INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

IT IS KNOWN THAT in a given heart, shocks of threshold defibrillation strength may or may not succeed in terminating (defibrillation) ventricular fibrillation (VF) (6, 24). The lack of a preset, fixed defibrillation threshold (DFT) (i.e., a shock below threshold will never defibrillate and one above it will always defibrillate) led to the proposal of an alternative approach, "the probability of defibrillation" for a particular defibrillation shock (1, 4). The mechanism of the probabilistic nature of the DFT remains poorly understood. It is not known why in the same heart a given electrical shock strength can be successful in some episodes and fail in others.

It was suggested that shock outcome depends on the electrical state of the heart at the instant of the shock (2), VF regularity (8, 11), and VF voltage amplitude (12). Recent optical mapping studies (28) suggest that failed defibrillation does not result from incomplete termination of VF wavefronts (32) or incomplete resynchronization of repolarization (7, 20). The demonstration of the presence of a postshock optical equivalent of an isoelectric window after near-threshold failed shocks argues for complete cessation of all activation wavefronts and subsequent reinitiation of VF after an isoelectric window (28). The optical equivalent of an isoelectric window supports the earlier electrode-based proposal that subthreshold shock terminates VF but fails to achieve successful defibrillation because the same shock reinitiates VF (3). For successful defibrillation, the shock strength must reach the upper limit of vulnerability (ULV) (3).

The ULV hypothesis states that after the complete cessation of all VF wavefronts, the failed near-threshold shock induces VF by the same mechanism that induces VF when the shock is applied during the vulnerable period of a regularly driven beat (31). A shock induces VF when applied during the vulnerable period by initiating a reentrant wavefront (5). Similarly, it is hypothesized that a failed near-threshold defibrillation shock also reinitiates VF by the formation of reentry (9). Because the pattern and the sequence of wavefront activation during VF are irregular and change constantly (21), the amount of fibrillating myocardium in its vulnerable period also changes. Owing to this inherent irregularity of activation wavefronts during VF, it is therefore likely that at the instant of a given shock, the amount of the myocardium in its vulnerable period will be different compared with another instant of shock. The aim of the present study was to test the hypothesis that near-threshold defibrillation shocks fail because the amount of the myocardium in its vulnerable period is higher than the amount of the vulnerable myocardium at the instant of successful defibrillation.

	METHODS AND MATERIALS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

The right ventricle (RV; 30 ± 2.5 g) of 10 farm pigs was isolated and perfused through the right coronary artery with oxygenated Tyrode solution at 37 ± 0.5°C in a tissue bath with the endocardial surface facing up (18). VF was induced by burst pacing [50-ms cycle length (CL) for 3-5 s] with 10 times diastolic current threshold. The 50% successful defibrillation (DFT₅₀) was determined by an up-down method (23) with a pair of 4-cm coil electrodes (model 1228, CPI) mounted in the tissue bath. One of the two electrodes was placed 1 cm away from the left edge of the RV and the second 1 cm away from the right edge of the tissue. In each isolated RV, at least five successes and five failures were obtained with shock strengths equal to the DFT₅₀. After each shock, 3-5 min of recovery time were allowed before another shock was tested. Shocks consisted of biphasic (6 and 6 ms) waves that delivered truncated exponential waveform shocks with fixed pulse duration and a variable tilt (model HVS200, Ventritex. Sunnyvale, CA). The biphasic waveform was generated by a positive-polarity truncated exponential waveform followed by a negative-polarity truncated exponential waveform. Leading edge voltage of the second phase was half the residual value of the first phase (13). The endocardial surface of five isolated RVs was mapped with 477 bipolar electrodes 1.6 mm apart spread over a 3.8 × 3.8-cm plaque that virtually covered the entire isolated, perfused RV endocardium. Less than 10% of the cut edges of the isolated RV remained outside the mapping plaque. The number of sites activated (sampling frequency 1 kHz) at each of the 477 recording electrodes was determined in 1-ms steps during the 100-ms VF interval that preceded the shock. Each activated site first became red, then pink, then yellow, then green, and then blue before fading to black. The persistence of each color was 5 ms (15, 18). This method of data analysis and display allowed us to determine the recovery time of each of the 477 sites at the time of the shock onset. Although we scanned the 100-ms interval preceding the shock with a 1-ms decrement, the sites showing recovery intervals of 5 ms were grouped together. To relate recovery times to the transmembrane action potential phases during VF, in five isolated, perfused swine RVs we continuously recorded with a glass microelectrode (16, 18) endocardial transmembrane action potentials for ~1 min at two different endocardial cells that were spaced 2-3 cm from each other. No activation mapping was performed in these five RV tissues. The total number of responses analyzed in each RV was ~1,500 beats (total of ~7,500 in all 5 RVs). The action potential duration (APD) for 90% repolarization (APD₉₀) of the regenerative action potentials (i.e., amplitude >40 mV) was measured. Lower-amplitude (i.e., <40 mV) responses were considered nonregenerative (i.e., graded responses; Refs. 10, 17) and were excluded. Ten to one hundred milliseconds of recovery time after the upstroke of each regenerative action potential during the VF was related to the phase of the action potential and to the range of transmembrane voltages attained.

Statistical analyses. In each RV tissue, differences between successful and failed defibrillation shocks relative to energy, voltage, and impedance were calculated with paired t-tests. An ANOVA test with Bonferroni correction was used to compare the number of activated sites at multiple preshock intervals of 5 ms each. Differences were considered significant if P < 0.05. Data are presented as means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Defibrillation shock parameters. A total of 26 trials of successful and 26 trials of failed defibrillation shocks were analyzed. There was no significant difference in the level of energy (1.52 ± 0.24 vs. 1.56 ± 0.26 J), voltage (192 ± 13 vs. 192 ± 15 V), and impedance (44.5 ± 6  vs. 45.5 ± 6.2 ) between successful and failed defibrillation shocks.

VF cycle length. There was no significant difference in the mean VF CL between successful and failed defibrillation shocks (79.7 ± 25.7 vs. 78 ± 24.7 ms; P = 0.5).

Amount of tissue recovery at shock onset. At shock onset, the number of sites and the recovery time at each of these sites were computed during successful and failed defibrillation shocks. The number of sites with 45-55 ms of recovery time was significantly (P < 0.04) lower for the successful defibrillation shocks than for the failed shocks. The mean number of sites with 45-55 ms of recovery time in five RVs was 23.3 ± 8.8 for the successful episodes versus 28.4 ± 6.7 for the failed episodes (P < 0.04). The numbers of sites presented represent the mean numbers of sites within this specific (i.e., 45-55 ms) recovery interval. The numbers of sites during successful and failed shocks correspond to 4.8% and 5.9% of the total recording surface area, respectively. Figure 1 illustrates an example of VF activation snapshots preceding a successful shock. It is apparent that at the instant of the shock there were nine sites in the entire mapped region with 45-55 ms of recovery time. Figure 2 shows an example of VF activation snapshots preceding a failed defibrillation shock in the same tissue as in Fig. 1 with the same shock strength as the successful shock (Fig. 1). In this case of failed defibrillation shock, there were 31 sites in the mapped region that had 45-55 ms of recovery time at the instant of the shock (Fig. 2). Outside this relatively narrow range of recovery time, no statistically significant difference could be detected in the amount of tissue (i.e., number of sites) with recovery intervals lasting <45 ms or >55 ms between successful and failed defibrillation shocks. The number of sites with 0- to 45-ms and 55- to 80-ms recovery times were 27.6 ± 10.1 (successful) versus 27.1 ± 11.3 (failed) and 21.8 ± 9.6 (successful) versus 22.3 ± 11.3 (failed) (P = 0.50), respectively.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Snapshots of activation patterns during ventricular fibrillation (VF) preceding a successful shock with 477 bipolar electrodes (top). Each activated site first becomes red, then yellow, then green, and then blue before fading to black. The persistence of each color is 5 ms. The number above each frame represents the time (in ms) that preceded the shock. Middle: continuous display (in steps of 1 ms) of the number of sites activated within 5 ms for the entire 100-ms period preceding the shock. There are 9 sites (red) that show 45 ms of recovery at the instant of the shock. Bottom: bipolar electrogram (i.e., pseudo-ECG). The electrical parameters of the shock were 1.5 J, 190 V, and 45 .

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Snapshots of activation during VF preceding an unsuccessful shock with 477 bipolar recording electrodes (top). This figure is from the tissue shown in Fig. 1 and is organized in the same manner as Fig. 1. There are 31 sites that show 45 ms of recovery at the instant of the shock. The electrical parameters of the shock were the same as in Fig. 1.

Relation of recovery to transmembrane action potentials. The mean VF CL (measured as the time between 2 consecutive action potential upstrokes) was 79 ± 22 ms, which was not significantly different (P = 0.6) from the VF CL measured with the bipolar electrodes of the mapping system. Nonregenerative responses (graded responses) with amplitudes <40 mV (10, 18) were observed with an incidence of 4.2 ± 0.5%. The graded responses were excluded from VF CL calculation (Fig. 3). During VF, the mean action potential amplitude was 67 ± 5 mV and the mean APD₉₀ was 63 ± 26 ms (range 34-88 ms; Fig. 3). The transmembrane voltage range that corresponded to 45- to 55-ms recovery was between 15 and 60 mV in 92% of regenerative action potentials.

View larger version (73K): [in this window] [in a new window]   Fig. 3.   Transmembrane action potential recordings in 5 isolated and perfused swine right ventricles (A-E). Each row shows 2 different recordings (cell 1 and cell 2) from 2 different cells 2-3 cm apart in each isolated ventricle during 2 different episodes of induced VF. Arrows indicate nonregenerative, graded responses, and the horizontal lines next to each panel represent zero transmembrane potential.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

New findings. The variable (probabilistic) nature of ventricular defibrillation shock outcome with shock strengths equivalent to the DFT₅₀ results from the variations of the amount of tissue in its vulnerable period at the instant of the shock. The results of the study show that the amount of myocardium in its vulnerable phase was significantly smaller during successful defibrillation shocks than during failed shocks with identical near-threshold shock strengths.

Vulnerable period, fibrillation, and defibrillation. The two critical prerequisites for the induction of VF by an electrical stimulus during regular pacing or sinus rhythm are stimulus timing (i.e., during the vulnerable period) and stimulus strength (25). During regular activation, both ventricles normally proceed quasisynchronously through the vulnerable period. The quasisynchronous nature of ventricular recovery during regular activation in normal hearts eliminates myocardial mass as a variable, making the shock strength and shock timing the sole determinants of shock outcome.

Amount of tissue in its vulnerable period versus shock outcome. The mechanism(s) by which myocardial tissue mass in its vulnerable period influences defibrillation shock outcome remains poorly understood. In vitro (10) and in vivo (5) studies of shock-induced VF during regular pacing often develop in the presence of a fixed amount of myocardium in its vulnerable period. Shocks applied during phase 3 prolong the duration of the action potential (19) and cause wave break at the site of the prolonged action potential duration in a repeatable and predictable manner (10) while allowing conduction at sites without prolongation of repolarization. However, this is not the case during VF. Because of the complexity of the activation pattern during VF (14, 21), the amount of myocardium in its vulnerable period changes constantly. Our results suggest that if the mass of myocardium in its vulnerable period at the instant of the shock is relatively high, the shock will fail to defibrillate. It appears that a certain amount of myocardial tissue (critical mass) must be present for the shock to induce and maintain the VF. Because a shock induces VF in normal ventricles by first initiating a reentrant wavefront of excitation, it appears that a critical mass of "vulnerable" myocardium must be present for reentry to be formed (stage I VF) (30). Reentry then evolves to stage II VF (30) by breakup (29) and fibrillatory conduction (33). Optical mapping studies have shown that failed defibrillation shocks are associated first with the formation of a postshock reentrant wavefront (phase singularity) that then breaks up into multiple wavelets, signaling the onset of VF (9). However, the role of the tissue mass in its vulnerable period in the formation of postshock reentry was not addressed in that study (9). In the instance of shock-induced VF, it appears that in addition to the shock strength and shock timing, the amount of myocardium in its vulnerable period also plays an important role in the shock outcome.

Limitations of study. Our two-dimensional (2-D) mapping studies were carried out in an essentially three-dimensional tissue. The 2-D maps do not allow the determination of intramural cells in their vulnerable period at the instant of the shock. The number of (endocardial) sites may therefore not represent the true volume of tissue in its vulnerable period. However, the consistent finding of a successful defibrillation outcome with a lower amount of endocardial sites in their vulnerable period suggests that our high-resolution and complete endocardial mapping of the isolated RV provides a reasonably accurate estimate of myocardial tissue mass in its vulnerable period. Although the amount of myocardium in its vulnerable period is higher in failed shocks than in successful shocks, the relation of the earliest site of activity after a failed defibrillation shock to the vulnerable myocardium was not determined in this study.

Clinical implications. Timing the shock when the mass of the fibrillating ventricles in its vulnerable period is smallest might increase the probability of successful defibrillation without the need to increase shock strength. Such shock timing remains a future challenge.