Electrical Activation During Ventricular Fibrillation in the Subacute and Chronic Phases of Healing Canine Myocardial Infarction
Background Little information is available regarding the effects of myocardial infarction on the characteristics of ventricular fibrillation (VF). Epicardial activation during VF can be characterized by the cycle length and by the characteristics of activation wave fronts.
Methods and Results VF was induced by programmed stimulation in 6 dogs with subacute healing (1 week) myocardial infarction (MI), 5 dogs with chronic (8 week) healing MI, and 6 dogs without MI. Using a plaque electrode array with a 2.5-mm interelectrode distance, 112 electrograms were recorded and 91 vector loops were created for each cycle of VF from either the anterior (infarcted) or lateral (noninfarcted) wall. Direction of maximum epicardial activation was determined at each site for the first 10 cycles of VF (early) and for 10 cycles after 5 seconds of VF (late). Wave front size was determined based on a similarity in epicardial activation directions within a given area and by a statistical analysis that determined the degree of spatial linking at varying distances over the recording plaque. VF cycle length was defined as the mean interval of 10 consecutive local activation times. Differences among groups and differences between the anterior and posterolateral walls were determined by ANOVA. The mean wave front area was significantly larger in the presence of subacute MI (97±4 mm2, early; 78±3 mm2, late) or chronic MI (94±5 mm2, early; 78±5 mm2, late) than in noninfarcted animals (73±5 mm2, early; 61±3 mm2, late). The degree of linking of epicardial activation directions was similar in the three groups at distances of 2.5 and 5.0 mm but was lower at a distance of 7.5 mm among animals without infarction, confirming a smaller wave front size and suggesting less organization of activation. VF cycle length was significantly longer in the presence of infarction (98±5 ms, normal control animals; 121±13 ms, subacute MI; 127±13 ms, chronic MI). VF cycle length was significantly longer over the anterior than the lateral wall in the presence of subacute MI (131±8 ms, anterior; 109±5 ms, lateral) or chronic MI (136±9 ms, anterior; 119±6 ms, lateral) but not in noninfarcted animals (99±5 ms, anterior; 97±5 ms, lateral). The prolongation of VF cycle length among animals with infarction was associated with slower estimated conduction velocities during VF.
Conclusions During VF, in animals with subacute or chronic healing MI, (1) the size of activation wave fronts is larger, (2) the cycle length of VF is longer, (3) the conduction velocities are slower, and (4) the degree of organization is greater than in control animals. Thus, the characteristics of VF throughout the heart are altered by the presence of regional myocardial infarction. The implications of these findings for the initiation and maintenance of VF in the presence of different underlying myocardial substrates require further study.
Sudden cardiac death secondary to ventricular fibrillation (VF) is a problem seen primarily in patients with extensive coronary artery disease and previous myocardial infarction.1 2 Even in the absence of coronary artery disease, there is often significant left ventricular dysfunction and scarring present.3 4 Therefore, a better understanding of the characteristics of VF in hearts with chronic infarction is important. In addition it has been clinically recognized that the risk of sudden cardiac death decreases during the recovery period after myocardial infarction.5 6 Therefore, the healing of infarction may influence the vulnerability to VF.
VF has been described as a chaotic, disorganized rhythm.7 8 9 However, more recent work using extensive epicardial and three-dimensional mapping techniques has suggested that activation during VF has an underlying organization.10 11 12 13 14 These studies have provided some insight into the characteristics of VF. However, this work has had several limitations, the most important being that all of these studies were carried out in animal models without myocardial infarction. We have recently reported on the use of the technique of vector mapping for describing electrical activation during VF in a subacute (5 to 7 days old) canine myocardial infarction model.15 Our study showed that there is significant organization of activation during VF both at the onset of VF as well as during established VF, although the extent of organization decreased over time. Specifically, we found evidence for both “spatial and temporal linking” of epicardial activation directions. That is, activation direction at a given epicardial site was related to activation directions at closely spaced surrounding sites during the same and prior beats of VF.
In the present study, additional characteristics of VF were assessed and compared among animals with no infarction, subacute healing infarction (5 to 7 days old), and chronic healing infarction (8 weeks old). First, vector mapping was used to study the size of activation wave fronts during VF in these three models. Second, spatial linking of epicardial activation directions was evaluated at varying distances over the epicardial surface as a means of estimating wave front size and extending our previous observations on organization of activation. Finally, the cycle length of VF and the velocity of epicardial conduction during VF were determined in all three models.
Seventeen adult mongrel dogs weighing 10 to 25 kg were used in this study. Group 1 (n=6) consisted of control animals without infarction. Group 2 (n=6) consisted of animals with subacute healing infarction (5 to 7 days old). Group 3 (n=5) consisted of animals with chronic healing infarction (8 weeks old). The methods for creation of myocardial infarction have been previously described in detail.15 16 The experimental myocardial infarction was produced by ligation of the left anterior descending coronary artery. After anesthesia, the chest was opened by a left lateral thoracotomy in the fifth intercostal space, and the left anterior descending coronary artery was isolated at a point just proximal to the first major diagonal branch and ligated. The chest was then closed, and the animals were allowed to recover.
Electrophysiological Study and Mapping Protocol
The protocol used has been previously described in detail.15 Five to 7 days after creation of myocardial infarction, animals in group 2 underwent electrophysiological and epicardial mapping studies. The same protocol was performed in group 3 animals 8 weeks after infarction. Six control animals (group 1) were also studied. The animals were anesthetized with sodium pentobarbital 30 mg/kg IV and then intubated and mechanically ventilated. The chest was opened by a median sternotomy, and bipolar electrodes that were used for ventricular pacing and programmed stimulation were sutured to the epicardium of the right and left ventricles. A handheld plaque electrode array containing 112 unipolar recording electrodes arranged in a rectangular configuration (8×14) with 2.5-mm interelectrode spacing (Bard Inc) was placed on the epicardial surface of the anterior left ventricular free wall with the long axis of the plaque parallel to the left anterior descending artery. The 112 unipolar signals as well as the standard surface ECG limb leads were acquired and stored continuously in digitized form on videotape using a cardiac mapping system (Map Tech). The stored signals were later processed off-line to determine local activation times and to create vector loops.
Ventricular stimulation was performed using rectangular pulses of 2-ms duration at twice diastolic threshold delivered by a programmable stimulator (Bloom Associates). Recordings were obtained from the anterior wall and subsequently from the posterolateral left ventricular free wall (with the short axis of the plaque parallel to the left circumflex artery). VF was induced with the recording plaque initially on the anterior wall. In all dogs, an initial attempt was made to induce VF using standard programmed electrical stimulation techniques. After an eight-beat drive train at a cycle length of 280 to 300 ms, up to four premature stimuli were delivered from the left ventricular pacing site. If VF was not induced from this site, programmed stimulation was then repeated from the right ventricular pacing site. If this was unsuccessful, VF was then induced with rapid ventricular pacing at four to five times diastolic threshold with a paced cycle length as low as 100 ms.
After induction of VF, electrograms were recorded for a period of 10 seconds. The recordings were then stopped, and the animal was defibrillated using epicardial DC shock(s) beginning at 10 J and increasing the energy as needed to achieve defibrillation. Animals then underwent a second induction of VF after a recovery period of 10 minutes to ensure stability of the underlying rhythm as well as systemic blood pressure. We have previously demonstrated that multiple short inductions of VF are well tolerated and do not result in any changes in the characteristics of VF (with respect to the organization of epicardial activation).15 During the second induction, electrograms were recorded from the posterolateral left ventricular free wall. In the case of lateral wall recordings (as opposed to anterior wall recordings), the electrograms were obtained in the region of healthy muscle that had not undergone infarction (as determined by visual inspection).
After completion of electrophysiological testing and mapping, the animals were euthanized with sodium pentobarbital 30 mg/kg IV followed by intravenous supersaturated KCl. Sutures were placed at the corners of the recording plaque (while the plaque was on the anterior wall) to serve as a reference during subsequent pathological studies. The heart was removed and sliced in 1-cm sections from apex to base parallel to the atrioventricular groove. Histological sections were obtained from each slice at 2.5-mm intervals. Each section was characterized according to one of three patterns that were observed: (1) complete transmural infarction without a surviving subepicardial layer, (2) complete viability, or (3) combination of infarction and viable tissue. The last pattern typically resulted in an area of nontransmural infarction with a clearly demarcated surviving subepicardial layer. In some cases, however, there were areas of necrosis interspersed with viable tissue throughout the thickness of the myocardium. For each section with both necrotic and viable myocardium, the thicknesses of the viable and nonviable layers were determined. The extent of transmurality of infarction was determined as the ratio of the infarcted myocardium to the total wall thickness; the mean transmural extent of infarction among sections exhibiting this pattern was then calculated. Histology was also used to confirm that the lateral wall in animals with infarction and all of the heart in control animals were free of infarction. The degree and pattern of infarction were determined without information as to the electrophysiological characteristics by the pathologist (S.I.R.).
Activation Time Mapping Methods
After completion of the animal experiments, the data were analyzed as follows. The electrograms and ECG signals for each episode of VF were played back from videotape and displayed on an analog monitor. For each episode of VF, the first 10 cycles at the onset of VF (transition to VF) and 10 cycles after 5 seconds of VF (established VF) were analyzed. The data were filtered with a high-pass filter of 0.1 Hz and acquired with an eight-bit resolution at a sampling frequency of 1000 Hz. Bipolar electrograms were digitally created from adjacent overlapping unipolar recording sites (interelectrode distance, 2.5 mm) as described below in “Vector Methods” and displayed at high gain. Local activation was defined at the point of peak negative dV/dT in the unipolar electrograms. However, in order for an activation to be present, the peak absolute value of dV/dT in the unipolar electrograms must have exceeded 1.0 V/s15 17 or a distinct local bipolar electrogram must have been present. Activation times for bipolar electrograms were taken at the peak amplitude(s).
The methods for creating vector loops have been described previously in detail.18 19 20 A vector loop is created by summing two orthogonal bipolar electrograms (representing the x and y axes). The direction of the maximum vector of a given vector loop represents the direction of local activation in cardiac tissue. Orthogonal bipolar electrograms were created by summing the opposite “corners” of each group of four electrodes. Repeating this process for all 112 unipolar electrograms in the recording plaque resulted in the generation of 91 separate pairs of orthogonal bipolar electrograms for each cycle of VF. These electrogram pairs were summed and a plot representing the maximum vectors created. The time window representing a cycle of VF (for summing the local electrograms and generating the vector loops) was manually chosen by one of the investigators (R.D.) by scanning through all of the local unipolar electrograms for that cycle. While the individual cycles of VF may differ in duration, timing, and frequency in different regions of the myocardium, this was not a problem in the relatively limited regions of epicardium that were analyzed in this study.
One potential problem that arises with the use of the maximum vector method is that in cases where there are complex local activation patterns (for example, as may be seen in the case of conduction block), the local electrograms may be fractionated and the corresponding vector loop may be multidirectional, indicating the presence of more than one important activation direction.19 Use of the maximum vector in these situations may fail to accurately reflect the pattern of local electrical activation. We previously examined vector loops during VF, showing that for more than 85% of vector loops there is only a single direction identified and that for more than 95% of vector loops only a single predominant direction is present.15 Therefore, only maximum vectors were used in this study.
Activation Wave Fronts
Two separate methods were chosen for estimating the size of activation wave fronts. One was based on a visual assessment of wave front size by using plots of the maximum vectors; the second was based on the use of linking analysis at varying distances over the recording plaque. In both cases, a wave front was defined as a region of myocardium activated in a similar direction.
The first method for determining wave front size was based on a simple visual assessment of activation wave fronts. Plots of the maximum vectors were created for each cycle of VF. Activation wave fronts were then selected by one of the investigators (R.D.) based on the presence of four or more adjacent vectors with vector angles that could have occurred by random chance less than 1% of the time. That is, if the random probability of the observed angles for four or more adjacent vectors was less than 1%, then it was assumed that this represented an organized activation wave front. Based on this probability analysis, the specific criteria that were used (for different numbers of vectors) are as follows: (1) four vectors: any two adjacent vectors within 15 degrees (all vectors within 45 degrees), (2) five vectors: any two adjacent vectors within 20 degrees (all vectors within 60 degrees), and (3) six or more vectors: any two adjacent vectors within 30 degrees (all vectors within 90 degrees).
For each cycle of VF analyzed, the number of activation wave fronts present and the area of each wave front were determined. The mean number of activation wave fronts per cycle and the mean wave front area per cycle were determined for 10 consecutive cycles at the onset of VF and for 10 consecutive cycles of established VF for each episode of VF analyzed. Since areas without a surviving epicardial layer could not be analyzed, the mean number of activation wave fronts was normalized to the viable area that was evaluated so that comparisons could be made among animals.
The second method for assessing wave front size was based on linking analysis of epicardial activation directions, as has been described previously in detail.15 For each time window (transition to VF and established VF) of each episode of VF analyzed, a multivariate linear regression model was created in which the direction of activation at a given site during a given beat was predicted by the activation directions at eight adjacent sites for the current beat.15 21 22 The presence of a significant model (P<.05) indicated that activation was not random and that there was linking of activation directions. In addition, the strength of the model as reflected by the correlation coefficient (r) indicated the degree of organization or linking. In experiments where localized areas of epicardium were not activated, the number of sites used in the linking analysis had to be decreased as previously described because vector loops could not be created for these sites.15 In the present study, electrogram dropout was seen only in animals with infarction and only over infarcted tissue (the anterior wall). It occurred in four of six animals in group 2 and three of five animals in group 3.
Our hypothesis for the use of linking analysis was that if spatial linking of activation directions was present at a given distance over the recording plaque, then there must be activation wave fronts spanning at least that distance. Linking analysis was performed by using activation directions at sites located 2.5 mm, 5.0 mm, and 7.5 mm from the “site of interest” (the dependent variable in the regression analysis). Distances greater than 7.5 mm could not be assessed because this would have resulted in an incomplete set of surrounding predictor sites for any chosen site of interest. Using the methods described, correlation coefficients were calculated for each of the two windows (transition to VF and established VF) for each episode of VF analyzed.
Both methods of wave front analysis could tend to underestimate wave front size. A single large wave front that changes direction over a small distance would be considered as more than one wave front. In addition, linking analysis provides only a lower limit of wave front size, since some adjacent vectors will be part of two separate wave fronts and thus have divergent local activation directions and vector angles.
VF Cycle Lengths
VF cycle length was defined as the mean interval of 10 consecutive local activation times averaged over 28 equally spaced sites on the recording plaque. The sites used for this analysis were separated by 5.0 mm. Activation times were assigned using the criteria described above. In situations where apparent “double potentials” were present, a single activation time corresponding to the more rapid deflection was assigned for purposes of defining the VF cycle length. In addition, sites recording only distant activation were not included in the analysis. For each episode of VF analyzed, the cycle length was determined during the transition to VF and during established VF. Comparisons were then performed as described below.
In an attempt to gain insight into the physiology underlying observed differences, the conduction velocities of activation wave fronts during VF were determined. Although conduction in experimental myocardial infarction may be discontinuous and three-dimensional activation in this model cannot be excluded, we used a combination of activation time and vector methods to define regions in which an “apparent” conduction velocity could be calculated. Conduction velocities were calculated for any wave front that was at least 5.0 mm in each dimension. For each such wave front, the direction of activation was determined, and the conduction velocity parallel to the wave front was determined using the assigned activation times. If the “apparent” conduction velocity between two adjacent sites within a given wave front was greater than 0.8 m/s, then these sites were considered to represent “epicardial breakthrough” rather than being part of a discrete activation wave front. The lowest “apparent” conduction velocity that was permitted was 0.05 m/s, since this degree of slowing of conduction has been demonstrated previously in this model.23 24 In addition, if gross discontinuities of activation times were noted, then conduction velocity for that wave front was not determined. The mean conduction velocity over 10 consecutive cycles of VF was determined for the onset of VF and for established VF.
The presence of differences between the anterior and lateral walls and among animals with subacute healing infarction, chronic healing infarction, and no infarction was evaluated using two-way ANOVA.15 21 22 This analysis was performed for wave front number and area (as assessed by the visual method), VF cycle lengths, and conduction velocities. In addition, differences among the three groups and between sites in the correlation coefficients derived from the linking analysis (at each distance) were also determined using two-way ANOVA. Because the values for r are not normally distributed, a Fisher’s z-transformation was performed before the ANOVA.15 21 The analyses were performed separately for the onset of VF and for established VF. In addition, within each group, differences in wave front size and number, the extent of linking, and VF cycle length between the onset of VF and established VF were assessed using paired t tests.21 22
Data are expressed as mean±SD. A value of P<.05 was considered to be significant. Post hoc analyses were performed using Scheffe’s F procedure.21 22 All statistical analyses were performed using commercial software (statview, Abacus Concepts) on a Macintosh IIfx computer (Apple Computer).
Activation Wave Fronts
Fig 1⇓ shows a plot of the maximum vectors recorded from the anterior left ventricular epicardium for one cycle of VF in an animal without infarction (group 1). Eight separate activation wave fronts are identified. Twenty-three of 91 vectors did not fall into a broad activation wave front. Similar plots are shown in Fig 2⇓ for an animal with subacute healing infarction (group 2) and in Fig 3⇓ for an animal with chronic healing infarction (group 3). In groups 2 and 3, these represent recordings from the infarction zone. In Fig 3⇓, an area within the right center portion of the plaque is without vectors; this represents an area where local activations were absent and therefore vectors could not be created. Based on a qualitative visual assessment, in the examples shown, the wave fronts appear to be smaller in the animals without infarction compared with the animals with infarction.
The mean wave front area was significantly greater among animals with infarction compared with control animals over both the anterior and lateral walls (Table 1⇓). However, there was no significant difference between the anterior and lateral walls within a given group. These findings were seen both during the onset of VF and during established VF. Within each group, the mean wave front area was significantly greater at the onset of VF compared with established VF (P<.05 for all comparisons).
The results of the visual assessment of wave front number are summarized in Table 2⇓. The mean number of activation wave fronts per given area of epicardium (7 cm2, the area of the recording plaque) was less in animals with infarction than in noninfarcted animals over both the anterior and lateral walls. There was no significant difference between the anterior and lateral walls within a given group. These findings were seen both during the onset of VF and during established VF. Within each group, the mean wave front number was significantly greater at the onset of VF compared with established VF (P<.05 for all comparisons).
For animals with infarction (groups 2 and 3), significant linking was present over the anterior and lateral walls at distances of 2.5 mm, 5.0 mm, and 7.5 mm both at the onset of VF and during established VF. However, in noninfarcted animals (group 1), significant linking was present only at distances of 2.5 and 5.0 mm over both the anterior and lateral walls (both at the onset of VF and during established VF). The mean value of the correlation coefficient describing the extent of linking at 7.5 mm was significantly lower over both the anterior and lateral walls among noninfarcted animals compared with animals with infarction (Fig 4⇓). These findings suggest that the mean wave front dimension in animals without infarction is less than 7.5 mm, whereas in animals with infarction, it is greater than 7.5 mm. In addition, they suggest differences in the extent of organization between animals with and without infarction. Within each group at a given recording site, at each distance evaluated the mean value for r was significantly greater during the onset of VF compared with established VF (P<.05), confirming previous work that organization decreases over the first 5 to 10 seconds of VF.15
VF Cycle Lengths
Demonstrated in Fig 5A⇓ are unipolar recordings from the anterior epicardial surface and lateral epicardial surface in an animal without infarction (group 1) during the first 600 ms of VF. Local activations are marked by the vertical lines, and the activation time intervals are indicated between the vertical lines. For this example, the mean activation time interval or VFCL (VF cycle length) over the anterior wall was 95±5 ms. The mean activation time interval or VFCL over the lateral wall was 93±5 ms.
Demonstrated in Fig 5B⇑ are unipolar recordings from the anterior epicardial surface and lateral epicardial surface in an animal with subacute healing infarction (group 2) during the first 600 ms of VF. Again, in each recording, the local activations are marked by the vertical lines and the activation time intervals are indicated between the vertical lines. For this example, the mean activation time interval or VFCL over the anterior wall was 135±8 ms. The mean activation time interval or VFCL over the lateral wall was 125±7 ms.
Finally, demonstrated in Fig 5C⇑ are unipolar recordings from the anterior epicardial surface and lateral epicardial surface in an animal with chronic healing infarction (group 3) during the first 600 ms of VF. Again, in each recording the local activations are marked by the vertical lines and the activation time intervals are indicated between the vertical lines. For this example, the mean activation time interval or VFCL over the anterior wall was 144±14 ms. The mean activation time interval or VFCL over the lateral wall was 126±12 ms.
The VF cycle length was significantly longer in the presence of either subacute or chronic infarction than in noninfarcted animals over both the anterior and lateral walls, suggesting that the presence of infarction affects the VFCL throughout the heart (Table 3⇓). Within each group, the VFCL was not significantly different at the onset of VF compared with established VF (P>.05 for all comparisons).
Fig 6⇓ demonstrates an example of the calculation of mean conduction velocity in an animal with chronic healing infarction (group 3). Fig 6A⇓ shows the maximum vectors for a single cycle of VF. The activation wave fronts are outlined by dashed lines. Fig 6B⇓ shows the local activation times for the same cycle of VF. The electrode sites corresponding to the vectors in the activation wave fronts shown in Fig 6A⇓ are also outlined by dashed lines. For each activation wave front, the conduction velocity was calculated, and the mean conduction velocity for all wave fronts was then determined. For this example, the mean conduction velocity for this cycle of VF was 0.26 m/s; the mean conduction velocity for 10 consecutive cycles of VF was 0.25 m/s.
The data on conduction velocities of activation wave fronts are summarized in Table 4⇓. By two-way ANOVA, the mean conduction velocities during VF were significantly slower over both the anterior and lateral walls among animals with infarction compared with control animals.
In the animals with subacute healing infarction (group 2), the predominant histological pattern that was seen from the anterior wall was the combination of necrotic tissue with a surviving epicardial layer (Fig 7A⇓). This pattern was seen under 38% to 87% of the area under the recording plaque. The mean thickness of the surviving subepicardial layer was 2.1±1.0 mm. The mean overall wall thickness in the region beneath the recording plaque was 12.5±2.4 mm. Thus, the surviving layer occupied a mean of 16±8% of wall thickness. However, within animals, there was wide variability in the thickness of the surviving subepicardial layer. The mean minimum width of the surviving subepicardial layer averaged over the six animals was 0.2±0.1 mm. In these regions, the surviving subepicardial layer was only several cell layers thick. Unipolar and bipolar electrograms from the regions with a surviving subepicardial layer showed distinct local activations (Fig 8A⇓). The remainder of the sections beneath the recording plaque consisted of a combination of areas of complete transmural infarction (range, 11% to 62%; Fig 7B⇓) and areas of completely viable tissue (range, 0% to 34%). The areas of complete infarction corresponded to the areas that did not exhibit local electrical activation (Fig 8B⇓).
In four of the five animals with chronic healing infarction (group 3), the predominant histological pattern that was seen from the anterior wall was also the combination of necrotic tissue with a surviving epicardial layer. This pattern was seen under 21% to 46% of the area under the recording plaque. The mean thickness of the surviving subepicardial layer was not significantly different from that in group 2 (3.6±1.3 mm). The mean overall wall thickness in the region beneath the recording plaque was also similar to group 2 (13.8±3.5 mm). Thus, myocardial infarction occupied a mean of 30±16% of wall thickness. The final animal in this group also had a small infarct, with only 10% of the area under the recording plaque containing both viable and necrotic tissue; the majority of the sections under the plaque were completely viable, although small areas of completely transmural infarction were seen.
The mean surviving subepicardial layer thickness and percentage transmurality of infarction were then correlated with the observed mean conduction velocities and wave front sizes within each animal. There was no significant correlation seen between either of these histological parameters and either conduction velocity or wave front size (P>.05 for all comparisons).
The major finding of this study is that the presence of subacute or chronic healing myocardial infarction affects several characteristics of VF. In the presence of infarction, the mean size of activation wave fronts is larger and the number of activation wave fronts over a given area of myocardium is smaller over both infarcted and normal tissue. In addition, the mean cycle length of VF is prolonged in the presence of myocardial infarction. Slower conduction velocities within the activation wave fronts were also noted in animals with infarction. These findings demonstrate that the presence of infarction affects the characteristics of VF throughout the heart. Thus, the presence of infarction may be important to the development and maintenance of VF. Finally, the results of the linking analysis confirm that the degree of organization of epicardial activation and wave front size during VF is affected by the presence of infarction.
Although prior studies have suggested that organized activation wave fronts exist, especially early in VF, none have systematically examined VF in a chronic healing infarct model.10 11 12 13 14 15 25 The cycle length and the number and size of activation wave fronts during VF have also not been well studied. In a preliminary report, Hillsley et al26 evaluated the number and size of activation wave fronts during VF in a porcine model. Using activation time mapping methods, they were able to demonstrate that wave fronts as large as 100 mm2 (or greater) were present during VF. Later work from this group using a spatial correlation function again suggested the presence of wave fronts from 16 to 100 mm2.27 Again, however, these studies did not address wave front characteristics in an infarct model. Two previous defibrillation studies in the canine model have evaluated VF cycle length; however, neither of these studies was performed in an infarct model and therefore did not provide insight into possible differences between VF in hearts with and without infarction.28 29
In this study, vector mapping was used in combination with standard activation time mapping methods to assess several characteristics of epicardial electrical activation during VF in a chronic infarct model, in a subacute healing infarct model, and in animals without infarction. Five- to 7-day and 8-week infarctions were selected for study because the cellular electrophysiological findings differ in the two models and because of the evolution of arrhythmias after clinical myocardial infarction.30 31 In 5- to 7-day experimental infarctions, a decrease in the rate of action potential upstrokes is to some extent responsible for the decrease in conduction velocity30 32 (although changes in cell-to-cell coupling may also be responsible), whereas a change in cell-to-cell coupling appears to cause most of the slowing of conduction velocity in 8-week-old infarctions.
The first major finding of this study was that the size of activation wave fronts is greater and the number of activation wave fronts is fewer in the presence of subacute or chronic myocardial infarction compared with control animals. There are several possible explanations for these findings. One possibility is that, as suggested by most prior studies, VF is due to reentry in all three models but that the size, number, and/or characteristics of reentrant circuits differ between infarcted hearts and normal hearts.11 13 14 However, we cannot exclude the possibility that the mechanisms of VF differ between normal and infarcted hearts because we did not determine the mechanism of VF in the different models.
The second major finding of this study was that the cycle length of VF was longer in animals with subacute healing infarction or chronic healing infarction compared with control animals. The prolongation of cycle length in the presence of infarction may be due to slowing of conduction, prolongation of refractoriness, or a change in the characteristics of reentrant circuits. Previous work in the canine infarct model has demonstrated that there is slowing of conduction and prolongation of local refractoriness as infarction heals.30 31 In addition, recordings of monophasic action potentials during VF in both the canine model as well as humans have suggested that VF is characterized by incomplete repolarization and the absence of a resting membrane potential, implying that differences in refractoriness may be a cause for observed differences in cycle length.33 34 We were able to demonstrate slowing of conduction during VF in animals with infarction. However, since we did not define complete reentrant circuits or activation wave fronts emanating from a focal site, we cannot determine whether slowing of conduction was the only mechanism for the slowing of cycle length. Refractoriness during VF was not assessed in this study. Similar to our observations regarding wave front size and number, we observed that in the presence of infarction VF cycle length was longer over both normal and infarcted tissue (compared with control animals without infarction). However, within the two infarct groups, the mean cycle length was shorter over healthy tissue compared with infarcted tissue.
By defining broad activation wave fronts based on vector criteria, we were able to estimate conduction velocities during VF over the anterior and lateral walls of the left ventricle. Although the presence of transmural conduction altering conduction velocities could be completely excluded, prior studies have suggested that activation within a wave front of vectors of similar direction represents epicardial and not transmural conduction.18 19 Conduction velocities in the present study were determined at the center of activation wave fronts. When wave fronts are highly curved as they may be in some regions in some arrhythmia models, the effective conduction velocity may differ.35 These areas were not examined in the present study. Conduction velocities in normal canine epicardium during ventricular pacing generally range from 0.2 to 0.7 m/s, depending on the relation to fiber orientation.18 However, in experimental myocardial infarction, conduction velocities as low as 0.05 m/s have been identified.23 24 Because fiber orientation may be disorganized with chronic myocardial infarction, we did not correlate conduction velocity with fiber orientation in the present study.32 Mean conduction velocities during VF in control animals were 0.29 to 0.31 m/s, which is somewhat lower than the mean conduction velocity (at a random relation to fiber orientation), which might be expected in normal epicardium. Mean conduction velocities in animals with infarction were 0.23 to 0.26 m/s, which was significantly slower than conduction velocities in normal animals.
Taken together, the findings of this study imply that prolongation of the lines of conduction block due to either prolongation of refractoriness or anatomic substrate differences and the slowing of conduction are responsible for the differences in the characteristics of VF seen in the setting of myocardial infarction. If the mechanism of these findings was simply slowing of conduction, then one would expect that while cycle length might prolong, wave front dimensions should either remain unchanged (in the presence of an automatic focus or an anatomically defined reentrant circuit) or become smaller (in the presence of a functionally defined reentrant circuit). On the other hand, in the presence of a functional reentrant circuit, prolongation of refractoriness, by increasing the length of the lines of block, could lead to both prolongation of VF cycle length and an increase in wave front dimensions. Similarly, changes in the anatomic substrate in the setting of infarction also could lead to differences in the anatomic features of the putative reentrant circuits (ie, longer lines of block), resulting in the same findings. A final possibility is that there are fewer reentrant circuits during VF in the setting of infarction, allowing a given wave front to activate a larger area before colliding with another wave front. Further studies will be required to distinguish among these possibilities.
The final finding of this study is that the degree of organization of electrical activation during VF appears to be greater in animals with infarction. In our previous study, we did not detect any differences between animals with and without infarction in the extent of organization.15 However, this was based on linking analysis performed at a distance of only 2.5 mm. In the current study, there were again no differences noted at distances of 2.5 mm; in addition, there were no differences noted at 5.0 mm. However, at greater distances (7.5 mm), organization as defined by spatial linking appeared to be greater in animals with either subacute or chronic healing infarction compared with normal animals. These findings are consistent with the results of the visual analysis of wave front dimensions, which suggested the presence of fewer and larger wave fronts in the presence of infarction. At each distance evaluated in each model, the degree of spatial linking during VF decreased over time so that linking was less after 5 to 10 seconds of VF than just after its onset. This is consistent with the finding that the dimensions of activation wave fronts also decrease over time, confirming that the degree of VF “organization” decreases as VF persists.
The implication of the observations that wave front size and number, VF cycle length, and organization of activation are different throughout the heart in the presence of infarction is that the presence of myocardial infarction affects the characteristics of VF throughout the heart, not only over areas of infarction. Furthermore, this suggests that the initiation and maintenance of VF in hearts with previous myocardial infarction may be directly related to the presence of infarction rather than being an epiphenomenon. The mechanisms for the effect on VF throughout the entire heart were not defined in the present study. Electronic interactions during VF, changes in myocardial electrophysiological properties due to remodeling, or pressure effects on refractoriness36 are all potential explanations for the differences in lateral wall VF characteristics in infarcted animals as compared with control animals.
In the present study, we did not demonstrate a dependence of VF characteristics on the degree of underlying histological abnormalities. As noted in prior studies,32 37 38 the majority of the anterior wall in the present study consisted of subendocardial and midmyocardial infarction with a spared subepicardial border zone. Some prior studies32 have shown only a few surviving subepicardial cell layers in this infarct model, and the animals in the present study had a surviving layer that averaged 3 to 4 mm in depth. However, there was substantial heterogeneity in the depth of the surviving subepicardial layer in the present study, and in some regions, infarct depth was similar to that described by Ursell et al.32 The mean percent infarct in the present study was similar to the finding of two other studies.37 38 The lack of a relation between the depth of the surviving subepicardial layer and activation during VF suggests that the effects of infarction on VF are similar within the range of anatomic heterogeneity seen in the present study.
There are several limitations that must be considered in interpreting our findings. One is failure to evaluate activation in three dimensions and to define either reentrant circuits or foci of activation. We have made the assumption that the number of activation wave fronts reflects the number of foci or reentrant circuits present during VF and that the wave front size reflects the size of (putative) reentrant circuits. A second limitation is that we did not measure local refractory periods during VF. Direct measurements of refractoriness at multiple sites during VF is not feasible. However, indirect measurements using monophasic action potential recordings could be performed. In addition, although we cannot exclude the possibility that three-dimensional conduction affected our interpretation of apparent epicardial conduction velocities, the uniform orientation of vectors within the defined activation wave fronts and the relatively consistent differences in activation times between adjacent sites that we used for our analysis suggest that these estimates are accurate. Our method for quantifying “organization” was use of a linear regression model. It is possible that a more complex model rather than a simple linear model would better define organization.15 A final limitation is that we did not extend our analysis beyond the initial 6 to 7 seconds of VF. It is possible, as recently suggested by other investigators,27 that the degree of organization may sometimes change with more prolonged episodes of VF. However, other work from the same group has given somewhat conflicting data with regard to changes in the number and size of activation wave fronts during prolonged episodes of VF.25 26
The finding that some characteristics of VF throughout the entire myocardium are affected by the presence of myocardial infarction suggests that the nature of the underlying heart disease must be considered in experimental and clinical studies of VF. More importantly, further understanding the nature of VF has potential therapeutic implications. It may be important to consider the underlying heart disease in selecting therapies that prevent VF, such as antiarrhythmic drugs, or that terminate VF, such as implantable defibrillators. Clarification of these issues will require further study.
This study was supported by grant HL-40667 from the National Institutes of Health, Bethesda, Md.
Reprint requests to Alan H. Kadish, MD, Northwestern Memorial Hospital, 250 E Superior, Suite 524, Chicago, IL 60611.
- Received November 2, 1994.
- Revision received January 3, 1995.
- Accepted January 22, 1995.
- Copyright © 1995 by American Heart Association
Bashe WJ, Baba N, Keller MD, Geer JC, Anthony JR. Pathology of atherosclerotic heart disease in sudden death: the significance of myocardial infarction. Circulation. 1975;52(suppl III):III-63-III-69.
Schwartz CJ, Gerrity RG. Anatomical pathology of sudden unexpected cardiac death. Circulation. 1975;52(suppl III):III-18-III-26.
Davies MJ. Pathological view of sudden cardiac death. Br Heart J. 1981;45:88-96.
Marchlinski FE. Ventricular tachycardia associated with coronary artery disease. In: Zipes DP, Rowlands DJ, eds. Progress in Cardiology. Philadelphia, Pa: Lea & Febiger; 1988:231-253.
Sano T, Tsuchihashi H, Shimamoto T. Ventricular fibrillation studied by the microelectrode method. Circ Res. 1958;6:41-46.
Ideker RE, Klein GJ, Harrison L, Smith WM, Kasell J, Reimer KA, Wallace AG, Gallagher JJ. The transition to ventricular fibrillation induced by reperfusion after acute ischemia in the dog: a period of organized epicardial activation. Circulation. 1981;63:1371-1379.
Chen PS, Wolf PD, Dixon EG, Danieley ND, Frazier DW, Smith WM, Ideker RE. Mechanism of ventricular vulnerability to single premature stimuli in open-chested dogs. Circ Res. 1988;62:1191-1209.
Janse MJ, van Capelle FJL, Morsink H, Kleber AG, Wilms-Schopman F, Cardinal R, Naumann d’Alnoncourt C, 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. 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.
Chen PS, Wolf PD, Melnick SD, Danieley ND, Smith WM, Ideker RE. Comparison of activation during ventricular fibrillation and following unsuccessful defibrillation shocks in open-chest dogs. Circ Res. 1990;66:1544-1560.
Damle RS, Kanaan NM, Robinson NS, Ge Y, Goldberger JJ, Kadish AH. Spatial and temporal linking of epicardial activation directions during ventricular fibrillation in dogs: evidence for underlying organization. Circulation. 1992;86:1547-1558.
Richards DA, Blake GJ, Spear JF, Moore EN. Electrophysiologic substrate for ventricular tachycardia: correlation of properties in vivo and in vitro. Circulation. 1984;69:369-381.
Kadish AH, Spear JF, Levine JH, Hanich RF, Prood C, Moore EN. Vector mapping of myocardial activation. Circulation. 1986;74:603-615.
Kadish A, Balke W, Levine JF, Moore EN, Spear JF. Activation patterns in healed experimental myocardial infarction. Circ Res. 1989;65:1698-1709.
Kanaan N, Jenkins J, Levy R, Kadish A. Computerized vector mapping of myocardial activation. Comput Cardiol. 1990;94.
Kleinbaum DG, Kupper LL, Muller KE. Applied Regression Analysis and Other Multivariable Methods. Boston, Mass: PWS-KENT Publishing Co; 1988.
Glantz SA, Slinker BK. Primer of Applied Regression and Analysis of Variance. New York, NY: McGraw-Hill Book Co; 1990.
Dillon SM, Allessie MA, Ursell PC, Wit AL. Influences of anisotropic tissue structure on reentrant circuits in the epicardial border zone of subacute canine infarcts. Circ Res. 1988;63:182-206.
Kadish A, Spear J. Identification of conduction block in cardiac muscle: in vitro observations in canine epicardium. Cardiovasc Res. 1994;28:259-269.
Johnson EJ, Rollins DL, Wolf PD, Smith WM, Ideker RE. Mechanism of ventricular fibrillation as mapped with 524 closely spaced simultaneously recorded epicardial electrodes. Circulation. 1992;86(suppl I):I-820. Abstract.
Hillsley RE, Simpson EV, Rollins DL, Wolf PD, Smith WM, Ideker RE. Effect of isoproterenol on ventricular fibrillation wave fronts mapped by 506 closely spaced epicardial electrodes. J Am Coll Cardiol. 1993;21(suppl A):53A. Abstract.
Witkowski FX, Penkoske PA, Plonsey R. Mechanism of cardiac defibrillation in open-chest dogs with unipolar DC-coupled simultaneous activation and shock potential recordings. Circulation. 1990;82:244-260.
Zhou X, Kinsley SB, Wolf PD, Rollins DL, Smith WM, Ideker RE. Action potential prolongation by biphasic and monophasic shocks during fibrillation. PACE Pacing Clin Electrophysiol. 1991;14:689-699.
Spear JF, Michelson EL, Moore EN. Cellular electrophysiologic characteristics of chronically infarcted myocardium in dogs susceptible to sustained ventricular tachyarrhythmias. J Am Coll Cardiol. 1983;14:1099-1110.
Lesh MD, Spear JF, Moore EN. Myocardial anisotropy: basic electrophysiology and role in cardiac arrhythmias. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. Philadelphia, Pa: WB Saunders Co; 1990:364-376.
Ursell PC, Gardner PI, Albala A, Fenoglio JJ, Wit AL. Structural and electrophysiological changes in the epicardial border zone of canine myocardial infarcts during infarct healing. Circ Res. 1985;56:436-451.
Russel DC, Smith HJ, Oliver MF. Transmembrane potential changes and ventricular fibrillation during repetitive myocardial ischemia in the dog. Br Heart J. 1979;42:88-96.
Swartz JF, Jones JL, Fletcher RD. Characterization of ventricular fibrillation based on monophasic action potential morphology in the human heart. Circulation. 1993;87:1907-1914.
Pertsov A, Davidenko J, Salomonsz R, Baxter W, Jalife J. Spiral waves of excitation underlie reentrant activity in isolated cardiac muscle. Circ Res. 1993;72:631-650.
Calkins H, Maughan WL, Weisman HF, Sugiura S, Sagawa K, Levine JH. Effect of acute volume load on refractoriness and arrhythmia development in isolated, chronically infarcted canine hearts. Circulation. 1989;79:687-697.
Kavanagh K, Kabas J, Rollins D, Melnick S, Smith W, Ideker R. High-current stimuli to the spared epicardium of a large infarct induce ventricular tachycardia. Circulation. 1992;85:680-698.
Denniss A, Richards D, Waywood J, Yung T, Kam C, Ross D, Uther J. Electrophysiological and anatomic differences between canine hearts with inducible ventricular tachycardia and fibrillation associated with chronic myocardial infarction. Circ Res. 1989;64:155-166.