. However, the effects of civamide on the blood pressure are also consistent with the effect of L-type calcium channel blockers.
Effects of civamide on Purkinje fibers in vitro
Based on our in vivo observations in the intact dog, we sought further evidence for a calcium channel-blocking effect of civamide. Civamide (10−5M) significantly shortened the APD50 and APD90 in isolated superfused Purkinje fibers. These changes are consistent with the decrease of ventricular muscle refractory periods seen in the intact dog studies. Civamide affected mainly duration of phase 2 of the action potential, in which opening of the L-type calcium channels occurs (17). No effects on the APA or RMP were observed. The lack of effect on the APA and RMP suggests that the voltage-sensitive sodium channel and the inward rectifying potassium channel (IK1) are probably not involved. Previous reports produced conflicting results on the effects of racemic capsaicin on ventricular myocyte APD. Castle (4) found that capsaicin (10−5M) prolonged rat myocyte APD through inhibition of the potassium channels Ito, IK, and IK1, effects opposite our findings. D'Alonzo (5), in contrast, showed that capsaicin (10−4, 3 × 10−5, and 10−5M) shortened papillary muscle APD in guinea pigs. It is possible that species difference may be responsible at least in part for these discrepant results.
Effects of civamide in vitro in the presence of nifedipine
In an attempt to evaluate further whether civamide might exert a calcium entry-blocking effect, experiments were performed on isolated Purkinje fibers preexposed to the L-type calcium channel blocker, nifedipine. Nifedipine alone shortened the Purkinje APD, and addition of civamide did not produce any further shortening. This, along with effects of civamide on ventricular muscle in vivo and the Purkinje fibers in vitro, is suggestive of civamide acting through blockade of L-type calcium channels. However, confirmation that the mechanism of action of civamide is the result of calcium channel blockade awaits studies of single cells by patch-clamp/voltage-clamp techniques.
Although significant conduction delay was not noted in the vicinity of focus of origin of the ventricular tachycardia on the activation maps, it cannot be ruled out that microreentry could be a possible mechanism for the focal endocardial ventricular tachycardia seen in this study. This is a possible limitation for all mapping studies using intramural needles, as microreentry has been shown to occur in as small a volume of tissue as 0.2-0.5 cc (18,19).
This study and others (10,11) showed that the occurrence of inducible ventricular tachycardias with focal origin is not uncommon in the early period of ischemia/infarction. Civamide alters ventricular tachycardia of focal endocardial origin and might potentially be useful as an antiarrhythmic agents during acute ischemia, when arrhythmias with focal origin may be seen concurrent with reentrant mechanisms, as was demonstrated in experimental animal models. Inducible ventricular tachycardias with focal origin have been shown to occur in humans with scars caused by prior myocardial infarction (20), although it is unknown whether these occur in the early stages of ischemia/infarction because of the inability to perform activation mapping studies in humans under these conditions. It would be useful to find a marker for focal endocardial ventricular tachycardia because such might be amenable to ablative techniques on the endocardium. Further studies are needed to clarify whether a potential clinical role will emerge for civamide in the therapy of ischemic ventricular arrhythmias.
Acknowledgment: This study was supported in part by a grant from GenDerm Inc., Lincolnshire, Illinois. D.O. Arnar was supported by a Fellowship Award from the American Heart Association, Iowa Affiliate. We thank Linda Bang for expert secretarial assistance.
1. Fitzgerald M. Capsaicin and sensory neurons: a review. Pain
2. Holzer P. Capsaicin: cellular targets, mechanism of action and selectivity for thin sensory neurons. Pharmacol Rev
3. Dray A. Mechanism of action of capsaicin like molecules on sensory neurons. Life Sci
4. Castle NA. Differential inhibition of potassium currents in rat ventricular myocytes by capsaicin. Cardiovasc Res
5. D'Alonzo AJ, Grover GJ, Darbenzio RB, et al. In vitro effects of capsaicin: antiarrhythmic and antiischemic activity. Eur J Pharmacol
6. Bourne N, Bravo FJ, Bernslein DI, Myers MG, Stanberry LR. Efficacy of capsaicin and its less irritating isomer, civamide in the treatment of experimental genital herpes [Abstract]. Antiviral Res
7. Ruffy R, Lovelace D, Mueller TM, Knoebel SB, Zipes DP. Relationship between changes in left ventricular bipolar electrograms and regional myocardial blood flow during acute coronary occlusion in the dog. Circ Res
8. Lee H, Cai JJ, Arnar DO, Shibata EF, Martins JB. Mechanisms of alpha-2 adrenergic modulation of canine Purkinje action potential. J Pharmacol Exp Ther
9. Van Why K, Dexter EU, Rath T, Ben-Haim S, Iguchi A, Martins JB. Doxorubicin prevents inducibility of ventricular tachycardia with focal origin in dogs early after coronary artery occlusion [Abstract]. PACE Pacing Clin Electrophysiol
10. Pogwizd SM, Corr PB. Reentrant and nonreentrant mechanisms contribute to arrhythmogenesis during early myocardial ischemia: results using three-dimensional mapping. Circ Res
11. Caroni P, Villain F, Carofoli E. The cardiotoxic antibiotic doxorubicin inhibits the Na+
exchange of dog heart sarcolemmal vehicles. FEBS Lett
12. Van Why KJ, Dexter EU, Ben-Haim SA, Iguchi A, Rath TE, Martins JB. Polymorphic ventricular tachycardia inducible 1 to 3 hours after coronary artery occlusion in the dog has a focal mechanism [Abstract]. Clin Res
13. Vos MA, Gorgels APM, Leunissen JDM, van der Nagel T, Halbertsma FJJ, Wellens HJJ. Flunarizine allows differentiation of mechanisms of arrhythmias in the intact heart. Circulation
14. Vos MA, Gorgels APM, Leunissen JDM, van der Nagel T, Halbertsma FJJ, Wellens HJJ. Further observations to confirm the arrhythmia mechanism: specific effects of flunarizine. J Cardiovasc Pharmacol
15. Clusin WT, Bristow MR, Baim DS, et al. The effects of diltiazem and reduced serum ionized calcium on ischemic ventricular fibrillation in the dog. Circ Res
16. Hopson JR, Martins JB. Hemodynamic and reflex control of transmural activation and rate of ventricular tachycardia in ischemic and hypertrophic ventricular myocardium of the dog. Circulation
17. Task force of the Working Group on Arrhythmias of the European Society of Cardiology. The Sicilian gambit: a new approach to the classification of antiarrhythmic drugs based on their actions on arrhythmogenic mechanisms. Circulation
18. Antzelevitch C, Jalife J, Moe GK. Characteristics of reflection as a mechanism of reentrant arrhythmias and its relationship to parasystole. Circulation
19. Pogwizd SM, Corr PB. Reentrant and nonreentrant mechanisms contribute to arrhythmogenesis during early myocardial ischemia: results using three-dimensional mapping. Circ Res
20. Pogwizd SM, Hoyt RM, Saffitz JE, Corr PB, Cox JL, Cain ME. Reentrant and focal mechanisms underlying ventricular tachycardia in the human heart. Circulation