Shock-induced vulnerability and defibrillation have been mostly studied in structurally normal hearts. However, defibrillation therapy is normally applied to patients with diseased hearts, frequently those with prior myocardial infarction (MI). Shock-induced vulnerability and defibrillation have not been well studied under this condition. We sought to examine the mechanisms of shock-induced arrhythmogenesis and arrhythmia maintenance in a rabbit model of healed MI (4 wk or more postinfarction). Ligation of the lateral division or posterolateral division of the left coronary artery at a level of 40–70% from the apex was performed 53 ± 21 days before acute experiments. Shock-induced vulnerability was assessed in infarcted (n = 8) and structurally normal (n = 8) hearts by delivering internal monophasic shocks at different shock strengths and delivery phases. Electrical activities from the anterior epicardium during shock application and during shock-induced arrhythmias were optically recorded and quantitatively analyzed. Ligation resulted in a transmural left ventricular free wall infarction mainly located at the apical region with a consistent endocardial border zone (BZ) as confirmed by histological studies. There were significant increases in the incidence, severity, and duration of shock-induced arrhythmias in the infarcted hearts versus controls due to 1) postshock break-excitation wavefronts that frequently originated near the infarction BZ and 2) the existence of an infarction BZ that created an anatomic reentry pathway and facilitated arrhythmia maintenance. In conclusion, the infarction BZ contributes to both increased shock-induced arrhythmogenesis and arrhythmia maintenance in the rabbit model of healed MI.
- infarction border zone
- cardiac vulnerability
- optical imaging
arrhythmogenesis in postmyocardial infarction (MI) has been studied extensively in the canine epicardial border zone (BZ; surviving tissue neighboring the infarct) formed after occlusion of the left anterior descending artery (6, 35, 43, 47, 51, 58, 62). The epicardial BZ is characterized by decreased conduction velocity, increased anisotropy, and gap junction remodeling associated with electrical reentry and sustained ventricular arrhythmias (22, 23, 43, 57, 59). However, clinical observations suggest the presence of both endocardial/subendocardial BZ and epicardial BZ in healed human infarcts, contributing to recurrent sustained reentrant ventricular tachycardias (35, 58). We recently observed that an endocardial BZ could be created by ligation of divisions of the left coronary artery in the rabbit. Thus the rabbit healed MI is a useful animal model that closely approximates clinically observed conditions (39).
Electric shocks are commonly used to terminate life-threatening ventricular tachycardia/fibrillation in patients, yet the mechanisms of defibrillation are still not well understood. The recently formulated virtual electrode polarization (VEP) hypothesis of defibrillation correlates vulnerability/defibrillation with the transmembrane voltage (Vm) pattern of negative and positive polarization induced by an electric shock (21). It is based on numerous theoretical (27, 48, 49, 52) and experimental (11, 19, 20, 30, 37, 41, 54) studies. We (10) recently reported on shock-induced arrhythmogenesis during acute global ischemia in the context of the VEP hypothesis. However, most patients who receive implantable cardioverter defibrillators suffer from coronary artery disease and regional ischemia, frequently with a history of prior infarction. The infarct BZ is altered in structure (42) and electrophysiology (44). This provides a proarrhythmic substrate in the setting of chronic/healed MI (4 wk and older) and likely contributes to increased shock-induced vulnerability and defibrillation failure. However, shock-induced vulnerability under these conditions has not been well studied.
According to the upper limit of vulnerability (ULV) hypothesis of defibrillation (24, 9), defibrillation fails when a shock induces new arrhythmia. Defibrillation efficacy is inversely related to the probability of postshock arrhythmia formation. Thus the ULV is correlated with the defibrillation threshold (DFT), usually slightly lower than the DFT (9), which is demonstrated both in the canine model and in the clinical setting (9, 24, 32, 50). The goal of the present study was to use a combination of optical imaging and histological techniques to evaluate the mechanisms of initiation and maintenance of shock-induced arrhythmias in the setting of healed MI in rabbit hearts.
The experimental protocol was approved by the Animal Research Committee of the Cleveland Clinic Foundation. All animals used in this study received humane care in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Sixteen New Zealand White rabbits of either sex were used in the study, including eight rabbits with structurally normal hearts as controls and eight rabbits with healed MI. The average age of the animals before infarction was 112 ± 25 days. The infarction was created by ligation of the lateral division or posterolateral division of the left coronary artery at a level of 40–70% from the apex during in vivo survival surgery. Rabbits were anesthetized and intubated. Under sterile conditions, a left thoracotomy at the fourth intercostal space was performed, and the target artery was ligated. The ECG was monitored continuously during surgery. Immediately before ligation, 1 mg/kg of lidocaine was intravenously administered to minimize potential ventricular arrhythmias. A chest tube was inserted into the thoracic cavity. The lungs were inflated. The ribs on both sides of the incision were approximated by three evenly placed loops of heavy-duty thread. The thoracic cavity was then sealed via multiple suture layers. The chest tube was withdrawn while negative pressure was applied. An appropriate long-lasting antibiotic was given just before surgery, and an analgesic was given postoperatively for 3 days. Rabbits were allowed to heal for an average of 53 ± 21 days before the acute experiments. No survival surgery was performed on the control hearts.
The heart preparation during the acute experiments has been described previously (10, 11, 19, 20). After anesthesia, the control or infarcted heart was removed and placed into a temperature-controlled (37°C) Langendorff perfusion chamber, where it was perfused with oxygenated modified Tyrode solution. The electrophysiological activity from the anterior wall of the heart was recorded with optical mapping techniques. A modified 10-mm coil electrode was inserted into the right ventricular cavity through the pulmonary artery for shock delivery. The modified 5-cm shock delivery reference coil electrode was positioned in the bath 1–2 cm above and 1–2 cm behind the heart (4007L, Angeion). A pacing bipolar electrode was placed at the apex of the right ventricle. To eliminate motion artifacts, 20 μM cytochalasin D (Cyto D; Molecular Probes) was added to the perfusate. The heart was equilibrated in perfusate with Cyto D for 30 min and then stained with 10 μM di-4-ANEPPS (Molecular Probes) for 5–10 min.
Acute experimental protocol.
We used a S1-S2 protocol to measure the vulnerability of control (n = 8) and infarcted (n = 8) hearts. The heart was paced at a basic cycle length of 300 ms at twice the diastolic pacing threshold. After 20 basic stimuli (S1), a truncated exponential monophasic shock (S2; 8-ms duration, 61% tilt, both polarities) was delivered from a 150-μF capacitor clinical defibrillator (HVS-02, Ventritex) at the preset coupling interval (the time interval between the last S1 stimulus and S2 stimulus). The leading edge voltages of the shocks were ±100, ±130, ±160, ±190, and ±220 V. Shocks were delivered at a phase defined as the percentage of action potential duration (APD): (coupling interval − conduction delay)/APD × 100%. Conduction delay was defined as the time interval between the pacing stimulus and average activation time in the field of view. The APD was calculated using an APD at 90% repolarization definition as in our previous study (11). In every acute experiment, shocks were applied at 25%, 50%, and 75% APD.
Optical recording system.
The optical recording system has been previously described (19, 20). The heart was illuminated by excitation light with a wavelength of 520 ± 45 nm. Excitation and emission lights were separated by a 585-nm dichroic mirror. Emission light above 610 nm was collected by a 16 × 16 square photodiode array (Hamamatsu) coupled to a computerized data-conditioning and -acquisition system. The field of view was 17.5 × 17.5 mm in all experiments.
After the acute experiments, two regions of heart were sampled: 1) the left ventricular free wall with the papillary muscles and 2) the right ventricular free wall with the septum. The specimens were immediately embedded in tissue freezing medium (Sakura, International Medical Equipment) and frozen in isopentane cooled by liquid nitrogen. Frozen tissues were cryosectioned into 16-μm-thick sections, on which Masson's trichrome histology staining was performed to verify the existence of infarction and identify the pattern of the infarction BZ.
Analysis of optical data from structurally normal and chronic infarcted hearts.
Custom-made signal analysis software has been described previously (10, 11, 19, 20). This software facilitates calculation of activation, repolarization, APD, and calibrated Vm maps. Calibrations assumed a normal resting potential of −85 mV and an action potential amplitude of 100 mV. The gradient of the Vm and its derivative (dV/dt) was calculated using a five-point algorithm similar to that used in our previous report (45) to calculate the conduction velocity.
To quantitatively evaluate shock-induced vulnerability, we defined shock-induced arrhythmia as ≥1 postshock extra beat (EB), sustained arrhythmia as ≥6 postshock EBs (25), and long-lasting arrhythmia as >30 s postshock arrhythmia, which required a rescue shock to terminate. We also defined ULV under each shock-delivery phase as the highest tested shock strength with which sustained arrhythmia (≥6 EBs) was induced.
We further performed several qualitative and quantitative analyses on shock-induced long-lasting arrhythmia to explore arrhythmia maintenance in structurally normal and chronically infarcted hearts. First, we divided the long-lasting arrhythmias into two groups: 1) monomorphic arrhythmias with regular optical and ECG morphology repeating from beat to beat and 2) polymorphic arrhythmias with irregular optical and ECG morphology. We applied fast Fourier transform (FFT) analysis (46) to calculate the dominant frequency of the shock-induced long-lasting arrhythmias (>30 s), as described in our previous study (12). The length of the arrhythmia recordings analyzed was fixed at 1.5 s. The sum of the FFT curves of all 256 channels was called the composite FFT. In the composite FFT, the frequency component with the highest amplitude was considered the dominant frequency. We also performed phase analysis using the algorithms developed by Bray and Wikswo (4, 5) and Gray et al. (31) for each long-lasting arrhythmia. Using a phase representation, we identified phase singularities and calculated the average number of phase singularities over time in the field of view for each arrhythmia recording. We further traced the centers of the wavefronts during the long-lasting arrhythmias. At certain time, if the dV/dt recorded from a single channel reached 35% of its maximum depolarization (dV/dtmax) in the whole arrhythmia recording, it was considered inside the wavefront (upstroke of the action potential). A continuous region composed of such channels was considered one wavefront. The average number of wavefronts over the time was calculated from each arrhythmia recording. For each rabbit, the centers of the wavefronts for all the sampling time in all the arrhythmia recordings were plotted in a single panel to identify whether the arrhythmia wavefronts tended to aggregate around a landmark (for example, the BZ in the infarcted heart) or were uniformly distributed in the field of view. We chose to trace the wavefront centers instead of phase singularities for two reasons. First, in infarcted hearts, there might be numerous wavefront breakthroughs and/or retreats on epicardium. However, such breakthroughs or retreats did not form visible phase singularities but rather only increased the number of wavefronts. Second, the phase singularity algorithm was designed for two-dimensional (2-D) data. For the infarcted hearts, many arrhythmias might originate from or propagate through the surviving tissue in the endocardium and/or midmyocardium. Phase singularities might be formed in such surviving tissue in three dimensions (3-D) at a depth. However, currently available imaging techniques and algorithms cannot detect phase singularities in 3-D.
Group data are expressed as means ± SD. Statistical comparisons between group data were performed using the paired or unpaired t-test. Statistical comparisons between probabilities were performed using a 2 × 2, two-sided Fisher exact test. Differences were considered significant when P < 0.05.
Electrophysiological and histological characterization of infarction.
Figure 1, left, shows representative photographs of the infarcted heart preparation and the field of view used in this study. With transillumination, the infarcted apex was translucent compared with the other areas of the left ventricle, due to a remarkably reduced wall thickness and replacement of the myocardium with a fibrotic scar. To evaluate infarct size, we calculated the ratio of the translucent infarcted area to the entire ventricular anterior wall area in the transillumination photos. In Fig. 1, bottom left, the transparent infarcted area was 0.90 cm2, the whole ventricular anterior wall area was 8.12 cm2, and the ratio of the two values was 11%. For all eight rabbits with infarction, the ratio of the transparent infarcted area to the entire ventricular anterior wall area was 14 ± 5%. During optical imaging, we always included part of the infarcted area in the field of view unless infarction was only developed at the ventricular posterior wall. Figure 1, middle, shows action potentials recorded from 256 channels in the field of view. The infarct area was marked by the triangular action potential shape and a significantly reduced signal-to-noise ratio. The four expanded action potentials in Fig. 1, right, further illustrate these points.
Masson's trichrome staining was performed on the eight infarcted hearts to confirm the existence of the infarct and to examine the pattern of infarction after the acute experiment. Ligation of the lateral division or posterolateral division of the left coronary artery always resulted in a transmural left ventricular free wall infarction. In five of eight hearts, the infarction developed at the apex. In the other three hearts, the infarction developed in the middle of the left ventricular free wall. All 8 rabbits developed endocardial BZs, the thickness of which was 2–10 cell layers. The endocardial BZ always covered the entire inner surface of the transmural infarction. In addition, two of eight hearts developed epicardial BZs, with a thickness of less than three cell layers. Figure 2 shows one example of an apical infarction with an endocardial BZ. Figure 3 shows a representative example of an infarction in the middle of the left ventricular free wall with both endocardial and epicardial BZs. Both Figs. 2 and 3 illustrate that most infarcted areas are largely fibrotic scars, infiltrated by some surviving “islands” of myocardial fibers, which we named as midmyocardial BZ. Such myocardial fibers might spread across the infarcted area and connect with the noninfarcted tissue in one or both ends, providing pathways for reentrant arrhythmias. We observed midmyocardial BZ in all eight infarcted hearts.
Shock-induced vulnerability in structurally normal and chronically infarcted hearts.
Optical imaging studies were performed on structurally normal hearts (n = 8) and healed MI hearts (n = 8) to assess shock-induced vulnerability. We calculated the incidence of arrhythmia (≥1 EB), sustained arrhythmia (≥6 EBs), and long-lasting arrhythmia (>30 s) for each shock strength and shock delivery phase in four groups: 1) control heart during anodal shock application, 2) control heart during cathodal shock application, 3) infarcted heart during anodal shock application, and 4) infarcted heart during cathodal shock application. We also calculated the distribution of ULVs for each shock delivery phase in the above four groups during sustained arrhythmias. Results are summarized in Fig. 4 and Table 1. The incidence of arrhythmias increased significantly in the infarcted hearts versus control hearts during anodal shocks (Fig. 4A, top). This phenomenon was most obvious at 25% APD and 50% APD. During cathodal shocks, the incidence of arrhythmias at 50% and 75% APD in the infarcted hearts was similar to that in the control hearts. An increase of incidence was observed at 25% APD (Fig. 4A, bottom). The incidence of sustained arrhythmias increased significantly in the infarcted heart versus control heart for both shock polarities (Fig. 4B). Long-lasting arrhythmias were mainly observed in the infarcted hearts but rarely in control hearts for either shock polarity (Fig. 4C). Figure 4D illustrates the distribution of ULVs for sustained arrhythmias shown in Fig. 4B. In this study, we limited the highest shock strength to 220 V to avoid strong electroporation and tissue damage caused by larger internal shocks (1). In some cases, the ULV was above 220 V. We divided the ULV values into three ranges: 1) <100 V, 2) 100–220 V, and 3) ≥220 V. For both shock polarities and all the shock delivery phases, we observed an increase in the ULV in infarcted versus control hearts.
In summary, the incidences of arrhythmias, sustained arrhythmias, and long-lasting arrhythmias during cathodal shocks were consistently higher than those during anodal shocks for both control and infarcted hearts. The ratio of the sustained arrhythmias to total arrhythmias increased significantly in the infarcted versus control hearts for both shock polarities (63% vs. 27% for anodal shocks and 89% vs. 67% for cathodal shocks). This indicates that the arrhythmia duration is longer in infarcted hearts. The ratio of long-lasting arrhythmias to sustained arrhythmias also increased significantly in the infarcted hearts versus control for both shock polarities (92% vs. 22% for anodal shocks and 85% vs. 56% for cathodal shocks), indicating that arrhythmias are less likely to self-terminate in infarcted hearts. In general, the ULVs of the infarcted hearts were shifted to a higher shock strength compared with control hearts for the same shock polarity and delivery phase.
Mechanisms of initiation of shock-induced arrhythmias.
To explore the mechanisms of increased shock-induced vulnerability in infarcted hearts, we compared the initiation of shock-induced arrhythmias in control and infarcted hearts for both shock polarities. Figure 5 shows representative examples of the initiation of shock-induced arrhythmias in a control heart. Figure 5A shows the result from an anodal shock (+160 V, 8 ms) delivered at 75% APD. VEP was induced at the end of the shock, as shown in the postshock Vm map. The areas at the ends of the shock electrode were negatively polarized (hyperpolarized/deexcited, blue), and areas away from the shock electrode were positively polarized (depolarized/excited, red). As a result, postshock excitation (1 EB in this case) immediately originated from the boundaries between the depolarized and hyperpolarized areas after the shock withdrawal. It propagated through the hyperpolarized areas and exited from both the top and the bottom of the field of view (activation map, 0–6 ms in scale). After 73 ms after the shock withdrawal, the arrhythmia wavefront reentered the field of view from the top right corner. It was possible that the wavefront of the immediate postshock arrhythmia propagated along the boundary of the refractory area out of the field of view and reentered it when the region was recovered. Reversal of shock polarity resulted in a reversal of the VEP pattern as shown in the postshock Vm map (Fig. 5B). The areas at the ends of the shock electrode were now depolarized/excited (red) and became the driving force for the immediate postshock arrhythmia. In this case, a −160-V shock induced sustained arrhythmia lasting more than six EBs. The areas away from the electrode were hyperpolarized/deexcited (blue). Immediately after shock withdrawal, the arrhythmia wavefront propagated quickly across the deexcited area and out of the field of view. The wavefront reentered the field of view from both top and bottom portions and formed a figure-of-eight reentry, of which only right half could be observed in the field of view. Thus, consistent with our previous reports (11, 19, 61), the virtual electrode-induced phase singularity was responsible for the initiation of shock-induced arrhythmias in structurally normal hearts.
Figure 6. shows representative examples of the initiation of shock-induced arrhythmias in a heart with a healed infarct, with the same shock strength and shock delivery phase as those in Fig. 5. During anodal shock application (Fig. 6A), a hyperpolarized/deexcited area (blue area in the postshock Vm map) only existed in the vicinity of the top end of the shock electrode. There was no pronounced hyperpolarized/deexcited area at the other end of the electrode because of the existence of the apical infarction. As a result, the immediate postshock arrhythmia wavefront only propagated from the depolarized/excited area (red area in the postshock Vm map) to the hyperpolarized/deexcited area (blue area) at the top left corner and reentered the field of view at 80 ms after shock withdrawal. Block of arrhythmia wavefront propagation was observed at the infarction BZ (the boundary between the infarcted area and adjacent noninfarct area), possibly contributing to the formation of long-lasting postshock arrhythmia as illustrated partially in the optical traces. During cathodal shock application (Fig. 6B), the VEP pattern was reversed. Again, near the infarct area, the VEP was diminished compared with that in Fig. 5B. After shock withdrawal, two excitation wavefronts appeared at the infarction BZ (activation map, 0–20 ms in scale: 2 white areas at the bottom of the field of view) and propagated across the hyperpolarized/deexcited area, forming a long-lasting reentrant arrhythmia. Thus this shock-induced arrhythmia did not initiate from the boundary between depolarized and hyperpolarized areas in the top of the field of view as that in Fig. 6A. Instead, the infarction BZ altered the VEP pattern and provided the substrate for postshock arrhythmia initiation.
Figure 7. further illustrates this point. In this case, a shock of −100V was applied during the plateau phase. Again, the shock produced a VEP pattern with a depolarized/excited area (red) near the shock lead and an adjacent hyperpolarized/deexcited area (blue) as shown in the postshock Vm map. The maximum Vm gradient (dark blue areas, middle of the third row) was located at the BZ (the boundary between the infarcted area and adjacent noninfarct area) as well as at the boundary between the depolarized and hyperpolarized areas. However, the resulting break-excitation wavefront (pink area, right of the third row) only originated at the BZ gradient and propagated toward the base, forming a long-lasting reentrant arrhythmia (postshock activation map).
We further systematically examined all arrhythmias in control and infarcted hearts to evaluate the general patterns of postshock arrhythmia initiation. The results are summarized in Tables 2 and 3, respectively. In 8 control hearts (Table 2), a total of 33 arrhythmias was induced during anodal shock applications, of which most (29 of 33) were initiated from the VEP boundary, as shown in Fig. 5A. Only a few postshock arrhythmias (4 of 33) were initiated in the bulk of myocardium, appearing as breakthroughs on the epicardium with a delay after shock withdrawal. We further counted the incidence of postshock figure-of-eight reentries in the first postshock cycle (as shown in Fig. 5B) of VEP-induced arrhythmias. In 9 of 29 VEP-induced arrhythmias, 2 wavefronts reentered from both the top right and bottom right corner and formed the figure-of-eight reentry. In the other 20 arrhythmias, a single wavefront reentered from either the top or bottom of the field of view after delay. For cathodal shock applications, a total of 81 arrhythmias was induced in control hearts. Only two arrhythmias (2 of 81) initiated in the bulk of the myocardium appeared as breakthroughs on the epicardium. In contrast, 79 of 81 arrhythmias were initiated from the VEP boundary, of which 13 were of figure-of-eight morphology. In the rest of the 66 arrhythmias, a single wavefront reentered from either the top or bottom of the field of view after a delay. For both shock polarities, the arrhythmias appearing as breakthroughs on the epicardium were predominately induced at early shock delivery phases (25% APD), and the figure-of-eight reentries were predominately induced at a later shock delivery phase (75% APD).
In the eight infarcted hearts (Table 3), two developed infarctions on the posterior wall and could not be observed in the field of view. We examined all arrhythmias induced in the other six hearts. For anodal shock applications, a total of 58 arrhythmias was induced: 29 of 58 arrhythmias were initiated from the VEP boundary (as shown in Fig. 6A), 17 of 58 were initiated from the BZ (as shown in Fig. 6B), and the remaining 12 arrhythmias were initiated from both the VEP boundary and BZ. For cathodal shock applications, the incidence of arrhythmias initiated in the BZ dramatically increased. In all 68 shock-induced arrhythmias, 47 of them were initiated from the BZ, 13 were initiated from the VEP boundary, and 8 were initiated from both the BZ and VEP boundary.
In summary, the primary mechanism of shock-induced arrhythmia initiation in the control hearts was via virtual electrode-induced phase singularity. In some occasions, arrhythmias initiated in the bulk of the myocardium appearing as breakthroughs on the epicardium at early shock delivery phases (25% and 50% APD). The later shock delivery phase (75% APD) facilitated the formation of figure-of-eight reentry. In the infarcted hearts, for anodal shock applications, at early shock delivery phase, postshock arrhythmias predominantly initiated from the BZ. At the later shock delivery phase, both the BZ and VEP played important roles in postshock arrhythmia initiation. For cathodal shock applications, VEP became less important and most of the postshock arrhythmias were initiated from the BZ area.
Mechanisms of maintenance of shock-induced arrhythmias.
To explore why shock-induced arrhythmias in infarcted hearts were more sustained and thus more frequently required rescue shocks, we analyzed the characteristics of all of the long-lasting arrhythmias (>30 s) in control and infarcted hearts. The results are shown in Fig. 8. and Table 4 and in the on-line data supplement (movies 1–3; http://ajpheart.physiology.org/cgi/content/full/01253.2004/DC1). In the 8 control hearts, a total of 32 long-lasting arrhythmias was induced during 240 shock applications with both polarities at all 5 shock strengths and 3 shock delivery phases (13.3% incidence). On the contrary, in the 8 infarcted hearts, a total of 105 long-lasting arrhythmias was induced (43.8% incidence). In the control hearts, all of the long-lasting arrhythmias were induced in only 3 of the rabbit hearts; 2 of 32 were induced with anodal shocks, and 30 were induced with cathodal shocks (Fig. 4C). For the infarction group, long-lasting arrhythmias were induced in all eight hearts.
We examined the morphology of all the long-lasting arrhythmias based on visual inspection of the ECG and optical recordings. In the control group, all the long-lasting arrhythmias were polymorphic in nature, consistent with our recent findings with Cyto D as an excitation-contraction uncoupler (12). In the infarction group, 9 long-lasting arrhythmias were monomorphic and 96 were polymorphic. Interestingly, we observed an ECG alternans pattern (online data supplement movie 2) in all nine monomorphic arrhythmias. These monomorphic arrhythmias were only induced in two infarcted hearts, of which no polymorphic arrhythmias were induced. We calculated the dominant frequency and numbers of phase singularities and wavefronts averaged over the recording time for all of the long-lasting arrhythmias (Table 4). There was no significant difference between the average numbers of phase singularities and wavefronts in control and infarcted hearts. However, the dominant frequency of long-lasting arrhythmias in the control hearts was slightly but significantly higher than in the infarcted hearts. We further calculated the average number of wavefronts in the infarcted hearts for two different conditions: 1) with (n = 6 hearts) or 2) without (n = 2 hearts) visible infarction in the field of view. We found that the average number of wavefronts in infarcted hearts without visible infarction in the field of view decreased slightly but significantly compared with that in control hearts (1.97 ± 0.30 vs. 2.17 ± 0.34, P < 0.05). Also, without infarction in the field of view, the average number of wavefronts was slightly but significantly smaller that that with visible infarction in the field of view (1.97 ± 0.30 vs. 2.16 ± 0.34, P < 0.05). There was no significant difference between the average numbers of wavefronts in control and infarcted hearts with visible infarction in the field of view (2.17 ± 0.34 vs. 2.16 ± 0.34, P = 0.93).
Furthermore, for each heart, we plotted the centers of the arrhythmia wavefronts of all the long-lasting arrhythmia recordings in a single panel to evaluate the distribution of wavefronts. The results are shown in Fig. 8. We divided all the hearts with long-lasting arrhythmias into four groups: 1) control hearts (n = 3); 2) infarcted hearts without visible infarction in the field of view (n = 2); 3) infarcted hearts with visible infarction in the field of view in which only monomorphic arrhythmias were induced (n = 2); and 4) infarcted hearts with visible infarction in the field of view in which only polymorphic arrhythmias were induced (n = 4). Figure 8A shows the examples from all the groups. In the control heart, wavefront centers of long-lasting arrhythmias distributed uniformly in the field of view (Fig. 8A, first column). In the infarcted heart without visible infarction in the field of view (Fig. 8A, second column), the distribution of long-lasting arrhythmia wavefront centers was very similar to the pattern in the control heart. In the infarcted heart with visible infarction in the field of view and monomorphic arrhythmias induced, the wavefront propagation pathways were highly repetitive (Fig. 8A, third column). We did not find a specific relationship between the position of the infarction BZ and the distribution of the monomorphic arrhythmia wavefronts. In contrast, for infarcted hearts with polymorphic arrhythmias, two aggregations of arrhythmia wavefront centers were observed in the vicinity of the infarction BZ (Fig. 8A, fourth column), implying that the arrhythmia wavefronts tended to attach to, break through, or retreat from the BZ. To elucidate whether such patterns were repeated in each group, we calculated the distances of all the wavefront centers to a certain landmark for all the hearts and plotted the histogram of the distances for each group in Fig. 8B. For both the control and infarcted hearts without visible infarction in the field of view, because there was no obvious anatomic landmark, the distances of all the wavefront centers to the field of view center were calculated. For the infarcted hearts with visible infarction in the field of view, the infarction area boundaries on the epicardium were chosen as the anatomic landmarks (the dashed curves in Fig. 8A). If a wavefront center was inside of the dashed curve, the distance was considered negative. For the control and infarcted hearts without visible infarction in the field of view, the number of wavefront centers increased linearly with the distance to the field of view center, suggesting a uniform distribution (Fig. 8B, first and second columns). For the infarcted hearts with visible infarction and monomorphic arrhythmias induced, there was no obvious distribution pattern (Fig. 8B, third column), maybe because of the high repetitive and specific wavefront propagation pathways of each rabbit heart. For the infarcted hearts with visible infarction in the field of view and polymorphic arrhythmias induced, a peak at distance = 0 (in the vicinity of the infarction BZ) was observed, suggesting that in the other three hearts, the wavefront distribution patterns were repeated as in Fig. 8A, fourth column. The wavefronts of the long-lasting arrhythmias tended to attach to, break through, or retreat from the infarction BZ.
To further elucidate how the long-lasting arrhythmia wavefronts propagated, we show three representative examples of long-lasting arrhythmias in the on-line data supplement (http://ajpheart.physiology.org/cgi/content/full/01253.2004/DC1): 1) polymorphic arrhythmia in a control heart (on-line movie 1); 2) monomorphic arrhythmia in an infarcted heart (on-line movie 2); and 3) polymorphic arrhythmia in an infarcted heart (on-line movie 3). In on-line movie 3, we frequently observed the arrhythmia wavefront breakthroughs near the infarction BZ. This suggests that the midmyocardial BZ provides pathways for observed more sustained and severe arrhythmias in infarcted hearts. Furthermore, wavefront propagations along the infarction BZ were commonly observed. In comparison, in the control heart, the long-lasting arrhythmia was mainly observed as functional reentry(s) in the field of view. In on-line movie 2, alternans were observed in the ECG recording. This example was picked from the same rabbit as that shown in Fig. 8A, third column. The infarction was formed in the middle of the left ventricle, as shown by the dashed black line. During each beat, a global arrhythmia wavefront swept from the apex to base of the heart. However, when the global arrhythmia wavefront reached the top left corner of the field of view, an isolated residual wavefront could be observed near the infarction BZ and then retreated from the epicardium. The exit site of this residual wavefront slightly varied from beat to beat, likely contributing to the ECG alternans. The slightly varying exit sites of residual wavefronts might result from propagations through different midmyocardial BZ pathways (see Figs. 2D and 3, D and F).
The main findings of this study were that, first, the rabbit model of healed MI consistently reproduces the endocardial BZ and, second, a significant increase of shock-induced vulnerability is observed in infarcted compared with control hearts. This is reflected in increased incidence of all shock-induced arrhythmias as well as sustained arrhythmias and long-lasting arrhythmias. Third, the severity and duration of shock-induced arrhythmias were also increased in hearts with healed infarctions. This is evident from a significantly increased propensity for long-lasting arrhythmias requiring defibrillation shocks to terminate. Finally, regional MI modulates shock-induced VEP patterns in a different way than in structurally normal hearts. The infarction BZ is responsible for both the initiation and maintenance of shock-induced arrhythmias, which results in increased ULV.
Our healed infarcted rabbit heart model has several characteristics not possessed by the well-studied canine heart model of MI, specifically at the chronic infarction stage. First, in the canine models, the infarct BZ mostly develops at the epicardium (38, 40, 53, 56). Reports from only one group observed the development of endocardial BZ (28, 29). In clinical settings, both endocardial and epicardial BZs are frequently observed (3, 26). In our study, the infarcted rabbit hearts consistently developed endocardial BZ. Thus this model may provide us with additional insights that are not readily acquired from most of the canine infarct models. Second, in the canine model, ventricular arrhythmias can be induced during both the healing and healed phases of infarction (17, 22, 23, 33, 34, 36). However, after the first week, shock-induced vulnerability decreases (34, 36). Sometimes sustained arrhythmias are not inducible at all in the infarcted canine hearts (43). In human studies, ventricular tachycardias could be induced in about 10% of patients after 5 days of infarction but in 20% to 50% after 3 wk (15). In our study, sustained arrhythmias were induced in all rabbits with more than 4 wk post-MI, providing a chronic infarction model for electrophysiology study with a high incidence of reproducible ventricular arrhythmias. Finally, in the well-studied canine model of permanent left anterior descending artery occlusion, most of the arrhythmias initiated from the epicardial BZ (8, 22, 23, 38, 55). The location of the reentrant circuits correlates with the location of the epicardial BZ (18, 43, 57). In clinical observations, arrhythmias arise from both endocardial and epicardial BZs, with endocardial BZ initiation more frequently observed (58). In our study, postshock arrhythmias mainly initiated from endocardial and/or midmyocardial BZs with frequent breakthroughs on the epicardium near the BZ area, which is seldom observed in the canine model. Thus this rabbit model may enhance our understandings of clinically observed endocardially initiated and maintained arrhythmias.
The shock-induced arrhythmogenesis and arrhythmia maintenance in chronically infarcted hearts are very different from those during acute global ischemia (10, 2). In rabbit hearts with acute global ischemia, the basic pacing wavefront propagation pattern, shock-induced VEP pattern, and arrhythmia origination pattern (via VEP) are very similar to those in control hearts (10). However, widening of the vulnerable window was observed in ischemia hearts (10, 2), which leads to high incidence of arrhythmias (10), (58). Also, the ULV and DFT remained unchanged (2). This widening of the vulnerable window results most likely from the significant increased repolarization dispersion (2, 10, 58), which in turn is the result of severely disrupted repolarization due to ischemia. In the acute global ischemia hearts, the arrhythmias originated and maintained through functional heterogeneity rather than structural heterogeneity as in chronically infarcted hearts. However, there is a certain level of similarity between acute regional ischemia hearts and chronically infarcted hearts. In regional ischemia hearts, there is also a BZ developed, characterized by low tissue impedance and high excitability compared with central ischemia zones (16, 58). Arrhythmias arise preferentially from the ischemic BZ (14, 58). Ventricular fibrillations were also observed locally anchored around fixed lines of block, indicating that structural rather than functional changes underlie the substrate for arrhythmia maintenance (16, 58). Interestingly, such acute regional ischemia causes an increase of wavebreaks in the ischemic BZ and a decrease of wavebreaks in the central ischemia zone area without a significant change in the nonischemia zone area (63), which is very similar to what we observed of the heterogeneous distribution of wavefronts in chronically infarcted hearts.
In structurally normal hearts, our results further confirm previous observations from our group that shock-induced arrhythmias were initiated via virtual electrode-induced PS mechanisms (19). The increased incidence of arrhythmias during cathodal shock applications relative to anodal shock applications is also in agreement with our previous observations (10, 61). Furthermore, as previously reported (13), a small number of postshock arrhythmias were initiated from the bulk of myocardium appearing as breakthroughs on the epicardium when the shock strength was insufficient to create strong transmural VEP, usually when shocks were delivered at early phases of APD. We (61) have provided mechanistic insights regarding shock-polarity and coupling interval dependences of vulnerability in our previous study.
In infarcted hearts, the symmetric pattern of VEP on the epicardium no longer exists due to the existence of infarction area during both anodal and cathodal shocks. Still, for anodal shocks, in 71% of the shock-induced arrhythmias, postshock wavefronts were initiated at the remaining boundary between depolarization and hyperpolarization areas on the epicardium (Fig. 6A). However, for cathodal shocks, shock-induced arrhythmias tend to initiate from the infarction BZ, not from the remaining VEP boundaries (Figs. 6B and 7). The existence of the BZ facilitates breakthrough excitations originating from the depth. When the VEP on the epicardium is too weak at early shock delivery phases, breakthrough excitations in the vicinity of the BZ dominate. In later shock delivery phases, breakthrough excitations compete with the postshock wavefronts originating from the VEP boundaries and frequently overwhelm the latter wavefronts, especially during cathodal shock applications. However, due to technical limitations, we cannot directly observe how these breakthrough excitations initiate in the depth of myocardium. In the present study, we designed experiments so that we could record the electrophysiological activities such as arrhythmogenesis from intact hearts, a condition closer to those under clinical settings. For this reason, we did not cut and expose the endocardium of the heart for optical mapping. In future studies, we may further explore the arrhythmogenesis in this model by directly mapping endocardial electrical activities.
We frequently observed the initiation of shock-induced arrhythmias near the infarct BZs, mostly in the zones adjacent to the bulk of the noninfarcted area (Figs. 6 and 7). Because epicardial BZs are usually absent in our infarcted heart model (for example, no epicardial BZ was observed in the hearts shown in Figs. 6 and 7) and because the optical imaging techniques allowed us to map the averaged electrical activities of areas 500 μm–1 mm in depth, observed arrhythmias is most likely initiated from endocardial and/or midmyocardial BZs. It is possible that such postshock excitations are induced by the VEP in the endocardial or midmyocardial area and propagate to the epicardium surface by the “shortcuts” of midmyocardial infarction BZs. Furthermore, it is likely that the mechanism of such breakthrough initiations is reentry induced by VEP in midmyocardium or endocardium that is supported by the following evidences (7): 1) such breakthroughs can be repeatedly induced by a programmed stimulation and 2) we observed the reverse relationship between the coupling interval and the initiation of postshock breakthroughs. At 25% APD shock delivery phase, a long delay was observed between the shock applications and the breakthroughs in the field of view. In 75% APD shock delivery phase, such breakthroughs were immediately postshock, facilitating the breakthrough excitations to overwhelm the effect of the epicardial VEP, which is always immediately postshock.
We also studied the mechanisms of arrhythmia maintenance in control and infarcted hearts. In control hearts, long-lasting arrhythmias were polymorphic in nature, in agreement with our previous report using the same excitation-contraction uncoupler, Cyto D (12). In three of eight control hearts, long-lasting arrhythmias were induced by cathodal shocks for all shock strengths delivered at 50% and 75% APD (Fig. 4C, bottom left; all columns are of the same size). The vulnerabilities of these three control hearts were similar to those of the infarcted hearts. The dominant frequency of the long-lasting arrhythmias in the control hearts was slightly but significantly higher than that in the infarcted hearts. However, the long-lasting arrhythmias in both groups had dominant frequencies of <9 Hz and should be considered tachycardiac in nature, not fibrillation. Except for the difference of dominant frequency values, the numbers of PSs and wavefronts averaged over time were almost the same, which implies that the polymorphic long-lasting arrhythmias in the control and infarcted hearts are similar in morphology. In fact, the primary difference is that polymorphic arrhythmia wavefronts in the infarcted hearts were frequently observed to attach to, break through, or retreat from the infarction BZs. These wavefront breakthroughs or retreats did not form visible PSs on the anterior epicardium. However, such shortcuts along the infarction BZ facilitate the maintenance of arrhythmias. For the maintenance of long-lasting monomorphic arrhythmias in the infarcted hearts, we did not observe a significant influence of the infarction BZ on arrhythmia pathways, in which other mechanisms we do not know may play a role. However, only 9% of all shock-induced arrhythmias in the infarcted hearts are monomorphic. We can safely conclude that the infarction BZ facilitates the maintenance of postshock arrhythmias. We also observed that the dominant frequency of long-lasting arrhythmias was lower in infarcted hearts, which suggests that the wavefront conductions are slowed in the infarction BZ areas. Furthermore, the reentrant core size in anatomic reentry was usually larger than that in functional reentry, which may also contribute to the low dominant frequency observed in infarcted hearts versus control hearts.
In conclusion, the infarction BZ plays an important role in both increased shock-induced arrhythmogenesis and arrhythmia maintenance in the rabbit model of healed MI.
Our experimental field of view was limited to the anterior epicardium. We could not directly observe shock-induced responses at the posterior epicardium, septum, or endocardium, which may be important in postshock arrhythmogenesis and arrhythmia maintenance, especially in infarcted rabbit hearts. We did not attempt to map electrical activities from the endocardium via cutting the heart and exposing the endocardium, because one of our main objectives in this study was to investigate arrhythmogenesis in intact hearts of both control and infarction groups. Currently, there is no optical mapping technique available to assess the 3-D map of electrical activities in the whole heart. Although we did not map the endocardium directly, we still frequently observed wavefront breakthroughs and retreats near the infarction BZ using current mapping techniques.
This study was supported by National American Heart Association Grant 0235172N (to Y. Cheng) and National Heart, Lung, and Blood Institute Grant R01-HL-67322 (to I. R. Efimov).
We thank William Kowalewski for excellent technical support. We are grateful to Dr. David Van Wagoner for careful review of the manuscript and helpful suggestions on its improvement.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society