Role of the Purkinje System in Spontaneous Ventricular Tachycardia During Acute Ischemia in a Canine Model
Background A role for the Purkinje system in the development of spontaneous ventricular tachycardia (VT) during acute ischemia has been suspected but not proved. We used a three-dimensional activation mapping system incorporating Purkinje signals to characterize the mechanism and site of origin of spontaneous VT occurring in the first 30 minutes after coronary artery occlusion in a dog model.
Methods and Results The left anterior descending coronary artery was occluded in 48 dogs after instrumentation of the risk zone with 21 multipolar plunge needles, each recording 6 bipolar electrograms through the myocardial wall. VT of Purkinje origin was defined as a focal endocardial VT with a Purkinje potential identified before muscle potential on the electrode recording the earliest activity. Purkinje potentials were identified on an average of 10 of the 21 plunge needles. During atrial pacing at cycle lengths of 300 to 700 ms, a total of 25 VTs were observed from 18 of the 48 dogs (37.5%). Of the VTs, 15 (60.0%) were of focal Purkinje origin, 1 (4.0%) of focal endocardial origin, 2 (8.0%) of focal midmyocardial origin, and 2 (8.0%) of focal epicardial origin; 3 (12.0%) had a reentrant mechanism, whereas in 2 (8.0%), the mechanism could not be defined. The mean cycle length of all VTs was 265±17 ms (mean±SEM, n=25). Of the 25 VTs, 19 originated from an ischemic area as defined by significant decreases in voltages of muscle electrograms at the time of occurrence of the VT, 4 originated from an ischemic border zone, and the origin of 2 could not be determined.
Conclusions In this model, VT with a focal mechanism is commonly seen in the early ischemic period. Sixty percent of the VTs were of focal Purkinje origin as characterized by three-dimensional activation mapping. The results of this study indicate that Purkinje tissue may play an important role in the development of early ischemic VT.
Spontaneous VT is common in the acute phase of myocardial ischemia. This VT is thought to be reentrant because of electrograms demonstrating delay and fragmentation, suggesting slow conduction1 2 3 and continuous electrical activity between sinus complexes and the first complex of VT.4 5 6 However, transmural activation mapping studies have indicated that focal mechanisms may also be the cause of VT during this period.4 5 7 8 The mechanisms underlying VT of focal origin are speculated to include triggered activity and abnormal automaticity.
Purkinje potentials have been shown to precede ventricular muscle activity on intramural electrograms during premature ventricular complexes during early ischemia, suggesting a Purkinje origin.7 9 In this study, a custom-built mapping system that incorporates Purkinje potentials was used to characterize the mechanism of spontaneous VT in the first 30 minutes after coronary artery occlusion in a dog model. The purpose was to evaluate the role of the Purkinje system in the development of this ischemic VT.
Healthy adult mongrel dogs of either sex, weighing 18 to 24 kg, were studied. The protocol was 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 to 200 mg/kg IV bolus). Anesthesia was maintained with a continuous infusion of α-chloralose dissolved in polyethylene glycol at 8 mg · kg−1 · h−1 IV. The animals were intubated orotracheally and ventilated on a volume-controlled ventilator (Harvard Apparatus). Tidal volume was adjusted to achieve an arterial Pco2 within physiological range (25 to 35 mm Hg), and positive end-expiratory pressure was applied to maintain Po2 in normal range (80 to 150 mm Hg) with oxygen-enriched air. NaHCO3 was infused as necessary to maintain the pH within physiological range (7.30 to 7.45). The serum electrolytes K+ (3.6 to 5.0 mEq/L), Mg2+ (1.5 to 3.0 mg/dL), and Ca2+ (8.5 to 10.5 mg/dL) were periodically measured and were always within normal limits. Arterial pressure was continuously monitored with a Statham P23dB transducer with a fluid-filled polyethylene catheter placed in the aortic arch via the femoral artery. The femoral vein was cannulated for infusion of drugs and saline.
The heart was exposed through a median sternotomy and the pericardium incised and sutured to the wound edges to support the heart. The LAD was dissected proximally, and a snare was placed around the vessel immediately distal to the first septal perforator. The sternal wound was covered by a plastic sheet, and an infrared heating lamp was directed to the heart to maintain intrathoracic temperature at 37°C. After the experiment, the dogs were euthanized by induction of VF.
A bipolar electrode was used to pace the right atrium at two times diastolic threshold with pulses of 2-ms duration and cycle lengths from 300 to 700 ms. Each dog was paced at only one cycle length. The region of the sinus node was clamped permanently to control the rate. Surface ECG leads I, II, III, aVR, aVL, aVF, and V5R were recorded. Twenty-one multipolar plunge needles were inserted into and surrounding the risk zone of the LAD occlusion (Fig 1⇓). Needles were inserted into the myocardium perpendicular to the epicardial surface, except electrodes 1 through 5, which were inserted diagonally into the septum and electrodes 19, 20, and 21, which were inserted at an angle to involve the septum in addition to the anterior wall of the right ventricle. Spacing between needles was 10 mm, although this could vary by a few millimeters between experiments to adjust for differences in coronary anatomy.
Electrograms were recorded simultaneously on two computers, one for the three most endocardial bipoles and the other for the three most epicardial. The three endocardial electrodes were recorded on a custom-built system consisting of a 33-MHz 80486 computer (Gateway 2000) coupled with a DAP 2400/6 (Microstar Laboratories) high-speed acquisition board, an analog signal multiplexer, and 64 independent amplifier circuits. Signals were amplified by a gain of 100, band-pass–filtered between 3 and 1300 Hz, and digitized with a 12-bit AD converter at 3.2 kHz. The epicardial electrograms were recorded with a commercial system (Bard Electrophysiology) on a Compaq Deskpro 286 computer with a 40-kilobase memory and a 80287 math processor. The latter recorded 64 channels at 12-bit resolution with a sampling frequency of 1 kHz/channel and band-pass filtering from 30 to 300 Hz. Presampling of the data allowed for acquisition of electrograms for 4 seconds before an event with both systems. 3D activation maps were constructed from multiplexed signals of up to 14 seconds with data from both acquisition systems. A common surface ECG (lead V5R) recording pacing spikes allowed alignment of signals from both computers.
Six bipolar electrograms were recorded from each needle selected from 16 electrodes of insulated tungsten wire separated by 1 mm and wound about the circumference of the needle. Optimal, noise-free bipolar electrograms (usually spaced at 1 mm) were chosen to maximize the capability of Purkinje recording. This adjustment was performed by sequential recordings on a storage oscilloscope for each potential bipole. A switching box was used to connect the selected bipoles to each amplifier. The most endocardial bipole was used to record Purkinje potentials when they could be identified according to previously published criteria from this laboratory, including 0.5-mV spikes lasting 1 to 2 ms preceding the larger and longer muscle spike and the surface QRS on the lead recording the earliest activity.10 The most epicardial bipole recorded an electrogram from the epicardium, and intermediate bipoles recorded electrograms through the myocardial wall.
Fig 2⇓ shows a diagram of the activation maps, which were constructed by placing all epicardial activation times in one plane, all subepicardial maps in the next plane, and so on through the ventricular wall. Activation times were chosen in the following manner: A computer algorithm searched for the most rapid electrogram dV/dt over an interval of 20 ms before the surface QRS recording the earliest activity to 20 ms before the onset of the next surface QRS. This computer-chosen site of the most rapid dV/dt was displayed superimposed on the electrogram for verification or correction. This correction was aided by high-gain analysis (up to a maximum of 2 μV/cm). The Purkinje activations were manually assigned, and the layer reflected only Purkinje activation even if muscle activation was observed on the electrogram. All activation times were referenced to the earliest surface QRS. Twenty-millisecond isochrones were hand-drawn in each plane.
VT was defined as three or more abnormal QRS complexes in a sequence. VT was considered to be sustained if it lasted longer than 30 seconds or either pace termination or cardioversion was required because of hemodynamic collapse. The cycle length of VT was averaged over the first three complexes. Monomorphic VT was characterized by uniform QRS complexes with a reproducible axis and bundle-branch block morphology. Polymorphic VT complexes changed axis or bundle-branch morphology on a complex-to-complex basis.
VT was designated to have a focal origin when no electrical activity could be recorded on all adjacent sites in 3D between the latest activation of one QRS complex and the earliest of the next QRS. Moreover, conduction from the site of earliest activity to adjacent electrodes could not manifest conduction delay, which might account for a majority of the cycle length of the VT.
Purkinje origin of VT was defined as a focal endocardial mechanism with recording of a Purkinje potential before the QRS on the endocardial electrogram recording the earliest activity. Purkinje potentials had to be identified on electrograms during atrial pacing before and after coronary occlusion and during VT to be considered mechanistically involved.
If the electrode recording the earliest activity for a VT complex was not surrounded by electrodes recording subsequent activation, the possibility could not be excluded that either a focal mechanism or reentry from a remote site outside the electrode array could have been responsible for initiating the complex. In such cases, the mechanism could not be defined.
Mechanisms were 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. Reentrant mechanisms also demonstrated unidirectional and functional block to the subsequent earliest site of activation.
Ischemia was defined as a reduction in voltage of electrograms of >45% from baseline (before coronary artery occlusion) for the endocardium or Purkinje layer and more than 55% reduction from baseline for the midmyocardium and epicardium.11
After instrumentation of the myocardium with the multipolar plunge electrodes, dogs were observed for spontaneous VT for 45 to 60 minutes before coronary artery occlusion to exclude VT occurring from mechanical artifact. The LAD was then occluded by tightening of the snare. All spontaneous VTs occurring in the first 30 minutes after occlusion were recorded and stored on computer.
Data are expressed as mean±SEM. Student’s t test was used for comparison between groups. A value of P<.05 was considered significant.
Of the 48 dogs studied, spontaneous VT was observed in 18 (37.5%), and these 18 dogs were subjected to a more detailed analysis. There was no difference in the size of the ischemic zone, expressed as a percentage of electrograms demonstrating a muscle voltage decrease consistent with ischemia, between those dogs that had spontaneous VT and those that did not.
Purkinje Potential Recordings
High-frequency, low-amplitude Purkinje potentials were identified on an average of 10 electrograms, with a range of 5 to 17.
Activation Sequence During Atrial Pacing Before Coronary Artery Occlusion
After instrumentation of the myocardium with plunge needles, no spontaneous VT was observed before coronary artery occlusion in any dog. The mean transmural activation time before coronary artery occlusion was 32.0±2 ms. Fig 3⇓ illustrates an example of the activation of an atrial paced complex (cycle length, 300 ms). The earliest site of activity in the left ventricle was in the Purkinje layer toward the base of the heart (−5 ms). The activity then spread out within the Purkinje layer to the endocardium and then transmurally to the epicardium, where the latest activity was seen (32 ms). The total activation time from Purkinje layer to epicardium for this complex was 37 ms.
Activation Sequence During Atrial Pacing After Coronary Artery Occlusion
The mean transmural activation time during atrial pacing after coronary artery occlusion was 52.6±4 ms (P<.0001 versus activation time before occlusion). Fig 4⇓ shows an activation map of the atrial paced complex (cycle length, 300 ms) 7 minutes after coronary artery occlusion in the same dog as shown in Fig 3⇑. The earliest electrical activity seen in this atrial paced complex was −6 ms in the Purkinje layer toward the base of the heart. From this site, the activity spread out similar to that described for the preocclusion atrial paced complex, although transmural activation time was now prolonged to 52 ms because of delay in the epicardium. No substantial conduction delay (range, 1 to 9 ms) was seen between Purkinje layer and muscle before or after occlusion.
Incidence of Spontaneous Arrhythmias
A total of 18 of 48 dogs studied had spontaneous episodes of VT. In these 18 dogs, 25 episodes of VT were seen. The mean cycle length of all VTs was 265±17 ms. Twenty-three episodes of VT were nonsustained, two episodes were sustained, and both degenerated into VF. Of the VTs recorded, 22 were monomorphic and 3 were polymorphic. Complete 3D activation mapping was performed on a total of 75 VT complexes. In 20 of 25 VTs, the first complexes were found to have a focal origin, whereas 3 had a reentrant mechanism (Table 1⇓). The VTs with a reentrant mechanism occurred at a mean time of 8.0±2 minutes (range, 5 to 10 minutes), whereas the focal VTs had a mean time of occurrence of 14.4±1 minutes (range, 7 to 30 minutes) (P=.07). Interestingly, 6 of the focal VTs (including 5 focal Purkinje VTs) occurred during the 7- to 10-minute period after coronary artery occlusion.
VT With Focal Purkinje Origin
Of the first VT complexes mapped, 60% were of focal Purkinje origin. An example of an episode of VT with focal Purkinje origin is shown in Fig 5⇓. An atrial paced complex is followed by 3 VT complexes. Surface lead II is shown at the top, and below are recordings from the Purkinje and endocardial layers. Electrogram E-F (endocardial-focus) was recorded from the site that was subsequently shown to be the site of origin of this VT. The other electrograms were recorded from the adjacent endocardium, with north (E-N) toward the base of the heart in relation to E-F and east (E-E), south (E-S), and west (E-W) located accordingly. Electrogram O-F was recorded from endocardium immediately overlying the Purkinje focus.
For the atrial paced complex, Purkinje potentials (small and large arrows) preceded muscle activity, appearing simultaneously before the surface QRS. In contrast, during the first VT complex, these Purkinje potentials were seen markedly earlier than the surface QRS, with the earliest recorded on E-F. There was a sequential delay from E-F to the adjacent Purkinje potentials.
Fig 6⇓ shows the activation map of the first VT complex as shown in Fig 5⇑. The earliest activity seen was again in the Purkinje layer (−14 ms) but located toward the apex in the ischemic zone. The activation spread out in all directions within the Purkinje layer and to the overlying endocardium without evidence of conduction block or delay. The activation then proceeded transmurally to the epicardium.
All three reentrant VTs had an epicardial circuit. Fractionated activity and double potentials were seen only on epicardial electrograms. Fig 7⇓ shows recordings of a VT with a reentrant mechanism. Both conduction delay and block were noted in the epicardium during atrial pacing preceding initiation of VT (Fig 8⇓). Figs 9⇓ and 10⇓ show slow activation in the epicardium proceeding in a double-loop pattern. In contrast to the atrial paced complex, the VT complex showed Purkinje activity late within the surface QRS.
Ischemia at the Site of Origin of VTs
Of the 15 Purkinje VTs, 12 had a ≥45% decrease in voltages of Purkinje potentials at the site of earliest activity as well as in the voltages of the immediately overlying muscle potential, indicating ischemia. The remaining three focal Purkinje VTs had neither Purkinje layer nor overlying muscle electrograms that decreased substantially. These three did, however, have immediately adjacent needles demonstrating “ischemic” voltage changes on muscle electrograms, so these three VTs could have originated from an ischemic border zone. Likewise, the one VT of focal endocardial origin did not show ischemic voltage changes at the site of origin but was adjacent to ischemic sites. The electrograms at the origin of other focal and reentrant VTs did demonstrate ischemia.
Maintenance of VTs
More than half of the VTs were maintained by the same focus and mechanism as the first complex was (Table 2⇓). This was the case in 10 of 15 focal Purkinje VTs and all 3 reentrant VTs. Others could be initiated at one focus and maintained by another. Initiation by one mechanism and maintenance by another was not seen, however, in the first three complexes of the VTs.
Two of the VTs of Purkinje origin degenerated to VF. In one of these, only the Purkinje system was involved in the maintenance of VT before VF occurred. In the other episode, the VT was initiated in the Purkinje layer but maintained by other intramural foci before degeneration to VF.
Effect of Atrial Pacing Cycle Length on Occurrence of VT
In an attempt to study whether the occurrence of VT was dependent on the basic atrial paced cycle length, the atrium was paced at different cycle lengths from 300 to 700 ms. Spontaneous VT did occur at all cycle lengths tested. Fig 11⇓ shows a direct relationship between the basic atrial paced cycle length and the average coupling interval from the last atrial paced complex to the first VT complex.
The main results of this study are that spontaneous nonsustained VTs of focal origin are common during acute ischemia in this canine model and that 60% of all VTs seen are of focal Purkinje origin. These results imply that Purkinje tissue might be importantly involved in the genesis of early ischemic ventricular arrhythmias.
Ischemic VTs of Purkinje Origin
Previous studies have suggested that Purkinje tissue may be the earliest site of activity during ectopy, based on Purkinje activity preceding muscle activity on extracellular electrograms from canine hearts during acute ischemia.7 9 Purkinje tissue is also the origin of late VT (>24-hour-old myocardial infarction).4 12 13 14 15 Because of the location of Purkinje fibers on the endocardial surface of the heart, it has been postulated that they may survive during transmural infarction by nourishment from cavitary blood.15 Abnormal electrophysiological properties develop in these fibers because of partial ischemia or regional ischemic metabolites. Indeed, microelectrode recordings from Purkinje fibers under these circumstances have shown reversible decreased amplitude of the action potential and slow upstroke.3 15
Recording of Purkinje activity in the present study was facilitated by the length of the transmural needles, the circumferential design of the electrodes (excluding directionality), the high frequency of sampling, and the wide–band-pass filtering. The electrograms during VT were carefully analyzed for the existence of Purkinje potentials, which had been observed during atrial pacing before and after coronary occlusion. We believe that the results confirm the speculation that Purkinje tissue can provide the origin of ischemic VT.
In the present study, a decrease in amplitude of ≥45% was noted at the site of earliest activity in both Purkinje and overlying muscle potentials of 12 of 15 VTs at the time of occurrence. It therefore appears that both Purkinje tissue and muscle became ischemic at the site from which VT originated. The other 3 Purkinje VTs originated from an area without electrograms demonstrating ischemic voltage changes. This was observed, however, on muscle potentials of immediately adjacent needles. Therefore, these VTs may have originated from an ischemic border zone.
Possible Mechanism of Focal VTs
The underlying mechanism of the focal VTs seen in this study is speculative. Our use of the term focal implies that no continuous electrical activity was recorded between subsequent beats and that activation maps did not demonstrate evidence of conduction delay, which could account for the majority of the cycle length. Conduction delay during ischemic conditions may be cycle length–dependent, so that as the rate increases, the conduction delay becomes greater.5 16 Therefore, it might be expected that the coupling interval to the first VT complex would increase with decreased cycle length, which is the opposite of what was observed for the focal VTs. This argues against a reentrant mechanism for the focal VTs seen in this study.
Triggered activity due to afterdepolarizations might be a potential mechanism of focal VTs in this study. Early afterdepolarizations are less likely to be a mechanism in this study, because their induction tends to be dependent on bradycardia.17 The likelihood of induction of VT by DADs may increase with shorter cycle lengths of basic drive. The majority of the experiments described in this study were conducted with a basic drive cycle length of 300 to 500 ms. With shorter cycle lengths, there tends to be an increase in the amplitude of DADs, which may reach threshold and initiate triggered activity.18 Some have considered DADs unlikely to occur during severe myocardial ischemia, because they are not seen in isolated superfused papillary muscle under conditions of hypoxia.19 Nevertheless, DADs might conceivably occur in endocardial Purkinje tissue, because oxygenation from cavity blood may serve to prevent severe hypoxia. The linear relationship between the drive cycle length and the coupling interval to the first VT complex in this study may support, although not prove, that triggered activity due to DADs may be a mechanism of focal VTs in this study.
Abnormal automaticity is another possible mechanism of focal VTs. Extracellular potassium, which becomes elevated early in ischemia, has been shown to suppress abnormal automaticity, and this has led to speculation that this mechanism may be less likely to occur in ischemic myocardium.20
Although substantial conduction delay or block was not seen surrounding a focus of origin of the focal VTs, the possibility of a small reentrant circuit being the mechanism cannot be excluded. Microreentry refers to a reentrant circuit involving a small number of adjacent myocytes.6 Microreentrant circuits have been shown to occur in as small a volume of tissue as 0.2 to 0.5 cm3,4 so it is possible for microreentry to occur in such models as used in this study.
Nonreentrant mechanisms involving injury currents flowing from ischemic muscle in the endocardium to nonischemic underlying Purkinje tissue might also conceivably be involved in the initiation of focal Purkinje VT.5 This may in particular be a possibility in those 3 VTs in which the earliest Purkinje signals did not demonstrate ischemia but adjacent muscle electrograms did.
It has been demonstrated that early ventricular arrhythmias may occur during two periods after coronary artery occlusion.21 These periods have been called Ia, for arrhythmias occurring between 2 and 10 minutes after coronary artery occlusion, and Ib, representing arrhythmias occurring at 12 to 30 minutes of acute ischemia. During the Ia period, extracellular electrograms show changes consistent with slow and inhomogeneous conduction in the epicardium. On the basis of conduction slowing, it has been proposed that VT in the Ia period may be caused by reentry.1 2 Our observation of reentrant VTs occurring only during the period of 5 to 10 minutes after coronary artery occlusion is consistent with this.
During the Ib period, there is at least partial recovery of the epicardial conduction delay, suggesting that mechanisms other than reentry might be responsible for Ib arrhythmias or that reentrant pathways located deep in the myocardium or involving the Purkinje-muscle junction may be involved. Recently published data have suggested that Ib arrhythmias may be influenced by the loss of cell-to-cell electrical coupling.22 The focal VTs seen in this study occurred during the range of 7 to 30 minutes, with 5 of the focal Purkinje VTs and a focal midwall VT occurring during the Ia period. These data, therefore, suggest that focal VT, especially focal Purkinje VT, can also occur during the Ia period.
Incidence of Spontaneous Arrhythmias During Acute Ischemia
The incidence of spontaneous VT occurring in the first 30 minutes after coronary artery occlusion was 37.5% in this study. Ventricular ectopy, however, was observed in virtually every experiment but was not subjected to analysis. Only 2 of the 48 dogs developed VF. The canine heart has an extensive coronary collateral network and may resemble that of humans with chronic coronary artery disease. The coronary arteries of pigs and cats have a less-well-developed collateral circulation and as such are more similar to the normal human coronary circulation. The incidence of VT and VF after coronary artery occlusion may be greater in the pig and cat than the dog because of a resultant increase in the ischemia/infarct zone.23 Likewise, in the dog, occlusion of the left circumflex artery may have greater arrhythmogenic effects than ligation of the LAD, because the circumflex artery supplies a greater area of the left ventricle. It is therefore likely that the choice of both the dog model and occlusion of the LAD influenced the incidence of spontaneous arrhythmias observed in this study.
We did not insert electrodes and record activity from the whole heart. Instead, the focus was on the area at risk from the LAD occlusion. In addition to instrumentation of the risk zone, electrodes were also inserted in the adjacent risk zone in an attempt to surround the ischemic zone and thus determine that the VTs recorded truly originated from within the area at risk. Nevertheless, there were instances in which the earliest activity occurred in an electrode that was not surrounded by electrodes recording subsequent activation. It is possible that these instances might represent a reentrant or focal mechanism arising at a remote site outside the area covered by the electrode array. When this occurred, that particular VT complex was labeled as “mechanism not defined.”
These data, from a 3D activation mapping system to demonstrate the earliest focus of activity during arrhythmias, show that the Purkinje system can frequently be an origin of spontaneous VT occurring early after coronary artery occlusion in the dog. These VTs appear to have a focal origin, although the underlying mechanism remains speculative. It is possible that the role of the Purkinje system in ischemic ventricular arrhythmias may have previously been underestimated.
Selected Abbreviations and Acronyms
|LAD||=||left anterior descending coronary artery|
This work was supported by a Grant-in-Aid from the American Heart Association, Iowa Affiliate. Dr Arnar was supported by a Fellowship Award from the American Heart Association, Iowa Affiliate. The authors would like to thank Dr Hon-Chi Lee for his helpful review of the manuscript, Dr Shlomo A. Ben-Haim for technical assistance, and Linda Bang for expert secretarial assistance.
- Received December 19, 1996.
- Revision received May 19, 1997.
- Accepted May 28, 1997.
- Copyright © 1997 by American Heart Association
Waldo AL, Kaiser GA. A study of ventricular arrhythmias associated with acute myocardial infarction in the canine heart. Circulation. 1973;47:1222-1228.
Boineau JP, Cox JL. Slow ventricular activation in myocardial infarction: a source of reentrant premature ventricular contraction. Circulation. 1973;48:702-713.
Scherlag BJ, El-Sherif N, Hope R, Lazzara R. Characterization and localization of ventricular arrhythmias resulting from myocardial ischemia and infarction. Circ Res. 1974;35:372-383.
Pogwizd SM, Corr PB. Reentrant and nonreentrant mechanisms contribute to arrhythmogenesis during early myocardial ischemia: results using three-dimensional mapping. Circ Res. 1987;61:352-371.
Janse MJ, Van Capelle FJL, Morsink H, Kleber AG, Wilms-Schopman F, Cardinal R, d’Alnoncourt CN, Durrer D. Flow of ‘injury’ current and patterns of excitation during early ventricular arrhythmias in acute regional myocardial ischemia in isolated porcine and canine hearts: evidence for two different arrhythmogenic mechanisms. Circ Res. 1980;47:151-165.
Pogwizd SM, Corr PB. Mechanisms underlying the development of ventricular fibrillation during early myocardial ischemia. Circ Res. 1990;66:672-695.
Wu G, Littmann L, Svenson RH, Nanney EA, Tatsis GP, Tuntelder JR, Chuang CH, Thompson M, Dezern KR. Computerized three-dimensional activation mapping study of spontaneous ventricular arrhythmias during acute myocardial ischemia in dogs: evidence against macroreentrant mechanisms. J Electrocardiol. 1995;28:155-130.
Cable DG, Rath TE, Dreyer ER, Martins JB. Refractory period responses of cardiac Purkinje tissue to α-1 and α-2-adrenergic influences. Am J Physiol. 1994;267(Heart Circ Physiol 36):H376-H382.
Ruffy R, Lovelace D, Mueller TM, Knoebel SB, Zipes DP. Relationship between changes in the left ventricular bipolar electrograms and regional myocardial blood flow during acute coronary occlusion in the dog. Circ Res. 1979;45:764-770.
Horowitz LN, Spear JF, Moore EN. Subendocardial origin of ventricular arrhythmias in 24-hour-old experimental myocardial infarction. Circulation. 1976;53:56-63.
El-Sherif N, Mehra R, Gough WB, Zeiler RM. Ventricular activation patterns of spontaneous and induced ventricular rhythms in canine one-day-old myocardial infarction: evidence for focal and reentrant mechanisms. Circ Res. 1982;51:152-166.
Friedman PL, Stewart JR, Fenoglio JJ Jr, Wit AL. Survival of subendocardial Purkinje fibers after extensive myocardial infarction in dogs: in vitro and in vivo correlations. Circ Res. 1973;33:597-611.
Friedman PL, Stewart JR, Wit AL. Spontaneous and induced cardiac arrhythmias in subendocardial Purkinje fibers surviving extensive myocardial infarction in dogs. Circ Res. 1973;33:612-626.
Hope RR, Williams DO, El-Sherif N, Lazzara R, Scherlag BJ. The efficacy of antiarrhythmic agents during acute myocardial ischemia and the role of heart rate. Circulation. 1974;50:507-514.
Damiano BP, Rosen MR. Effects of pacing on triggered activity induced by early afterdepolarizations. Circulation. 1984;69:1013-1025.
Le Marec H, Dangman KH, Danilo P Jr, Rosen MR. An evaluation of automaticity and triggered activity in the canine heart one to four days after myocardial infarction. Circulation. 1985;71:1224-1236.
Coetzee WA, Opie LM. Effects of components of ischemia and metabolic inhibition on delayed afterdepolarizations in guinea pig papillary muscle. Circ Res. 1987;61:157-165.
Hoffman BF, Rosen MR. Cellular mechanism for cardiac arrhythmias. Circ Res. 1981;49:1-15.
Kaplinsky E, Ogawa S, Balke W, Dreifus LS. Two periods of early ventricular arrhythmias in the canine acute myocardial infarction model. Circulation. 1979;60:397-403.
Smith WT, Fleet WF, Johnson TA, Engle CT, Cascio WE, for the Experimental Cardiology Group, University of North Carolina. The IIb phase of ventricular arrhythmias in ischemic in situ porcine heart is related to changes in cell-to-cell electrical coupling. Circulation. 1995;92:3051-3060.
Wit AL, Janse MJ. Experimental models of ventricular tachycardia and fibrillation caused by ischemia and infarction. Circulation. 1992;85(suppl I):I-32-I-42.