Pacing After Shocks Stronger Than the Upper Limit of Vulnerability
Impact on Fibrillation Induction
Background—After upper-limit-of-vulnerability (ULV) shocks of the same strength and coupling interval (CI) during the T wave, (1) the epicardial activation pattern (EAP) for the first postshock cycle is indistinguishable between shocks that do (VF) and do not (NoVF) induce ventricular fibrillation (VF) and (2) ≥3 cycles in rapid succession always occur during VF but not during NoVF episodes. To study the role of these rapid cycles, rapid pacing was performed after a shock stronger than the ULV that by itself did not induce rapid cycles and VF.
Methods and Results—A 504-electrode sock was sutured to the heart in 6 pigs to map EAPs. The S2 shock strength and S1-S2 CI at the ULV were determined by T-wave scanning with an up/down protocol. Ten shocks 50 to 100 V above the ULV (aULV) were delivered at the same S1-S2 CI to confirm that VF was not induced. Then, the postshock interval after aULV shocks was scanned with an S3 pacing stimulus from the LV apex until the shortest S2-S3 CI that captured was reached. This was repeated for S4, S5, etc, until VF was induced. To induce VF, 3 pacing stimuli (S3-S5) with progressively shorter CIs were required; S3 or S3, S4 never induced VF. After cycle S5, which induced VF, 2 EAP types occurred: focal (74%) and reentrant (26%).
Conclusions—At least 3 cycles with short CIs are necessary for VF induction after aULV shocks. Cycles S3-S4 may create the substrate for cycle S5 to initiate VF.
Ventricular fibrillation (VF) induction is probabilistic and dependent on shock strength and timing.1 2 For shock strengths near the upper limit of vulnerability (ULV), the global dispersion of refractoriness immediately after the shock and the characteristics of the first postshock cycle may not be sole determinants of VF induction.3 Instead, the number and rapidity of activation cycles soon after the shock appear to determine whether secondary reentry occurs, leading to VF.3 These findings suggest that 3 ectopic, repetitive, overlapping cycles after near–ULV strength shocks are required to initiate VF.3
We tested the following hypotheses dealing with near-ULV shocks during the vulnerable period: (1) The key determinant of VF induction is the number and rapidity of postshock cycles, not the characteristics of the first postshock cycle. (2) At least 3 postshock cycles with overlapping activation fronts are necessary to initiate VF. The hypotheses were tested by delivery of shocks 50 to 100 V above the ULV, which never induced VF alone, followed by 1 to 5 pacing stimuli to see when VF was induced.
The shock above the ULV (aULV) was 550±58 V. The mean coupling interval (CI) for S1-aULV shocks was 224±13 ms. Diastolic pacing thresholds (DPT) at the endocardial right ventricular (RV) apex and the anteroapical left ventricle (LV) were 0.2±0.0 and 0.2±0.1 mA, respectively. Heart weight was 163±8 g. Of the 60 VF episodes induced by postshock pacing, 58 were analyzed. The other 2 episodes were not included because the last S1 stimulus did not capture.
The aULV shock alone never induced VF. Intercycle intervals (ICIs) of the first 4 ectopic cycles (Figure 1A⇓) were very long (517±126, 319±28, 518±162, and 474 ms, respectively). There was only 1 fourth and no fifth ectopic cycle in any NoVF episode. Wavefront conduction times (WCTs) (Figure 1B⇓) were relatively short and constant (56±16, 61±2, 66±5, and 63 ms for cycles 1 through 4, respectively). The overlapping index was always <1 (Figure 1C⇓), indicating no overlap in any NoVF episode.
During post-aULV pacing, VF was never induced by single (S3) or double (S3-S4) stimuli, even at the shortest coupling interval (CI) that captured (107±32 and 130±11 ms for S2-S3 and S3-S4, respectively). When the third pacing stimulus (S5) was added, it always induced VF for the shortest CI that captured (118±12 ms) and for up to 20 ms longer than the shortest CI that captured. For this 20-ms CI range, new activations after the S5 cycle always appeared spontaneously even though no further stimulus was delivered. For an S4-S5 CI that was >20 ms longer than the shortest CI that captured, VF was never induced and no or only 1 ectopic cycle appeared before sinus rhythm resumed.
For all VF episodes, ICIs for the first 5 postshock activations were 138±31, 137±13, 111±19, 141±25, and 132±25 ms (Figure 1A⇑). The first 3 cycles were paced (S3-S5), and the last 2 were spontaneous. ICIs of the S3-S5 cycles for VF episodes were shorter than for the first 3 spontaneous cycles for NoVF episodes (P<0.01). Because only 1 fourth and no fifth ectopic cycles appeared during NoVF episodes, these cycles were not compared. WCTs for the first 5 cycles for VF episodes were 89±10, 147±13, 191±31, 175±35, and 161±40 ms, respectively (Figure 1B⇑). The WCT of the S3-S5 cycles for VF episodes was longer than for NoVF episodes (P<0.01). Although there was no overlap between cycles 1 and 2 for VF episodes (Figure 1C⇑), overlap cycles occurred starting at the transition from cycle 2 to 3 (index >1). Because 3 postshock pacing stimuli always induced VF, no S6 or S7 stimuli were given.
In VF induced by sub-ULV shocks (Figure 2⇓), the first 5 spontaneous postshock cycles showed activation patterns similar to those of the post-aULV pacing-induced cycles. ICIs of cycles 2 (150±15 ms), 3 (124±28 ms), 4 (130±13 ms), and 5 (122±10 ms) were not different from those of the post-aULV pacing VF inductions. However, the ICI of cycle 1 (66±15 ms) was shorter for sub-ULV than for post-aULV pacing VF inductions (P<0.001, Figure 1A⇑). WCTs of the first 5 cycles (104±16, 138±22, 180±18, 182±22, and 171±9 ms, respectively) after sub-ULV VF inductions were not different from those in the post-aULV pacing VF inductions. Overlapping cycles were always found during the transition from cycles 2 to 5 (Figure 1C⇑). However, there was no overlap during the transition from cycle 1 to 2 in any episode. Maps of 1 VF episode induced by a sub-ULV shock in 1 animal are shown in Figure 2⇓, demonstrating that the site of earliest activation (SEA) of cycles 1 to 5 was in the anterior LV apex, where the S3-S7 pacing electrode was placed.
Activation With Post-aULV Pacing
Activation appearing spontaneously after the S5 paced cycle in all VF episodes exhibited 2 patterns: focal and reentrant. Of the 58 VF episodes, a focal pattern was found in 43 (74%) and a reentrant pattern in 15 (26%).
Reentry was always in the posteroapical RV toward the RV apex (Figure 3⇓). The 3 paced activations propagated across the ventricles in a focal pattern with progressive slowing from cycles 1 (S3) to 3 (S5) in the apical and posterior RV overlying the S2 electrode. Activation propagated anteriorly and posteriorly around the region of slow conduction, toward the RV base, collided at the posterobasal RV, and terminated for cycles 1 (frames 210 to 250) and 2 (frames 410 to 470). Cycle 3 (S5) appeared on the epicardium at the LV apex (frame 410) as cycle 2 (S4) activated the RV apex, causing overlapping cycles. Cycle 3 activation blocked (frames 530 to 570) on reaching the posterior RV apex, allowing the anterior and posterior fronts to create figure-8 reentry by reactivation of the posteroapical RV (frames 590 to 650, arrows), initiating the first nonpaced cycle (cycle 4). This pattern repeated in cycle 5 (frames 750 to 810), which was interrupted when a new ectopic focus appeared at the anterobasal RV (frame 830, white circle). This focal activation front collided with the reentrant front of cycle 5, interrupting reentry. Next, multiple foci appeared on the epicardium, degenerating into VF (Figure 3B⇓).
An example of an epicardial focal pattern is shown in Figure 4⇓. Focal activations of cycles 1 (S3, frames 170 to 250) and 2 (S4, frames 290 to 450) were similar to those in Figure 3⇑ but propagated faster because the S3 stimulus was delivered 20 ms later. Cycle 3 activation (S5) propagated toward the RV apex without blocking (frames 410 to 590). Before cycle 3 activated the entire epicardium, cycle 4 activation appeared spontaneously on the apical epicardium close to the pacing site (frame 530) and propagated in a focal pattern (frames 560 to 690). Cycle 5 arose from multiple ectopic foci (frame 670) before cycle 4 terminated. Overlap was found in all cycle transitions except from cycles 1 to 2. No reentrant activation was seen during the first 5 cycles.
The importance of the S5 stimulus and its CI is emphasized in Figures 5⇓ and 6⇓. When the S5 stimulus was not given (Figure 5⇓) or was given at an S4-S5 CI 30 ms longer than the shortest CI that captured (Figure 6⇓), no new spontaneous activation appeared after the last paced cycle, and a long pause occurred before sinus rhythm resumed. In a few cases, a single ectopic cycle occurred before sinus rhythm resumed.
Pacing Without aULV Shocks
VF was never induced when P1, P1-P2, P1-P2-P3, or P1-P2-P3-P4 stimuli were delivered at any CI after 10 S1 stimuli. When 5 stimuli (P1-P5) were delivered, VF was not induced at a strength of 5 to 10 times DPT. After the P5 cycle, there was either a pause or, occasionally, only 1 or 2 ectopic cycles before sinus rhythm resumed. However, in the 2 animals in which P1-P5 strength was increased to 20 times DPT, 4 episodes of VF (2 in each animal) were induced when all 5 stimuli were at the shortest CIs that captured (208±14, 138±7, 121±15, 116±17, and 126±31 ms, respectively). In all 4 VF episodes, 1 or 2 ectopic foci always appeared spontaneously on the epicardium close to the pacing site after the last paced (P5) cycle before degeneration into VF (Figure 7⇓). This pattern was similar to that in the VF induction episodes by 3 pacing stimuli after an S2 shock. However, without the S2 shock, spontaneous activations never appeared on the epicardium after the third or fourth pacing stimulus. Furthermore, without an S2 shock, VF induction was not reproducible, because the ectopic focus occurred only rarely after the P5 paced cycle.
Number and Rapidity of Postshock Activations Determine VF Induction
Our major finding is that ≥3 rapid activations are required to initiate VF after shocks 50 to 100 V stronger than the ULV. The first and second postshock cycles are not the sole determinants for VF induction, as shown by the similarity of these cycles for VF and NoVF episodes. The third cycle, however, led to VF initiation when it appeared so rapidly that it overlapped in time with the second cycle. When the third pacing stimulus was absent or was delivered ≥30 ms after the shortest CI that captured, VF was never induced. Thus, the short ICI and long WCT of the first 2 postshock cycles appear to create the substrate for VF initiation by the third cycle.
Without the S2 shock, VF was not induced by 3 premature pacing stimuli at any CI. Furthermore, in the absence of the shock, VF was only occasionally induced by 5 premature stimuli that required a greater strength than that used to induce VF by pacing after S2 shocks. This finding is consistent with a report by Hamer et al4 that showed that VF induction depended on the strength and timing as well as the number of premature stimuli.
Our study also demonstrates that VF can always be induced by 3 postshock stimuli with short CIs. Thus, in addition to the premature stimuli, the shock also plays a significant role in VF induction, perhaps because the strong shock field prolonged refractoriness in the posteroapical RV near the S2 electrode, which in turn caused postshock activation to slow and block in that region, creating the substrate for reentry.5 Without the shock, 3 cycles, no matter how rapid, are not sufficient to cause block.
Mechanism for VF Induction
Both focal6 7 8 and reentrant9 10 patterns have been reported to initiate VF, as observed in this study after 3 paced cycles. Because the spontaneous focal activations appeared close to the pacing site, they could have arisen from ectopic foci or Purkinje fibers excited by strong electrical stimulation.11 Alternatively, intramural reentry is possible, because 3D mapping was not performed.
Although epicardial reentry was seen in some cases in this study, it was not observed during the first 5 postshock cycles in our previous study, in which VF was induced by a ULV50 shock.3 This may be a result of the differences in the S2 shock strength used in the 2 studies. A shock 50 to 100 V above the ULV, but not a ULV50 shock, could be strong enough to cause slow propagation and block, allowing reentry to occur. Also, the local S3-S5 stimuli after aULV shocks may have had different effects than the spontaneously arising activation fronts after ULV50 shocks.
Whatever the cause of these spontaneous activations, this study emphasizes the importance of not just the first but of several cycles after shocks of near-ULV strength on VF induction. Future studies need to focus on the small arrhythmogenic area at the SEA in the low-voltage-gradient area near the LV apex,12 because this study suggests that the primary determinant of the outcome of shocks near the ULV could be limited to this small region, at least in normal hearts.
The requirement of 3 pacing stimuli for VF induction after the aULV shocks may not apply to all conditions. The number of stimuli required for VF induction could differ depending on shock strength, drugs, heart size, cardiac disease, and species.
Because only epicardial mapping was performed, intramural reentry cannot be ruled out. The mapping array had 4-mm interelectrode spacing, which may be too coarse to detect microreentry. Because optical mapping was not performed, the changes in action potentials and repolarization patterns were not investigated.
After shocks of near-ULV strength, the number and rapidity of postshock cycles induced by the shock are primary determinants of VF induction. The progressive decrease in ICI and increase in WCT of the first several cycles, as well as the effect of the shock, are essential for VF induction. These findings support the hypothesis that the immediate postshock activation pattern is not the sole determinant for the shock outcome for shocks near ULV strength in normal hearts and that a minimum number of cycles (3 in this study) are necessary to initiate VF after ULV shocks.3
Animal Preparation and Electrode Placement
Six healthy pigs (30 to 35 kg) of either sex were anesthetized and monitored as described previously.3 The chest was opened through a median sternotomy, and the heart was suspended in a pericardial cradle. At the end of the study, the animal was euthanized by fibrillation.
Shocking, pacing, and recording electrodes were described previously.3 Shocking electrodes (S2) were at the right ventricular (RV) apex and at the superior vena cava right atrial junction (Figure 8A⇓). S1 pacing was from the RV catheter tip. An anteroapical LV epicardial electrode (Figure 8A⇓) was used to pace after shocks (S3-S7). A 504-electrode sock recorded epicardial electrograms (Figure 8B⇓).
First, the mapping array was oriented.3 The intrinsic R-R interval and the diastolic pacing threshold (DPT) were measured. Pacing stimuli (S1, S3-S7) were unipolar, 5-ms, monophasic pulses. S2 shocks were biphasic (6-ms first phase, 4-ms second phase), truncated exponential waveforms delivered from a defibrillator (HVS-02, Ventritex Corp). Delivered voltage, current, and energy as well as the beginning, peak, and end of the T wave were determined as described previously.3
ULV shock strength was determined with a modified up/down protocol and T-wave scanning as described previously.3 Next, 10 shocks 50 to 100 V above the ULV (aULV) were delivered at the S1-S2 coupling interval (CI) that last induced VF during ULV determination (aCI) to confirm that (1) VF was not induced, and (2) the postshock interval, ie, the interval between aULV delivery and earliest activation after the shock, was >100 ms. If this had been <100 ms, then spontaneous activation would have occurred before S3 stimulation.
After S2 shocks of aULV strength and aCI, all stimuli (S3-S7) were 5 to 10 times the DPT delivered from the anteroapical LV. This site was chosen because previous studies1 3 13 showed that the site of earliest activation (SEA) after near–ULV strength shocks using the RV–superior vena cava electrode configuration arose in this low-potential-gradient region.12 Post-S2 pacing (Figure 8C⇑) started with an S2-S3 CI of 100 ms. If S3 captured, the S2-S3 CI was decreased in 10-ms steps. If S3 did not capture, the S2-S3 CI was increased in 10-ms steps. Scanning was continued until the shortest S2-S3 CI that captured was found.
S4 scanning was then performed in the same fashion as the S3 protocol until the shortest S3-S4 CI that captured was obtained, whereas the S2-S3 interval remained at the shortest CI that captured. Next, S5, S6, and S7 stimuli were added one at a time until VF was induced. The initial S4-S7 CIs were the CIs for that cycle observed during VF induction with sub-ULV shocks in our previous study (S4, 140 ms; S5, 130 ms; S6, 130 ms; and S7, 130 ms).3 The shortest S4, S5, S6, and S7 CIs that captured were determined by the same scanning protocol as for S3 stimulation. If any postshock pacing stimulus later ceased capturing at its shortest CI, its CI was increased in 10-ms steps until a capture resumed.
After each VF induction, the aULV shock was delivered after an S1 train without postshock pacing to confirm that the aULV shock itself did not induce VF. Ten VF episodes were induced by S3-S7.
Pacing Without S2 Shocks
To examine the effect of premature pacing without a shock, 5 pacing stimuli (P1-P5, 5 to 10 times DPT) were delivered after 10 S1 without an S2 shock. In 2 animals, pacing stimuli at 20 times DPT were also tested. The pacing and scanning protocol for P1-P5 was performed as described for the post–S2 shock pacing protocol, starting at a P1 CI after the last S1 stimulus of 250 ms. The initial CI was 200 ms for P2 and P3 and 150 ms for P4 and P5. When VF was induced, a rescue shock (20 to 30 J) was delivered from the S2 electrodes.
Activations were analyzed by visualization of an animation of the first derivative of the electrograms as described previously.3 The first 5 activation cycles after VF and NoVF induction episodes by S2 (aULV) shocks alone, by S2 (aULV) shocks followed by pacing, and by P1-P5 pacing alone were analyzed. If there was no VF induction, only ectopic (nonsinus) cycles were analyzed. The first 5 cycles of the last VF episode induced by a sub-ULV shock (10 to 20 V below the ULV) during ULV determination were also analyzed.
The SEA was defined as the location of the first electrode that registered activation of each cycle. The intercycle interval (ICI) was the interval between the SEA activation times for consecutive cycles. ICI for the first cycle was the postshock interval. Wavefront conduction time (WCT) was the interval between activation times at the SEA and at the site of latest activation for each cycle. Overlapping cycles in which earliest activation of the cycle (n+1) was observed before latest activation of the cycle (n) were detected by dividing the WCT of cycle (n) by the ICI of cycle (n+1) (the overlapping index). No overlap was present when the index was ≤1.
VF and NoVF induction episodes were compared by Student’s t test for paired and unpaired data. Values are shown as mean±SD. Differences were considered significant for values of P≤0.05.
This study was supported in part by National Institutes of Health research grants HL-28429 and HL-42760. Dr Fotuhi is a recipient of a Habilitationsstipendium der Deutschen Forschungsgemeinschaft.
The Methods section of this article can be found at http://www.circulationaha.org
- Received February 16, 1999.
- Revision received September 14, 1999.
- Accepted October 7, 1999.
- Copyright © 2000 by American Heart Association
Shibata N, Chen P-S, Dixon EG, Wolf PD, Danieley ND, Smith WM, Ideker RE. Influence of shock strength and timing on induction of ventricular arrhythmias in dogs. Am J Physiol. 1988;255:H891–H901.
Malkin RA, Idriss S, Walker RG, Ideker RE. Effect of rapid pacing and T-wave scanning on the relation between the defibrillation and upper-limit-of-vulnerability dose-response curves. Circulation. 1995;92:1291–1299.
Chattipakorn N, Rogers JM, Ideker RE. Influence of postshock epicardial activation patterns on the initiation of ventricular fibrillation by shocks near the upper limit of vulnerability. Circulation. 2000;101:1329-1336.
Frazier DW, Wolf PD, Wharton JM, Tang ASL, Smith WM, Ideker RE. Stimulus-induced critical point: mechanism for electrical initiation of reentry in normal canine myocardium. J Clin Invest. 1989;83:1039–1052.
Sano T, Scher AM. Multiple recording during electrically induced atrial fibrillation. Circ Res. 1964;14:117–125.
Sano T, Sawanobori T. Mechanism initiating ventricular fibrillation demonstrated in cultured ventricular muscle tissue. Circ Res. 1970;26:201–210.
Scherf D, Schott A. Extrasystoles and Allied Arrhythmias. London, UK: William Heinemann Medical Books Limited; 1973.
Han J, Moe GK. Nonuniform recovery of excitability in ventricular muscle. Circ Res. 1964;14:44–60.
Chen P-S, 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.
Tang ASL, Wolf PD, Afework Y, Smith WM, Ideker RE. Three-dimensional potential gradient fields generated by intracardiac catheter and cutaneous patch electrodes. Circulation. 1992;85:1857–1864.