As shown in Fig. 13.2, there is increased sympathetic discharges both preceding and following the onset of VF. The sympathetic nerve activity releases norepinephrine, which increases the intracellular Ca²⁺. Abnormally large Ca²⁺ transient is known to be a mechanism of cardiac arrhythmia [22]. Zoll et al. [1] showed that electrical defibrillation is effective in terminating VF. However, in the same manuscript, the authors showed VF may also recur spontaneously soon after electrical defibrillation. The mechanisms by which electrical shocks terminate VF remain unclear, so are the mechanisms by which VF may spontaneously recur after VF-defibrillation episodes. Since the strong electrical shocks alter the transmembrane potentials by creating a voltage gradient field, defibrillation was thought to be achieved exclusively by the changes of membrane potential through electrical current. However, recent experimental data have demonstrated that defibrillation shocks also change Ca_i, which may be of great importance to the outcome of the defibrillation.

To understand the mechanisms of defibrillation, it is necessary to also understand the mechanisms of defibrillation failure. The mechanisms of failed defibrillation depend on the strength of defibrillation shocks [23, 24]. At weak strength shocks (typically >100 V lower than the shock strength with 50 % probability of successful defibrillation (DFT₅₀)), reentrant activation immediately after the shock contributes to the defibrillation failure by maintaining continuous VF. Optical mapping studies have reported that there was no post-shock interval at weak shock strengths and more phase singularities remained after delivery of the weaker shocks [23, 24]. To achieve defibrillation, the virtual electrode polarizations (VEPs) created by the shock-induced electric field need to have sufficient amplitude to halt the activation wavefronts of VF. However, this is not the only reason of defibrillation failure at weak shock strengths. Even if the VEPs extinguish all existing VF activation wavefronts, it is not sufficient to ensure successful defibrillation, because the VEPs can themselves generate new activation wavefronts that reinitiate VF. Since VF consists of unsynchronized rapid activations, some portions of the ventricular myocardium are in their vulnerable period no matter when the shock is delivered. When a critical value of the potential gradient encounters a critical degree of refractoriness of the tissue during the vulnerable period, reentry is formed at the critical point, which reinitiates VF [25, 26]. Stronger electric shocks prevent the formation of the critical point and the re-initiation of VF, because of the presence of an upper limit to the strength of shocks given during the vulnerable period that could induce VF (i.e. the upper limit of vulnerability) [27, 28]. Another type of the critical point has been reported in a study by Efimov et al. [29]. In this pattern, the critical point is created during the action potential plateau by shock-induced VEPs of depolarization and hyperpolarization. The hyperpolarization caused by virtual anode de-excites the myocardium during the plateau of the action potential, leading to recovery of the excitability at the hyperpolarized region. This allows the adjacent depolarized region caused by virtual cathode to activate the hyperpolarized region, giving rise to a new activation wavefront and initiation of reentry (i.e. the virtual electrode-induced phase singularity) [29]. Electric shocks at a sufficiently strong strength also prevent the virtual electrode-induced phase singularity, because an increase in the shock strength enhances the degree of depolarization and hyperpolarization at the virtual cathode and anode, respectively, which accelerates the conduction velocity of the activation wavefront across the boundary of each VEP [30]. The electric shock can evoke an action potential in the excitable segment of the ventricular myocardium during VF. However, the weak strength shocks can exert little effects on the myocardium during a refractory phase of the action potential, which leave the ventricular repolarization unsynchronized. Therefore, the critical point of either mechanism induced by electric shocks can induce reentry in the tissue with heterogeneous repolarization at weak shock strengths.

On the other hand, the mechanism of defibrillation failure differs at stronger shock strengths near defibrillation threshold (DFT) (shock strength within 50 V of DFT₅₀). It has been reported that the near DFT strength shock applied during the refractory period is capable of extending the APD and the effective refractory period [31]. The degree of the APD extension depends on the shock coupling interval, thereby causing the myocardium to repolarize at a constant time regardless of the pattern of the pre-shock electrical activities [32]. This synchronized repolarization elicited by the strong electric shock prevents the genesis of the shock-induced critical point and reentry. The optical mapping study has identified that the number of post-shock phase singularities decreases continuously from pre-shock values to zero as the shock strength increases [24]. Importantly, even though reentry was not observed immediately after the near DFT strength shock, defibrillation does not necessarily succeed. In other words, immediate post-shock reentry is not responsible for defibrillation failure at near DFT shock strengths. In that case, activation is not recorded in both electrical and optical mapping studies for a period of 50–90 ms following the near DFT shocks (i.e. the isoelectric window), after which earliest activation always propagates centrifugally away from the early site in a focal pattern [23, 33–35]. If the shock fails to defibrillate, these foci arise rapidly and repetitively for several cycles, resulting that the activation wavefronts degenerates back into VF. If the shock successfully defibrillates, no foci are observed or only a few focal activations last until they stop spontaneously. One question is whether the shock extinguishes all VF activations during isoelectric window. Many electrical and optical studies mapped epicardial surfaces; hence one may assume that activation wavefronts are still present intramurally or in the Purkinje fibers during isoelectric window, and the focal pattern of the epicardial activation for the first post-shock beat represents epicardial breakthrough from the deeper layer of the ventricle. A simulation study [36] demonstrated that, during the isoelectric window, an activation wavefronts induced by VEPs within the deep ventricular wall propagated fully intramurally through an excitable tunnel, until it emerged onto the epicardium, forming the focal pattern of the first post-shock epicardial activation. However, three-dimensional whole-heart mapping studies with plunge electrodes showed no evidence of wavefront propagation or reentry during the isoelectric window, and a new activation wavefront emerged focally following the isoelectric window [33, 35]. It is possible that plunge needle electrodes could not detect slowly propagated graded responses or were too few recordings to demonstrate activations in the intramural regions or the Purkinje system. However, we found that the isoelectric window followed by a focal discharge was still observed using an optical mapping technique in rabbit hearts with endocardial cryoablation in which only a thin layer (≈0.5-mm) of surviving epicardial tissues [37]. Therefore, the experimental evidence available supports the idea that the near DFT shocks eliminate all VF activations. A new wavefront of the focal mechanism rather than reentry is re-induced by the defibrillation shock in failed defibrillation.

We simultaneously mapped epicardial membrane voltage (V_m) and Ca_i during attempted defibrillation at a near DFT shock strength in Langendorff-perfused rabbit ventricles [37]. We found that the heterogeneous post-shock distribution of Ca_i plays a key role in defibrillation failure. A focal activation following the isoelectric window in unsuccessful defibrillation always originates from the regions of low Ca_i surrounded by elevated Ca_i (i.e. Ca_i sinkholes) (Fig. 13.3a). No Ca_i sinkholes are present after type A successful defibrillation. We also determined the relationship between the Ca_i sinkholes and vulnerability to VF, since re-induction of VF by a defibrillation shock itself is responsible for defibrillation failure. The Ca_i sinkholes were also present after a shock on T-wave that induced VF (Fig. 13.3b) [37, 38]. What is the mechanism by which the first post-shock activation consistently arises from the Ca_i sinkhole? The first post-shock activation following the isoelectric window is focal, suggesting triggered activity as a potential mechanism. In some episodes of unsuccessful defibrillation, an earlier rise in Ca_i than V_m are observed at the early site for the post-shock focal activation. The ratio map shows that a rise in Ca_i precedes a rise in V_m at the Ca_i sinkhole site (i.e. Ca_i prefluorescence, arrow in Fig. 13.4). It is possible that the Ca_i prefluorescence might be attributable to spontaneous (voltage-independent) Ca²⁺ release from the sarcoplasmic reticulum (SR) that induces triggered activity by the reverse excitation-contraction coupling [39]. The low Ca_i in the sinkhole reflects more complete SR Ca²⁺ uptake. With the SR more Ca²⁺ loaded and more fully recovered, it may be more susceptible to spontaneous Ca²⁺ release from the SR, compared with regions with persistently high Ca_i. Consistent with this hypothesis, suppression of SR function by ryanodine and thapsigargin significantly reduced the upper limit of vulnerability and DFT [37]. Alternatively, EAD might be another explanation for the triggered activity. At the Ca_i sinkhole site, the Ca_i continues to decline, which promotes the recovery of I_Ca,L from inactivation. Reactivation of I_Ca,L increases both Ca_i and V_m, leading to EAD and triggered activity. Another possible mechanism is the electrotonic interaction with surrounding cells. In the Ca_i sinkholes, APD is shorter than in the surrounding sites, possibly because the reduced Ca_i level in the sinkholes may shorten the APD by the bidirectional coupling between Ca_i and V_m [40, 41]. Electrotonic currents from the surrounding depolarized regions might assist impulse formation in the sinkhole region in the similar way to phase 2 reentry or reflection [42].

The heterogeneity of Ca_i as well as that of V_m immediately after shock seems important in the mechanisms of defibrillation and vulnerability, because it facilitates the formation of Ca_i sinkholes. It is known that biphasic waveform shock is more effective than monophasic waveform shock in achieving successful defibrillation. We demonstrated that the greater efficacy of biphasic waveform is directly related to less heterogeneous post-shock Ca_i distribution, thus reducing the probability of Ca_i sinkhole formation and VF re-initiation [43]. The electric field created by defibrillation shock alters the V_m but it also significantly changes the Ca_i [44, 45]. Transient hyperpolarization at virtual anodes is expected to increase the driving force for Ca²⁺ entry through I_Ca,L, potentiating SR Ca²⁺ release, while transient depolarization at virtual cathodes is not expected to exert this effect [46]. It follows that a shock-induced APD prolongation is more accentuated at the virtual anode sites than at the virtual cathode sites, since the increased Ca_i prolonged APD by potentiating inward Na⁺-Ca²⁺ exchange current (I _NCX), consistent with positive Ca_i-APD coupling [43, 47]. During VF, the effects of VEPs on Ca_i are variable, because effects of shocks on Ca_i depend on the timing of shock application relative to the local repolarization state [43, 45]. Thus, the sites of the Ca_i sinkhole are not predictable by distribution of the VEPs, but the heterogeneous effects of VEPs on Ca_i would promote the Ca_i sinkhole formation. This may account for the more reduction in the Ca_i sinkholes by the biphasic waveform shocks which have a neutralizing effect on the VEPs.

Defibrillation failure can be viewed as an immediate recurrence of VF after the shock halts all VF activations. This type of VF recurrence generally occurs 50–100 ms after the shocks [23, 24, 37]. However, as documented by Zoll et al. in their first paper on electrical defibrillation [1], VF can also recur seconds after an electrical countershock that was initially successful. More recent studies show that spontaneous recurrences of VF can occur 1 s to 2 min after defibrillation for long-duration VF with resultant ischemia [48]. Also, defibrillation during global ischemia is often followed by immediate VF recurrences (<10 s after shock) [49]. Importantly, when the VF recurrences occur during ECG saturation caused by the strong electric shock, it masquerades as defibrillation failure. This implies that a shock-resistant VF, which is often observed in the out-of-hospital sudden cardiac arrest [50], may include this type of the pseudo-defibrillation failure. Furthermore, it has been reported that recurrent VFs in patients with a normal structural heart frequently arise from the distal Purkinje system [51]. In canine electrical mapping studies, Purkinje fibers are highly active and the earliest post-shock activation always originates from the Purkinje fibers after long-duration VF [48]. We recently found that triggered activity due to delayed afterdepolarization (DAD) is the mechanism of post-shock ventricular arrhythmias originating from the Purkinje fibers which have a higher Ca_i-V_m coupling gain due to a smaller I_K1 [52].

Recurrent spontaneous VFs after successful defibrillation are observed frequently in patients with heart failure [53]. We recently developed a rabbit model with pacing-induced heart failure which shows recurrent episodes of spontaneous VF [54, 55]. Simultaneous mapping of V_m and Ca_i in the failing hearts identified that marked APD shortening and disproportionally persistent Ca_i transient after successful defibrillation induce afterdepolarizations during late phase 3 of the action potential possibly by promoting the forward mode of I _NCX (i.e. CCTF or late phase 3 EAD) [9–11], which causes triggered activity and re-initiation of VF. The importance of intracellular Ca²⁺ handling in electrical storm is also supported by the study by Tsuji et al. [56]. In that study, the authors found abnormal protein phosphorylation may affect the Ca²⁺ handling and contribute to the electrical storm in rabbits with complete heart block. In addition to abnormal Ca_i accumulation, shortened APD is also a major factor that promotes electrical storm. We discovered that in heart failure, there is a significant upregulation of apamin-sensitive potassium current (I _KAS) conducted through the SK channels. Because of the Ca_i accumulation during VF, this channel is activated to conduct I _KAS and shorten the APD. Inhibiting I _KAS by apamin can suppress recurrent VF in our model (Fig. 13.5). Patch clamp studies confirmed the upregulation of I _KAS in failing rabbit ventricles (Fig. 13.6). Our recent studies documented that SK channel is upregulated in failing human ventricles [57]. Further studies of I _KAS may contribute to the better understanding of the mechanisms by which VF begets VF.