High-Frequency Periodic Sources Underlie Ventricular Fibrillation in the Isolated Rabbit Heart
Abstract—The mechanism(s) underlying ventricular fibrillation (VF) remain unclear. We hypothesized that at least some forms of VF are not random and that high-frequency periodic sources of activity manifest themselves as spatiotemporal periodicities, which drive VF. Twenty-four VF episodes from 8 Langendorff-perfused rabbit hearts were studied using high-resolution video imaging in conjunction with ECG recordings and spectral analysis. Sequential wavefronts that activated the ventricles in a spatially and temporally periodic fashion were identified. In addition, we analyzed the lifespan and dynamics of wavelets in VF, using a new method of phase mapping that enables identification of phase singularity points (PSs), which flank individual wavelets. Spatiotemporal periodicity was found in 21 of 24 episodes. Complete reentry on the epicardial surface was observed in 3 of 24 episodes. The cycle length of discrete regions of spatiotemporal periodicity correlated highly with the dominant frequency of the optical pseudo-ECG (R2=0.75) and with the global bipolar electrogram (R2=0.79). The lifespan of PSs was short (14.7±14.4 ms); 98% of PSs existed for <1 rotation. The mean number of waves entering (6.50±0.69) exceeded the mean number of waves that exited our mapping field (4.25±0.56; P<0.05). These results strongly suggest that ongoing stable sources are responsible for the majority of the frequency content of VF and therefore play a role in its maintenance. In this model, multiple wavelets resulting from wavebreaks do not appear to be responsible for the sustenance of this arrhythmia, but are rather the consequence of breakup of high-frequency activation from a dominant reentrant source.
Ventricular fibrillation (VF) is the leading cause of sudden cardiac death in the industrialized world. It is estimated that VF results in ≈300 000 deaths each year in the United States alone.1 Nevertheless, despite years of intensive research, its mechanism remains poorly understood. Recent studies from our laboratory2 3 4 5 and others6 7 8 9 have shown that VF is not entirely random and that complex spatiotemporal organization underlies this enigmatic phenomenon. Yet there is still some controversy as to whether fibrillation in the ventricles is due to (1) multiple unstable wavelet reentry,10 11 (2) the destabilization of 3-dimensional scroll waves and their breakup,12 13 14 15 or (3) the fibrillatory conduction from a single or a small number of ongoing reentrant circuits.2 16 Computer simulations have suggested that single or multiple 3-dimensional rotor(s) (ie, scroll waves) may be the underlying mechanism of VF.12 13 17 18 In 1995, Gray et al2 provided direct experimental evidence that at least some cases of VF can be due to a single rotor moving rapidly throughout the heart. However, in many studies, rotors are not consistently observed on the epicardial surface.15 19 20
We hypothesized that at least some forms of VF are not random and that high-frequency periodic sources of activity manifest themselves as spatiotemporal periodicities, which drive VF. To this end, we used frequency analysis, in conjunction with high-resolution video imaging and phase mapping,4 (1) to demonstrate spatiotemporal periodicity (STP) of wavefronts, (2) to determine its contribution to the frequency content of VF, and (3) to elucidate the relative importance of multiple wavelets in the maintenance of VF. Overall, the results from our study strongly suggest that a single or a small number of sources of periodic activity are responsible for the maintenance of VF and that fibrillatory conduction away from such sources results in multiple short-lived wavelets.
STP: a minimum of 4 sequential waves propagating into or in our field of view (1) from the same location (edge, breakthrough, or rotor) and (2) with a timing that varies by no more than ±1 frame (4.17 ms) from a mean period.
Phase singularity points (PSs): the break points in the wavefront that act as organizing centers around which rotation occurs (for a more technical definition see below and Reference 44 ).
Wavebreak: break that occurs when any wave of excitation hits a functional or anatomical obstacle and PSs are formed.
Wavelet: a segment of excitation wavefront, regardless of size, bounded on its ends by PSs, a PS and a boundary, or 2 boundaries.
Rotor: a wave of excitation rotating around a PS for 1 or more cycles.
Breakthrough: a wavefront that emerges to the epicardium from deeper muscle layers; ie, it does not arise from an epicardial fractionation or collision event, or propagate from the edge of the mapping region.
Fibrillatory conduction: breakup of waves of activity emanating from a source, where the tissue surrounding the source is not activated in a 1:1 manner.
Materials and Methods
Langendorff-Perfused Rabbit Heart and Optical Mapping
The experimental protocol is similar to those published elsewhere.3 23 Briefly, isolated Langendorff-perfused rabbit hearts were continuously perfused through a cannula in the aortic root with warm (36.5±1°C) oxygenated Tyrode’s solution. The voltage-sensitive dye di-4-ANEPPS (2 mL; 15 μg/mL) was perfused through the coronary arteries for 1 to 2 minutes. VF was induced by burst pacing. Light at 535 nm was shone on the epicardial surface of the heart; the emitted fluorescence was acquired by a charge-coupled device camera at 240 frames per second.24 To reveal the signal, the background fluorescence was subtracted from each frame and low-pass spatial filtering was applied. All optical recordings were ≈3 to 4 seconds in duration. An adjustable glass wall was used to gently compress and restrain the heart, and no electromechanical uncoupling agents were used.
Electrocardiographic and Frequency Analysis
Global volume–conducted bipolar ECGs were obtained at a sampling rate of 2400 Hz. Pseudo-ECGs 22 were constructed from optical recordings by integrating the transmembrane fluorescence signal over the entire mapped region of the ventricle. Fast Fourier transformation (FFT) was performed on both the global bipolar ECGs and pseudo-ECGs using Welch’s25 method.
Two-Dimensional Phase Maps
We quantified the patterns of wave propagation during VF using phase mapping,4 a recently developed technique, which highlights the formation of wavebreaks and the resulting PSs. In Figure 1A⇓, the fluorescence (F) changes recorded by a single camera pixel (asterisk in Figure 1C⇓) during VF are presented as a function of time. In Figure 1B⇓, the fluorescence of this pixel at time t, F(t), was plotted against the fluorescence of the same pixel offset by a time interval τ=2 frames. A cyclic return map of F(t) versus F(t–τ) was constructed. This allowed a new parameter, the phase θ(t), to be defined as the angle of the coordinate [F(t), F(t–τ)] around the mean fluorescence for that given pixel, with values between –π and π, represented as a continuous color scheme from red to purple. After the transformation, a new phase, θ(t), movie was produced including all pixels, whereby the upstroke of the action potential, and hence the activation wavefront, corresponded to the color green, whereas the plateau of the action potential corresponded to the colors blue and purple. The refractory tail of the action potential corresponded to the colors red and yellow. A PS was defined at the point where all phases converged. Figure 1C⇓ is a single snapshot (phase map) of a phase movie with 3 wavelets, each bounded by a PS or a PS and a boundary.
In 8 experiments, 24 optical recording episodes were analyzed for STP. Only 10 randomly chosen episodes were analyzed for lifespan analysis; all 24 episodes were analyzed for waves entering and leaving.
Correlation of frequencies was performed using simple linear regression analysis. Comparisons were also performed using standard ANOVA. A P<0.05 was considered to be statistically significant.
An expanded Materials and Methods section is available online at http://www.circresaha.org.
To evaluate the frequency content of episodes of VF, we first performed spectral analysis on global and local measures of ventricular electrical activity (ie, global bipolar ECG and pseudo-ECG, respectively). For all episodes analyzed, the dominant frequency (DF) of the global bipolar electrograms (EGs) ranged between 9.0 and 19.2 Hz, with a mean of 13.2±3.4 Hz (cycle length, 76±20 ms). The DF of the pseudo-ECGs ranged between 7.0 and 20.1 Hz, with a mean of 12.6±4.0 Hz (cycle length, 79±25 ms).
A representative example is shown in Figure 2⇓, with its corresponding optical data. The pseudo-ECG (Figure 2A⇓) shows rapid, irregular activity, and its FFT (Figure 2B⇓) has a dominant peak at 12.2 Hz, with a smaller peak at 10.4 Hz. The global bipolar EG for this same episode is illustrated in Figure 2C⇓, with its corresponding FFT shown in Figure 2D⇓. In this single example, a DF peak is seen at 12.1 Hz, and 2 smaller peaks are seen at 11.0 and 13.5 Hz. For the purposes of this study, the dominant frequencies of the pseudo-ECGs and global bipolar ECGs can be considered identical, as they are within the given spectral resolution of our system. Subsequent examination of the optical recordings from the same episode revealed the source of this DF in the form of STP (Figure 2E⇓). Figure 2E⇓ shows 4 sequential isochrone maps of activation for this same episode of VF over a period of 333 ms (see horizontal bars in Figure 2A⇓ and 2C⇓). All 4 maps contain a very similar periodic activity, with a wavefront emerging repetitively and periodically from the upper left corner. This spatiotemporal periodic activity continued throughout the entire 4 seconds of recording. The cycle length of these periodic waves was ≈83 ms, which corresponded to the DF peak of the FFT (12.1 Hz) seen in both the pseudo-EGs and global bipolar EGs, clearly demonstrating that these periodic waves were responsible for the DF of the arrhythmia in this specific example. As can be seen in activations 3 and 4 of Figure 2E⇓, these periodic waves did not always propagate across the mapping field undisturbed. More often than not, these waves did not activate the rest of the mapped region in a 1:1 manner after entering the mapping field but followed complex pathways with multiple spatially distributed conduction delays and sites of block (ie, fibrillatory conduction). Figure 2F⇓ shows a single pixel recording from the region where the STP enters the mapped region (black asterisk in Figure 2E⇓). The FFT of this pixel recording (Figure 2G⇓) shows a single narrow peak at 12.2 Hz, which corresponds to the DF of both the optical pseudo-ECG and bipolar ECG, in addition to the cycle length between this spatiotemporal periodic activity. At a site distant from this region (white asterisk in Figure 2E⇓; map at t=249 ms), fibrillatory conduction results in activity shown in Figure 2H⇓. At this pixel location (white asterisk), the DF (Figure 2I⇓) is 10.3 Hz, with a prominent secondary peak at 12.3 Hz. This spectral pattern correlates well with patterns of activation at this location, where an approximate sequence of 6:5 could be demonstrated. Thus, the representative isochrone map of activation patterns (Figure 2E⇓), in conjunction with the 2 peaks seen in Figure 2I⇓, shows that the faster peak (12.3 Hz) represents the input frequency and the slower peak represents the output frequency at that site. Spatial distribution of many such patterns of complex input:output relations is the hallmark of fibrillatory conduction. Overall, the data presented strongly suggest the presence of a periodic source outside the mapping region at 12.1 Hz that is driving this complex arrhythmia.
After having demonstrated the direct relationship between the frequency of periodic activity and the global VF DF, we then proceeded to quantify such a relationship in all episodes. Figure 3⇓ shows the correlation between the frequency of the periodic activity (STP) and the DF of the pseudo-EG. As shown in Figure 3A⇓, in 24 episodes from 8 hearts, we obtained a strong correlation of R2=0.75 between these 2 variables. In Figure 3B⇓, a similar correlation for the DF of the global bipolar EG and the frequency of the STP region produced an R2 value of 0.79. In many cases (52%), the frequency of the STP region was higher than the DF of the pseudo-ECG but was still represented as a significant peak in the frequency spectra. As seen in Figure 2⇑, this is most likely reflective of the fact that activity from the periodic site was not propagating in a 1:1 fashion to the majority of the mapping field, thus representing breakup of periodic activity and fibrillatory conduction. Overall, the high correlation between the periodic sources and the global DF strongly suggests that these spatiotemporal periodic wavefronts are the source for the majority of the frequency content of VF.
Sources and Nature of Periodic Activity
In 21 of 24 episodes, STP was found. In many cases, however, sources (ie, rotors and/or breakthroughs) alternated with STP occurring from the edge of the mapping field. That is, a source would persist for several activations and either drift out of the mapped region or terminate, in which case STP from the edge would occur. These sources would reappear later on in the movies with the concomitant termination of STP from the edge. For those select episodes (n=16) in which the source of STP was clearly identified, 12.5% of the time (2 episodes) this source occurred exclusively in the form of a rotor (Figure 4⇓), 81.25% of the time (13 episodes) it occurred exclusively in the form of a breakthrough (Figure 5⇓), and 6.25% of the time (1 episode), the source was seen to alternate between a periodic breakthrough and a rotor (Figure 6⇓). In specific cases, these sources were seen to persist during the entire episode. As can be seen in Figure 4B⇓, a rotor acted as a source of 2 distinct periodic waves. This rotor persisted for 52 rotations, and both waves emanating from it were seen to exit the field of view in a spatially and temporally similar fashion throughout the episode. As seen in the series of 2-dimensional phase maps illustrated in Figure 4A⇓, this mother rotor (represented by PS 1) gave rise to multiple short-lived wavelets, in addition to the 2 broad wavelets that eventually left the field of view. Two-dimensional phase mapping better enabled us to study the generation of multiple wavelets. At t=0 ms, this source wave, bounded by PS 1 and PS 2, is clearly seen as a reentrant phenomenon on the ventricular epicardium. At t=8 ms, wavelet 1-2 breaks up into wavelets 1-3 and 4-2, each bounded by a PS. At t=16 ms, wavelet 4-2 persists, and wavelet 1-3 breaks up into wavelets 3-6 and 1-5. At t=24 ms, wavelet 4-2 continues to move toward the left ventricular free wall, whereas wavelet 6-3 extinguishes itself on refractory tissue (red). Wavelet 1-5 breaks up into 2 distinct wavelets (not shown, as it is outside of the mapping field), appropriately named wavelet 1-boundary and wavelet 5-boundary. Finally, at t=32 ms, wavelet 4-2 moves upward and to the right, and wavelet 5-boundary moves downward and to the left. Both waves eventually exit the mapping region. Wavelet 1-boundary, also characterized as mother rotor PS 1, continues to rotate, acting as a source of new wavelets for the remainder of the movie. In Figure 4B⇓, the isochrone map for this activation sequence (ie, three fourths of a rotation) depicted in the phase maps t=0 through t=32 ms is shown.
High-frequency periodic sources were also observed to occur in the form of a breakthrough (Figure 5⇑). Two-dimensional phase maps are shown illustrating one such source that persists throughout the movie. At t=0 ms, no depolarization wavefront is seen on the ventricular epicardial surface. At t=4 ms, wavelet 1 to 2, bounded on each side by PS 1 and PS 2, creeps into the mapping region from the midmyocardial or endocardial muscle layers of the right ventricle. Four milliseconds later, the full wavelet emerges and begins to propagate toward the left ventricle. At t=20 ms and t=28 ms, wavelet 1 to 2 now moves toward the apex and the left ventricular free wall. At t=36 ms, wavelet 1 to 2 breaks up into wavelets 2 to 3 and 1 to 4. Eventually wavelets 2 to 3 and 1 to 4 exit the mapping field (not shown).
The isochrone maps depicted in Figure 6A⇑ illustrate an example in which the source alternated between a breakthrough and a rotor. In Figure 6B⇑ and 6C⇑, respectively, the pseudo-EG from the same episode and its FFT are shown. At t=1.483 seconds (first dashed line from left in Figure 6B⇑), a periodic breakthrough is shown that persisted for 43 activations acting as a source of new wavelets. The frequency of activation of this source accelerated and decelerated. Most commonly, however, the cycle length at the breakthrough site was ≈61 ms, which corresponded highly to the DF of the FFT (16.4 Hz; Figure 6C⇑). The cycle length of this breakthrough varied from ≈50 to ≈69 ms, corresponding to peaks in the range of 14.5 to 20.2 Hz. At t=3.554 seconds (second dashed line from left in Figure 6B⇑), the breakthrough transformed into a rotor, which persisted for 10 rotations and meandered in a cycloid fashion, finally ending up in the location depicted at t=3.912 seconds. The rotation period at t=3.554 seconds (≈50 ms) showed a small increase to 54 ms at t=3.912 seconds. Both of these rotation periods are represented by major peaks in the FFT at 20.2 and 18.5 Hz, respectively. At t=3.912 seconds, the rotor continued for 2 more activations before finally transforming back to a breakthrough at t=4.096 seconds. A reasonable interpretation of these results is that the overall arrhythmia resulted from high-frequency activation by a single nonstationary scroll wave, of which the filament changed orientation repeatedly with respect to the ventricular epicardium, and most likely resulted in the source alternating between a rotor and a breakthrough.
Figure 7⇓ illustrates STP from the edge of the mapping region alternating with STP in the form of a breakthrough. In the isochrone map shown in Figure 7A⇓, at t=1.708 seconds (first dashed line from left in Figure 7B⇓), a wavefront activates the apex of the left ventricle and proceeds toward the base. This activity occurred in a spatially and temporally similar fashion for 14 activations. The cycle length of this activity was ≈100 ms, which was equal to the DF of the FFT (9.9 Hz; Figure 7C⇓) of the optical pseudo-EG (Figure 7B⇓). At t=3.108 seconds (second dashed line from left in Figure 7B⇓), STP from the edge stopped, and STP was then seen to occur in the form of a breakthrough. This source persisted for 7 activations at a cycle length of ≈100 ms and gradually drifted toward the upper right corner of the mapped region. Finally, at t=3.808 seconds (third dashed line from left in Figure 8B⇓), the breakthrough drifted out of the field of view, and STP (cycle length ≈100 ms) occurred once again from the edge.
Analysis of Wavelet Lifespan During Sustained VF
To investigate the role of multiple wavelets in the sustenance of this complex arrhythmia, we measured the lifespan of PSs and hence indirectly measured the lifespan of wavelets during VF. Figure 8⇑ shows the lifespan of PSs in milliseconds and rotations. The lifespan distribution was skewed to the left, with 51% of PSs lasting only 8 ms or less. The mean lifespan of PSs was short, 14.7±14.4 ms, with a range that varied from 4.17 (1 frame) to 100 ms. Because the average rotation period of a rotor was ≈80 ms, 98% of PSs were found to exist for <1 rotation. The majority of these short-lived PSs were seen to be the result of breakup from broad spatiotemporal periodic waves (ie, fibrillatory conduction; see Figure 4⇑).
Waves Entering and Leaving the Field of View
We hypothesized that if wavebreaks and their resulting wavelets were not maintaining this arrhythmia, the number of waves entering the mapping field should exceed or be equal to the number of waves leaving it. In Figure 9A⇓, the mean number of waves entering the mapping region was 6.50±0.69, whereas the mean number of waves leaving was 4.25±0.56 (P<0.05) for all experiments. In Figure 9B⇓, the ratio of entering to leaving waves (E:L ratio) for each episode varied from 0.6 to 5.0 with a mean of 1.92±1.03 (hatched bar). In 21 of 24 episodes (85.0%), the E:L ratio was >1; 3 episodes (15.0%) had an E:L ratio <1. In 2 of the 3 cases in which the E:L ratio was <1, periodic breakthroughs were observed. In the other example, an ongoing rotor was seen acting as a source of new wavelets. It is important to note that in all cases in which a periodic source of wavelets (ie, a rotor or breakthrough) was present, the number of wavelets leaving the mapping region exceeded those entering. However, as only the last 50 frames (≈200 ms) of each episode were analyzed because of the laborious time of analysis, periodic sources were not always present. Hence, during this time frame, the number of wavelets leaving did not always exceed those entering. In those episodes in which we could see a periodic source of activity (n=16), this source was often intermittent in nature, alternating with STP from the edge (Figure 7⇑).
In this study, we have demonstrated organization during VF in the form of sequences of wave propagation that activated the ventricles in a spatially and temporally similar fashion over time. Furthermore, the frequency of such periodic activity corresponded highly to the dominant peak in both the global bipolar EG and the optical pseudo-ECG, strongly suggesting that the source(s) of this spatial and temporally periodic activity is the dominant source maintaining VF in this model. The lifespan of PSs, and hence indirectly the lifespan of wavelets during VF, was short; 98% of them lasted <1 average rotation period. In the vast majority of cases, the number of waves entering the mapping region exceeded the number leaving, despite a large number of wavebreaks and new wavelet formation. This suggests that the breaking of waves, although effective in producing a large number of wavelets in the mapping region, was not a robust mechanism of long-term new wavelet production in the ventricles as a whole. Overall, these results strongly suggest that periodic sources of activity are responsible for the majority of the frequency content of VF and are therefore important for the maintenance of this arrhythmia. Fibrillatory conduction from such sources results in the formation of short-lived wavelets and the complex activation patterns characteristic of VF.
Two-dimensional phase mapping is a powerful technique to analyze wave propagation in VF. However, certain constraints need to be considered. Irregularities in small-amplitude fluorescence signals near the phase singularity may cause limit cycle trajectories in the return map, which do not envelop the center of the map (Figure 3B⇑ online; see http://www.circresaha.org). Arbitrarily choosing the mean fluorescence as the center of the map minimized the occurrence of such trajectories. On the basis of sample recordings (not shown), we have estimated the error in localizing the PSs to be ≈0.5 mm. Therefore, owing to the highly coupled nature of the cardiac tissue, it is difficult to conceive of erroneous PSs produced as a result of this phenomenon. Rarely, wavelets can survive annihilation of their PSs during failed vortex shedding,26 and therefore some wavelets may have gone undetected by our analysis. We only mapped 40% of the epicardial surface of the ventricles and thus have no information about the midmyocardial and the endocardial muscle layers. The high preponderance of epicardial breakthroughs in our experiments, in addition to recent computer simulations performed on a realistic 3-dimensional representation of the heart,12 13 17 18 strongly suggest that intramural reentry may be the source for VF, but the hypothesis requires experimental validation. As a whole, we do not believe these limitations in our technique affected the results of this study. Lastly, our experiments were performed on an isolated Langendorff-perfused rabbit heart. The relevance of these data to human VF remains to be determined.
Previous Studies on VF
Several theories regarding the underlying mechanisms of VF have been proposed, as follows: (1) multiple unstable wavelet reentry, where multiple wavelets were thought to move randomly throughout cardiac tissue and a critical number of wavelets was required for the sustenance of this arrhythmia10 11 ; (2) 3-dimensional scroll waves and their destabilization and breakup12 13 14 15 ; and (3) uninterrupted periodic activity of discrete reentrant sites with the subsequent breakup of waves, ie, fibrillatory conduction.2 16 One important factor that is unaccounted for by the first 2 hypotheses is the presence of organized activity in the form of persistent STP, as demonstrated in most of our episodes of VF. In fact, some recent studies have also demonstrated spatial organization during VF. Bayly et al8 showed a correlation between recordings obtained by closely spaced unipolar electrodes. Furthermore, Damle et al7 previously showed spatial and temporal linking of epicardial activation patterns in a canine model of VF. More recently, elegant studies by Huang et al27 and Rogers et al20 quantified organization during VF using a concept very similar to STP. They used a parameter termed “multiplicity,” which measures the number of different pathways in an overall activation. Altogether, these studies provide strong evidence that VF is not an entirely random phenomenon. Our study is the first to demonstrate organization during VF in the form of STP and to relate such activity to the frequency content of VF.
Mechanism Underlying the Periodicity
It has generally been accepted that the mechanisms underlying the maintenance of VF are reentrant in nature.28 For those select episodes (n=16) in which the source of STP was observed within our field of view, 12.5% of the time (2 episodes) this source occurred exclusively in the form of a rotor (Figure 4⇑), 81.25% of the time (13 episodes) it occurred exclusively in the form of a breakthrough (Figure 5⇑), and 6.25% of the time (1 episode) the source was seen to alternate between a rotor and a periodic breakthrough (Figure 6⇑). These results are compatible with the idea that complex 3-dimensional vortex-like reentry17 (ie, scroll waves) is the most likely underlying mechanism of VF in this model. According to contemporary hypotheses, 3-dimensional reentry is organized around a central filament that forms the rotation axis of reentry. The evolution of the filament influences the dynamics of the arrhythmia, and the filament orientation determines the nature of the epicardial activation patterns. Our results do not provide information about activity inside the ventricular wall. However, as breakthrough patterns were the most common form of periodic activity seen, we speculate that, more often than not, the filament was probably not aligned perpendicularly to the epicardial surface. Using plunge electrodes, Chen et al14 were the first to demonstrate pairs of mirror-image scroll waves, of which the filaments were aligned perpendicularly to the ventricular epicardial surface (ie, transmural reentry). In such a case, the 2-dimensional manifestation of the scroll wave on the epicardial surface would be a spiral wave. On the other hand, it is not difficult to perceive instances in which the filament is not aligned perpendicularly to the epicardial surface (ie, intramural reentry). Recent work by Berenfeld and Pertsov17 has provided a mechanistic explanation for the greater prevalence of intramural reentry over transmural reentry, whereby the scroll wave filament shows a tendency to align along the local myocardial fiber orientation and thus is not perpendicular to the epicardial surface. According to that study,17 stable intramural reentry would be manifest as periodic breakthroughs of activity on the epicardial surface, thus providing an explanation for the predominance of breakthroughs over rotors as a source of STP.
Duration of Spatiotemporal Patterns and Number of Apparent Sources
In certain episodes, STP could be seen throughout the entire 3- to 4-second recording. However, in most episodes it was transient (4 to 51 activations). In all cases, when the wavefront entered the mapping field, breakup of activity occurred, with independent short-lived wavelets being produced. In the vast majority of episodes, periodic activity would last for 4 or more activations, thereafter it would stop, and then it would return later on at the same location and with the same frequency (ie, intermittent STP). This could reflect propagation from a single rotor with intermittent block. In addition, it was noted that in some episodes there was spatiotemporal periodic activity that stopped, propagated into our mapping field from a different spatial orientation, and then once again returned to its original spatiotemporal pattern of propagation. This phenomenon could be explained by epicardial activation patterns from a single source via multiple select routes of ongoing 1:1 propagation or propagation from a rotor that is drifting back and forth. Alternatively, it could also be explained by 2 distinct independent sources of activity at a similar frequency; yet the first 2 scenarios seem more plausible in this case.
It is important to note that the frequency of periodic activity was not always the same. As clearly demonstrated in Figure 6⇑, the rotation period of the source (ie, rotor in this case) may change as it meanders, and thus periodic activity emanating from this source may exit the mapped region with slight changes in frequency. In the faster and more complex episodes of VF, such a phenomenon was not uncommon. However, in the slower and more organized episodes of VF, such an occurrence was never found (Figure 7⇑). Nevertheless, in the rare case in which there was more than one distinct site of STP in the same episode, the frequency of these spatiotemporal periodic regions was usually the same (Figure 7⇑), which might reflect sites of activation in a 1:1 manner from a mother source. Finally, as the rotation period of a mother rotor may accelerate or decelerate, multiple peaks may be seen in the FFT (Figure 6C⇑), thus representing different rotation periods of that rotor at different points in time. Varying degrees of intermittent block of waves propagating from the periodic sources (Figure 2⇑) resulting in spatially distributed input:output frequency relations (eg, 6:5, 4:3) is also a mechanism for the multiple distinct peaks seen in the frequency spectra of VF.
Are Multiple Wavelets Important in the Maintenance of VF?
To answer this question, we measured the lifespan of PSs, and hence indirectly that of wavelets. The lifespan of PSs was short; 98% of them lasted <1 rotation. According to the multiple wavelet hypothesis, cardiac fibrillation is maintained by spontaneous wavebreaks that constantly generate randomly wandering daughter wavelets.10 11 If the lifespan of wavelets is very short, then these wavelets will have a decreased chance to give rise to new wavelets, in a dynamic equilibrium that maintains fibrillation, as originally postulated by Moe et al10 in 1964. In fact, if we had sampled at a rate faster than 240 Hz (≈4 ms), it is quite probable that the lifespan of many wavelets would have been even shorter, giving further credence to our hypothesis.
To test the relevance of the multiple wavelet hypothesis to VF in our model, we measured the number of wavelets entering and leaving our mapping field as an estimate of the number of new wavelets produced. If multiple randomly occurring wavelets are the “engine” that maintains fibrillation, there should be an equal/neutral or positive balance between the number of wavelets created and the number destroyed so that a critical number of wavelets still remains. In such a schema, the number of wavelets leaving our mapping field should be equal to or greater than the number of wavelets entering that field. In our experiments, despite a large number of wavebreaks, multiple wavelets were in general constrained to the mapping field (E:L ratio >1), except when a periodic source of activity (ie, rotor or breakthrough) was found. This suggests that the breaking of waves, although effective in producing large numbers of wavelets in the mapping region, was not a robust mechanism of new wavelet production for the ventricle as a whole.
The question then becomes, if multiple wavelets are not maintaining this complex arrhythmia, then what is? Our data are the first to show that spatiotemporally periodic waves are important contributors to the frequency content of VF and hence are important for the maintenance of VF. Moreover, these spatiotemporal periodic sources were always cycling at the fastest frequency so that the rest of the tissue could not keep up in a 1:1 manner. The coexistence of short-lived wavelets with periodic activity mostly in the form of breakthroughs strongly suggests that fibrillatory conduction away from relatively stable intramural scroll waves17 underlies VF in our experimental model.
This work was supported in part by a Grant PO1-HL-39707 from the National Heart, Lung, and Blood Institute; an American Heart Association New York State Affiliate Fellowship awarded to R.M. and O.B.; and a North American Society for Pacing and Electrophysiology Fellowship awarded to A.C.S. In addition, we thank Jiang Jiang, Fan Yang, and Clara Wu for their technical assistance.
- Received August 5, 1999.
- Accepted October 13, 1999.
- © 2000 American Heart Association, Inc.
Myerburg RJ, Kessler KM, Interian A Jr, Fernandez P, Kimura S, Kozlovskis PL, Furukawa T, Bassett AL, Castellanos A. Clinical and experimental pathophysiology of sudden cardiac death. In: Zipes DP, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 1st ed. Philadelphia, Pa: WB Saunders Co; 1990:666–678.
Gray RA, Jalife J, Panfilov AV, Baxter WT, Cabo C, Davidenko JM, Pertsov AM. Mechanisms of cardiac fibrillation. Science. 1995;270:1222–1225.
Gray RA, Jalife J, Panfilov AV, Baxter WT, Cabo C, Davidenko JM, Pertsov AM. Nonstationary vortex-like reentry as a mechanism of polymorphic ventricular tachycardia in the isolated rabbit heart. Circulation. 1995;91:2454–2469.
Damle RS, Kanaan NM, Robinson NS, Ge YZ, 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.
Witkowski FX, Kavanagh KM, Penkoske PA, Plonsey R, Spano ML, Ditto WL, Kaplan DT. Evidence for determinism in ventricular fibrillation. Physiol Rev Lett. 1995;75:1230–1233.
Moe GK. On the multiple wavelet hypothesis of atrial fibrillation. Arch Int Pharmacodyn. 1962;149:1–2.
Chen PS, Wolf PD, Dixon EG, Danieley ND, Frazier DW, Smith WM, Ideker RE. Mechanism of ventricular vulnerability to single premature stimuli in open chest dogs. Circ Res. 1988;62:1191–1209.
Berenfeld O, Jalife J. Purkinje-muscle reentry as a mechanism of polymorphic ventricular arrhythmias in a 3-dimensional model of the ventricles. Circ Res. 1998;82:1063–1077.
Rogers JM, Huang J, Smith WM, Ideker RE. Incidence, evolution, and spatial distribution of functional reentry during ventricular fibrillation in pigs. Circ Res. 1999;84:945–954.
Pertsov AM, Davidenko JM, Salomonsz R, Baxter W, Jalife J. Spiral waves of excitation underlie reentrant activity in isolated cardiac muscle. Circ Res. 1993;72:631–640.
Skanes AC, Mandapati R, Berenfeld O, Davidenko JM, Jalife J. Spatiotemporal periodicity during atrial fibrillation in the isolated sheep heart. Circulation. 1998;98:1236–1248.
Mandapati R, Asano Y, Baxter WT, Gray R, Davidenko JM, Jalife J. Quantification of the effects of global ischemia on the dynamics of ventricular fibrillation in the isolated rabbit heart. Circulation. 1998;98:1688–1696.
Welch PD. The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans Audio Electroacoust. 1967;AU-15:70–73.
Cabo C, Pertsov AM, Davidenko JM, Baxter WT, Gray RA, Jalife J. Vortex shedding as a precursor of turbulent electrical activity in cardiac muscle. Biophys J. 1996:70:1105–1111.
Epstein AE, Ideker RE. Ventricular fibrillation. In: Zipes DF, Jalife J, eds. Cardiac Electrophysiology: From Cell to Bedside. 2nd ed. Philadelphia, Pa: WB Saunders Co; 1995:927–933.