| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Cardiac Rhythm Management Laboratory, 1Department of Medicine, 2Department of Biomedical Engineering, 3Department of Biomathematics and Biostatistics, and 4Department of Physiology, University of Alabama, Birmingham, Alabama 35294-0019
Submitted 11 August 2003 ; accepted in final form 29 October 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 min after induction when activation briefly became more organized. Thus a quantitative description of activation does not uniformly decrease as VF progresses, but undergo rapid changes and exhibit a brief interval of increased organization after 1 min of VF. Further studies are warranted to determine whether these changes, particularly the increased organization of stage iii, have clinical consequences, such as an alteration in defibrillation efficacy.
cardiac mapping; electrophysiology
VF is highly dynamic with activation patterns that change quickly. To better understand the nature of the activation sequences and how they change during VF, activations should be analyzed continuously and described by several variables that represent the important characteristics of the wavefronts. Accordingly, we (20, 33–35) have recently developed a set of robust algorithms that analyze large volumes of electrical mapping data automatically and quantify many of the electrophysiological aspects of VF.
The objective of this study was to describe quantitatively the activation fronts beneath a large electrode array covering 20% of the canine epicardium and to characterize continuous temporal changes during the first 10 min of VF. We applied cluster analysis to divide the activation patterns into stages. We hypothesized that the quantification of the evolution of VF obtained from electrical activation would not be the same as that from mechanical contraction obtained by Wiggers and would provide additional information toward better understanding the mechanisms of VF maintenance. While the first 20 s of VF, as analyzed in our previous study, is the time during which shocks from implantable defibrillators are delivered, most external shocks for sudden cardiac arrest are delivered after several minutes of VF. Although rare, survival is occasionally possible even after 10 min of VF (38). Therefore, the knowledge about how VF changes during the first 10 min is important.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Animal preparation. Six 27–32 kg (30 ± 4 kg, means ± SD) mongrel dogs were anesthetized with 25 mg/kg iv thiopental sodium, intubated, and mechanically ventilated with 2 to 3% isoflurane in 100% oxygen. Core body temperature, arterial blood pressure, arterial blood gases, surface ECG lead II, and serum electrolytes were monitored and maintained within normal ranges throughout the study.
The heart was exposed through a median sternotomy and supported in a pericardial sling. A 504-electrode (24 x 21) plaque was sutured to the anterior lateral right ventricular and anterior left ventricular epicardium. Care was taken to avoid occluding epicardial vessels. The electrodes were 2 mm apart and, in total, covered 20.16 cm2, which is 20% of the epicardial surface. This spatial resolution has been shown to be appropriate for mapping epicardial activation patterns during VF (3). A stainless steel wire (Cooner Wire; Chatsworth, CA) was attached to the right leg for use as a ground reference for the unipolar recordings from the plaque.
VF was induced with 30-V, 60-Hz alternating current through a defibrillation catheter in the right ventricular apex. The first three episodes of VF were mapped for 20 s before a rescue shock was given. The fourth run of VF was continually mapped for >10 min. The first 20-s VF episode served as control data in another study (21), whereas the last VF episode was analyzed in this study.
Mapping and data acquisition. Unipolar electrograms were amplified, band-pass filtered between 0.5 and 500 Hz, sampled at 2 kHz, and recorded continuously during VF for 600 s. Each VF episode was divided into 600 epochs of 1 s each for analysis.
Quantitative analysis of VF activation. The activation patterns in each epoch were quantified with a series of descriptors based on a decomposition of VF into its constitutive wavefronts. This decomposition is described in detail elsewhere (35). Briefly, all 504 electrograms were differentiated using a five-point digital filter. The discrete samples of the differentiated signals (one for each electrode every 0.5 ms) were stored in a three-dimensional array: two of the dimensions correspond to the dimensions of the plaque and the third to time. Samples for which the rate of voltage change (dV/dt) was less than –0.5 V/s were considered active. This threshold was chosen previously with the use of a cryoablation technique to distinguish local activation from deflections due to distant activations (6) and is commonly used to detect activations during VF (26, 28, 45).
Individual wavefronts were identified and isolated by grouping together active samples that were adjacent in space and time. During VF, wavefronts frequently fragment into multiple new wavefronts or collide with other wavefronts. In our definition of a wavefront, a wavefront ends when it fragments, at which time the two or more resultant wavefronts begin. Conversely, when multiple wavefronts collide and coalesce, the original wavefronts end, and the resulting wavefront or wavefronts begin.
From this decomposition of the overall activation pattern, we computed 12 descriptors for each of the 600 VF epochs in each animal. Descriptor 1 is the total number of wavefronts, which is the number of wavefronts completely or partially within the mapped region (35). Descriptor 2 is the mean area swept out by the wavefronts. The area swept by a wavefront is the number of electrodes that recorded an activation from that wavefront, multiplied by the area of the grid represented by each electrode (4 mm2 for our 2-mm inter-electrode spacing) (35). When reentry was present, electrodes were only counted once. Thus the maximum area that could be swept out by a wavefront is the area of the plaque. Descriptor 3 is block incidence, or the fraction of wavefronts that was terminated without fractionating, colliding with other wavefronts, or propagating out of the mapped region (20). Descriptor 4 is breakthrough/focal incidence. This is the logical opposite of block. A wavefront that appears in the mapped region without arising from a fractionation or collision event or without propagating in from an edge of the mapped region either has broken through to the epicardium from deeper layers or has arisen de novo (20). Descriptor 5 is the multiplicity of the activation pattern. Multiplicity is a parameter that measures the number of distinct wavefront activation pathways (33); it is a descriptor of the spatiotemporal complexity of the activation sequences. Descriptor 6 is repeatability of the activation pattern (20). This descriptor expresses, in an average sense, how many wavefronts propagated along each of the distinct families of pathways indicated by the multiplicity parameter. For example, if an activation pattern contains eight wavefronts, four of which follow one pathway, and four of which follow another, then the multiplicity of the pattern is two and the repeatability is four. If each pathway represents part or all of a reentrant circuit, then repeatability expresses the number of cycles for which reentry persists. Descriptor 7 is mean propagation speed of the wavefronts. We estimated the propagation velocity of a wavefront by computing the location of the centroid of the wavefront. The velocity of the centroid approximates the velocity of the wavefront (20). Descriptor 8 is fractionation; the fraction of wavefronts that fragmented into two or more child wavefronts (35). Descriptor 9 is collision. This is the fraction of wavefronts that collide and coalesce with one or more other wavefronts (35). Descriptor 10 is negative peak dV/dt of VF activations, or the mean rapidity of the peak downslope of activation, which may be related to the ability of the activation front to excite adjacent tissue. Descriptor 11 is peak frequency of VF activation. The VF electrograms were analyzed by a fast Fourier transform analysis. The peak frequency of the VF spectrum at each epicardial site was computed with a frequency resolution of 0.49 Hz and averaged over the entire array. Descriptor 12 is incidence of reentry. This descriptor differs from the other 11 in that it was computed for 150 consecutive 4-s epochs (34) rather than 600 1-s epochs. For each epoch, wavefronts were grouped into families consisting of wavefronts connected temporally by fragmentation or collision events. If a family contained a sequence of wavefronts that activated tissue more than once, it was deemed reentrant. The incidence of reentry is the fraction of wavefront families containing a reentrant sequence of wavefronts.
Statistical analysis. k-Means clustering analysis (32) was employed for partitioning the 600 1-s epochs into subgroups (descriptor 12, incidence of reentry, was excluded from this analysis because it was computed for a different set of epochs). To partition epochs x1, x2,..., xn where n = 600 k cluster centers a1, a2,..., ak were chosen to minimize
 |
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Examples of VF activation sequences during the five stages are shown in Fig. 1. The broad picture of the dynamics are the following: stage i, regular and repeatable activation pattern (Fig. 1A); stage ii, large spatial and temporal variation (Fig. 1B); stage iii, return to a highly repeatable activation pattern (Fig. 1C); stage iv, irregular and fragmented wavefronts; and stage v, small, slow moving, short-lived wavefronts. The time course of the multiple parameters was consistent for five of the six dogs studied. One dog did not show the increased organization of stage iii.
|
Wavefront number, fractionation, and collision. There were rapid temporal fluctuations in activation (Fig. 2). The number of wavefronts in stage i increased rapidly (Fig. 3). The number of wavefronts maintained a high level during stage ii with large spatial and temporal fluctuations. The maximum difference for the number of wavefronts between sequential 1-s epochs in a single animal was 69. The number of wavefronts continuously declined during stages iv and v. Also, the variation in the number of wavefronts between successive 1-s epochs decreased as VF continued. The changes in the number of fractionations and collisions were similar to that of the number of wavefronts (not shown).
|
|
|
|
|
|
|
Propagation speed and peak frequency. Wavefronts propagated quickly at the beginning of VF but slowed as VF continued, except for an increase in conduction velocity during stage iii, when the number of wavefronts decreased and repeatability increased (Fig. 5A).
The changes in peak frequency were relatively smooth with a rapid increase during stage i, reaching a maximum during stage ii and decreasing during stages iii and iv (Fig. 5B). However, the peak frequency increased at the end of stage iv and throughout most of stage v. The high peak frequency during stages iv and v appeared to be caused by small rapid undulations in the signal rather than the larger slower complexes that may have represented true activation on the epicardium (Fig. 1, D and E).
Breakthrough/focal and block incidence. The changes in breakthrough/focal and block incidence over time were similar (Fig. 6). There were fewer wavefronts that blocked or broke through to the epicardium during the first three stages. The incidence of block and breakthrough increased during stage iv as VF continued.
Area swept out by wavefronts. The pattern of evolution for the area swept out by VF wavefronts was similar to that of propagation speed. In general, the wavefronts activated smaller epicardial areas as VF continued. Consistent with the decreased number of wavefronts and increased propagation speed during stage iii, the area swept out increased in this time period (Fig. 7A).
Peak negative dV/dt. The peak dV/dt was most negative during the first three stages after VF induction and then gradually became less negative during stages iv and v (Fig. 7B).
Reentrant circuits. Only 9.4% of wavefront families formed reentrant circuits during the first stage of VF on the epicardium. Throughout the 600 s of VF, 16.3% of wavefront families were involved in reentrant pathways with the highest rate, 23%, during the second, third, and early fourth stages of VF (Fig. 8).
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 min after VF induction and lasting 30 s, 2) a high variation for most descriptors from second to second during VF, especially during the second stage, and 3) an increased incidence of reentry beginning 40 s after VF induction and lasting for 2 min. To our knowledge, the present study is the first to make continuous observations of dynamic VF changes using multiple descriptors that quantify several aspects of the electrophysiological features during VF. As discussed below, our five stages based on observation of the electrically mapped epicardial activation patterns during VF are not completely consistent with Wiggers' four stages based on observation of epicardial movement.
Comparison of present study with previous studies. Several studies have been performed to observe the evolution of VF. In 1940, Wiggers (40) reported that four stages of electrically induced VF were visible by cinematography. The first was the undulatory or tachysystolic stage, which consisted of three to six undulatory contractions lasting <1 s. The second stage, called the convulsive incoordination stage, lasted for 15–40 s. This stage was characterized by more frequent waves of contractions that sweep over smaller sections of the ventricles. In this stage, synchronous contractions of the ventricles were lost. The third stage, that of tremulous incoordination, lasted 2 to 3 min, during which the surface of the ventricles appears to be broken up into progressively smaller and smaller independently contracting areas. The final stage was that of atonic fibrillation, with complete failure of contractility.
On the basis of electrical wavefront activation, we partitioned VF into five stages, which somewhat interlace with Wiggers' four stages. Our stage i is longer than Wiggers' stage I, in which the clustering algorithm not only included the VT-like activation period, but also included the first part of Wiggers' stage II. Although a wavebreak is present during this time (Fig. 6) (27), it is fairly constant, as is the number of different activation patterns as indicated by multiplicity (Fig. 4). Our stage ii covers the remainder of Wiggers' stage II and part of stage III. The number of wavefronts was greatest during this period. Several of the quantitative variables exhibited a plateau throughout stage ii. Stage iii was a period of increased organization as indicated by a decrease in repeatability (Fig. 4), an increase in conduction velocity (Fig. 5), and an increase in the area swept out by each wavefront (Fig. 7). Wiggers' stages do not include such a period of increased organization.
Electrical and optical mapping studies have demonstrated that reentry is present during Wiggers' stages I and II VF but does not involve the majority of wavefronts (8, 9, 14, 18, 27, 44). Our previous study showed that 30% of wavefront families exhibit reentry (34). With the use of a robust reentrant analysis algorithm for continuous VF analysis in the present study, we demonstrated a large variation from second to second in the incidence of reentrant excitation during VF. The present study found that the highest incidence of reentry occurred 40 s after induction, which is beyond the time period examined by most previous studies.
Spectral correlation and coherence analysis (4, 22, 36), as well as nonlinear dynamical approaches (4, 17), have been used to study VF organization. Recently, the application of the theory of wave propagation in excitable media to the study of VF (43) with high-resolution mapping techniques (11, 18) has enabled investigators to demonstrate that self-organization of nonlinear electrical waves with both deterministic and stochastic components exists in VF. Altogether, these studies provide strong evidence that VF is not an entirely random phenomenon. In this study, we continuously traced activation pathways and quantified the similarity of activation sequences and found that many wavefronts traveled similar pathways during VF for approximately the first 5 min of VF (stage i–iv, Fig. 4).
Activation patterns during VF time course. The evolution of activation patterns during the different stages of VF could be caused by several factors, including 1) alterations independent of changes in the intrinsic cardiac state such as accommodation and restitution of electrophysiological characteristics, 2) alterations in the cardiac state caused by changes in autonomic tone due to hypotension and cardiac stretch, and 3) alterations in the cardiac state caused by ischemia due to lack of coronary perfusion. The quick dynamic increase in wavefronts and other descriptors during stage i of VF may be independent of any underlying change to the intrinsic state of the heart. This possibility is supported by computer modeling studies of excitable systems, in which the underlying properties of the media are held constant (10, 13, 15, 25, 29, 30).
There was a rapid change in wavefront conduction and organization between 63 and 86 s of VF in stage iii, which returned those descriptors to near the stage i values. This may be related to the "dip" in extracellular K+, which was shown by Wilde et al. (42) during global ischemia in isolated guinea pig hearts. A decrease in K+ efflux combined with an unaltered active K+ influx caused by the Na+/K+ pump and by a release of endogenous catecholamine (42) during this particular period would greatly reduce the extracellular K+, which in turn would increase the resting potential and improve propagation by increasing excitability. The decreased multiplicity and the increased repeatability, velocity, and area swept out by each wavefront during stage iii support the assumption that activation temporarily improves. However, the dip in extracellular K+ was observed after 10 min of global ischemia in isolated guinea pig hearts. Because ischemic myocardium exhausts its energy stores more rapidly during VF than during sinus rhythm because of the rapid activation rates in VF (24), it is possible that the dip in extracelluar K+ would occur much earlier during VF. However, this interpretation is highly speculative and additional experiments are needed to verify the cause of the increased organization of stage iii.
The continuous deterioration in metabolic state caused by ischemia as VF progresses may increase asynchronous depolarization and decrease the size of the activation fronts, indicated by the decreased area swept out by the wavefronts (Fig. 7). By stage v, 70% of wavefronts appear on the epicardium in a breakthrough/focal pattern and 70% of wavefronts are extinguished by conduction block (Fig. 6). The Purkinje fibers are more tolerant to ischemia than is the working myocardium (2) and triggered activity has been shown to occur in Purkinje fibers exposed to high K+ and acidosis (31, 39). Thus it is possible that after 5 min of VF, most activation fronts breakthrough to the epicardium after arising intramurally from Purkinje fibers, either by triggered activity or reentry within the specialized conduction system.
Clinical relevance. This study shows that there are rapid temporal and spatial changes in the number of wavefronts and all other parameters during VF. This may explain in part the probability of defibrillation success. Although the exact mechanism for defibrillation is still unclear, evidence suggests that a certain minimum strength shock electric field must be achieved in the region of the heart exposed to the weakest electrical field intensity to halt activation in this region and avoid reinitiation of VF (23). The characteristics and timing of activation in the weakest electrical field region immediately before the shock may be critical for the success of defibrillation (12, 19). When a shock is delivered during the interval when this region is vulnerable to the stimulation of new activation fronts that reinitiate VF, the shock will fail to defibrillate. Thus defibrillation may require a relatively lower shock energy to defibrillate when VF is more organized so that the entire region exposed to the weak shock field spends a smaller fraction of time in its vulnerable period.
Evidence of a critical time for resuscitation success during prolonged VF has been provided by Geuze et al. (16). After 60–90 s of VF in the dog, the time of the first rapid fall in the number of wavefronts in our study (Fig. 3), spontaneous recovery of blood pressure after successful defibrillation was rare. The declining number of wavefronts and other concordant electrophysiological changes may mirror a decline in cardiac mechanical function, raising the possibility that if these electrophysiological changes can be detected in the ECG they may be used to predict the state of cardiac mechanical function following defibrillation during cardiac resuscitation. Furthermore, the scaling structure of the ECG, a measure of organization, has been investigated to estimate the duration of VF (7), which could provide a useful stratifying variable in clinical studies of outcome after sudden cardiac death for which the onset of VF was not witnessed. Quantifying the organization and reentry circuits during VF could contribute to a better understanding and interpretation of the scaling exponent measured from the surface ECG. Because our results indicate that organization does not monotonically decrease throughout the time course of VF.
Increasingly sophisticated computer models of normal and abnormal cardiac electrophysiological states are being developed. The quantitative results of our study provide a benchmark for the mathematical simulation of ventricular fibrillation.
Limitations. The limitations for extracellular electrical mapping studies from epicardial arrays described in our previous study (21) are also present in this study. We mapped only 20% of the epicardial area and performed no transmural mapping so that complete activation pathways could not always be determined and all reentrant circuits were probably not detected. The evolution of VF described in this study used anesthetized dogs with normal hearts, which may differ from clinical VF, which most often occurs in cardiac ischemia or heart failure patients without anesthesia. The evolution of VF in diseased hearts may differ from that in normal hearts.
|
ACKNOWLEDGMENTS |
|---|
Present address of K. P. Singh: Department of Biostatistics, University of North Texas Health Science Center, 3500 Camp Bowie Blvd., Fort Worth, TX 76107.
GRANTS
This study was supported in part by National Heart, Lung, and Blood Institute Grants HL-28429 and HL-64184.
|
FOOTNOTES |
|---|
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.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Masse, T. Farid, P. Dorian, K. Umapathy, K. Nair, J. Asta, H. Ross, V. Rao, E. Sevaptsidis, and K. Nanthakumar Effect of global ischemia and reperfusion during ventricular fibrillation in myopathic human hearts Am J Physiol Heart Circ Physiol, December 1, 2009; 297(6): H1984 - H1991. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Dosdall, P. B. Tabereaux, J. J. Kim, G. P. Walcott, J. M. Rogers, C. R. Killingsworth, J. Huang, P. G. Robertson, W. M. Smith, and R. E. Ideker Chemical ablation of the Purkinje system causes early termination and activation rate slowing of long-duration ventricular fibrillation in dogs Am J Physiol Heart Circ Physiol, August 1, 2008; 295(2): H883 - H889. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. A. Trayanova The Long and the Short of Long and Short Duration Ventricular Fibrillation Circ. Res., May 23, 2008; 102(10): 1151 - 1152. [Full Text] [PDF]
|
|
|
|
|
|
L. Li, Q. Jin, J. Huang, K.-A. Cheng, and R. E. Ideker Intramural Foci During Long Duration Fibrillation in the Pig Ventricle Circ. Res., May 23, 2008; 102(10): 1256 - 1264. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Huang, K.-A. Cheng, D. J. Dosdall, W. M. Smith, and R. E. Ideker Role of maximum rate of depolarization in predicting action potential duration during ventricular fibrillation Am J Physiol Heart Circ Physiol, October 1, 2007; 293(4): H2530 - H2536. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. F. Huizar, M. D. Warren, A. G. Shvedko, J. Kalifa, J. Moreno, S. Mironov, J. Jalife, and A. V. Zaitsev Three distinct phases of VF during global ischemia in the isolated blood-perfused pig heart Am J Physiol Heart Circ Physiol, September 1, 2007; 293(3): H1617 - H1628. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. W. Kay, G. P. Walcott, J. D. Gladden, S. B. Melnick, and J. M. Rogers Lifetimes of epicardial rotors in panoramic optical maps of fibrillating swine ventricles Am J Physiol Heart Circ Physiol, October 1, 2006; 291(4): H1935 - H1941. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
