Effect of acute global ischemia on the upper limit of vulnerability: a simulation study

Blanca Rodríguez , Brock M. Tice , James C. Eason , Felipe Aguel , José M. Ferrero Jr., Natalia Trayanova

Abstract

The goal of this modeling research is to provide mechanistic insight into the effect of altered membrane kinetics associated with 5–12 min of acute global ischemia on the upper limit of cardiac vulnerability (ULV) to electric shocks. We simulate electrical activity in a finite-element bidomain model of a 4-mm-thick slice through the canine ventricles that incorporates realistic geometry and fiber architecture. Global acute ischemia is represented by changes in membrane dynamics due to hyperkalemia, acidosis, and hypoxia. Two stages of acute ischemia are simulated corresponding to 5–7 min (stage 1) and 10–12 min (stage 2) after the onset of ischemia. Monophasic shocks are delivered in normoxia and ischemia over a range of coupling intervals, and their outcomes are examined to determine the highest shock strength that resulted in induction of reentrant arrhythmia. Our results demonstrate that acute ischemia stage 1 results in ULV reduction to 0.8A from its normoxic value of 1.4A. In contrast, no arrhythmia is induced regardless of shock strength in acute ischemia stage 2. An investigation of mechanisms underlying this behavior revealed that decreased postshock refractoriness resulting mainly from 1) ischemic electrophysiological substrate and 2) decrease in the extent of areas positively-polarized by the shock is responsible for the change in ULV during stage 1. In contrast, conduction failure is the main cause for the lack of vulnerability in acute ischemia stage 2. The insight provided by this study furthers our understanding of mechanisms by which acute ischemia-induced changes at the ionic level modulate cardiac vulnerability to electric shocks.

  • ionic channels
  • computer simulations
  • arrhythmias

electrical defibrillation is recognized as the most effective therapy against the malignant arrhythmias that lead to sudden cardiac death. However, although the majority of the patients that undergo defibrillation typically suffer from some form of coronary disease, little is known about the effect of acute ischemia on defibrillation efficacy. Experimental studies provide conflicting evidence: some report an increase in defibrillation threshold (DFT) (3, 29, 38, 43), whereas others find no change (5, 20, 22, 32) or even a decrease (2) in DFT during acute ischemia.

Numerous studies (5, 7, 18) have demonstrated that the DFT is strongly linked to the upper limit of vulnerability (ULV) of cardiac tissue to electric shocks. Much research has focused on investigating the mechanisms of cardiac vulnerability in an attempt to better understand how failed defibrillation shocks reinitiate cardiac arrhythmias. During the past decade, experiments using optical mapping techniques (8, 12, 13, 46) and computer simulation studies (25, 30, 37, 41) have significantly improved the understanding of the mechanisms of shock-induced arrhythmogenesis, culminating in the virtual electrode polarization (VEP) theory. As demonstrated in these studies, an electric shock induces both positive and negative changes in transmembrane potential (Vm) in cardiac tissue, which set the stage for the ensuing postshock activity. However, the majority of these theoretical and experimental studies investigate the mechanisms of shock-induced arrhythmogenesis in the normal myocardium and seldom during acute myocardial ischemia. Knisley and Holley (23) were the first to present evidence of VEP on the anterior left ventricular epicardium during 15-min occlusion of the left anterior descending artery. Recently, the optical mapping experiments of Cheng et al. (9) demonstrated that VEP was pivotal in shock-induced arrhythmogenesis after 30 min of 75% reduction in coronary flow.

Understanding postshock activity in the ischemic heart is, however, hampered by limitations in the current experimental modalities. These remain inadequate in providing information regarding the shock-induced VEP and postshock activity confined in the depth of the myocardium. Our group has recently developed a sophisticated three-dimensional (3-D) stimulation/defibrillation model based on canine ventricular geometry and fiber architecture (17, 42). Simulations of arrhythmia induction with this model have the unique ability to provide mechanistic insight into cardiac vulnerability to electric shocks under normal and pathological conditions, including acute myocardial ischemia.

The goal of this research is to employ our realistic model of stimulation/defibrillation to investigate the mechanisms underlying the changes in the ULV occurring from 5 to 12 min of acute global ischemia, before the occurrence of cellular uncoupling. This study focuses on global ischemia as a first step in understanding the mechanisms that underlie clinical ischemia related to coronary heart disease. We hypothesize that modifications caused by acute ischemia in both membrane electrophysiological properties and the effect of the shock are responsible for the change in the ULV. To test this hypothesis, we incorporate, homogeneously throughout the tissue, the effects of hyperkalemia, acidosis, and hypoxia on the membrane dynamics of cardiac cells. We consider two degrees of global acute ischemia: stage 1, corresponding to 5–7 min after the onset of ischemia and stage 2, representing 10–12 min of coronary occlusion (6, 14, 21, 33, 44, 45). For each degree of ischemia, we determine the ULV, and we examine the mechanisms that result in its change compared with normoxia.

METHODS

Computational model.

We used the canine ventricular defibrillation model described in detail previously (17, 42). Briefly, ventricular anatomy and fiber orientation were based on the Auckland data (24, 28). The ventricles were placed in a box representing the perfusing bath. Because modeling postshock activity in the entire canine ventricles is computationally intractable, two planes were passed through the ventricles and the bath to isolate a 4-mm-thick slice from the location shown in Fig. 1. Thus the slice included both cardiac tissue and surrounding conductive media (portions of the original perfusing bath and the blood cavities). The top and bottom surfaces of the slice were insulated as if pressed between glass plates. The electrical activity in the myocardial slice was simulated using the bidomain equations (15). The potentials in the blood and perfusing bath were governed by Laplace's equation. The Luo-Rudy phase II action potential model (26, 48) was used to represent the nonlinear kinetics of the membrane currents. To stabilize the Luo-Rudy phase II model for high-strength defibrillation shocks, the model's sodium current (INa), potassium current, and intracellular calcium concentration were all altered as in previous studies (39, 42). Simulations were performed using a semi-implicit finite element method with a variable time step as previously described (41).

Fig. 1.

Model geometry. The slice is 4 mm thick and located at the widest part of the canine heart. Fiber orientation and pacing and defibrillation electrodes are also shown. The surface area occupied by the external anode measures 10 × 2 mm, whereas the right ventricular (RV) cathode area is 2.82 × 2 mm.

Simulation of acute global ischemia.

Acute global ischemia was represented by the effects of its three major components, hyperkalemia, acidosis, and hypoxia, on the electrophysiological activity of cardiomyocytes. Two degrees of early acute ischemia were considered as stages 1 and 2. To represent these two degrees of ischemia: 1) the extracellular K+ concentration ([K+]o) was increased from its normal value of 5.4 mmol/l to 8.7 and 12.0 mmol/l in stages 1 and 2 of acute ischemia, respectively (6, 44, 45); 2) the maximum conductances of the INa and L-type Ca2+ current (ICa,L) were decreased to 87.5 and 75% of their normal values in ischemia stages 1 and 2, respectively, to simulate their inhibition due to intracellular and extracellular acidosis (6, 21, 33); and 3) a formulation of the ATP-dependent K+ current (IK,ATP) (14) was incorporated into the membrane model; its activation was regulated by the intracellular [ATP] and [ADP], which were assigned values of 6.8 mmol/l and 15 μmol/l, 5.7 mmol/l and 57 μmol/l, and 4.6 mmol/l and 99 μmol/l, in normoxia, ischemia stage 1, and ischemia stage 2, respectively (14, 45). Note that because the study focuses on early acute ischemia, gap junctional conductance was not altered as per experimental evidence (6).

Protocol for determining ULV.

The slice was paced at a basic cycle length of 250 ms at the anterior junction of the septum and the wall of the left ventricle (Fig. 1). This fast pacing rate was chosen based on the finding of Malkin et al. (27) that at these rates, the ULV better mimics the DFT. Truncated exponential monophasic shocks of 8 ms duration and 65% tilt were applied after the seventh pacing stimulus between a cathode located in the posterior right ventricle and a distant anode located in the bath to the left of the left ventricle (Fig. 1), as in experimental models of defibrillation (12, 13). Shocks of various strengths were applied at different coupling intervals (CI) to determine, for each CI, the highest shock strength above which sustained arrhythmia was no longer induced. The global maximum among these values represents the ULV. The ULV and the CI at which the ULV occurs (CIULV) were determined in normoxia and in stages 1 and 2 of early acute ischemia. Sustained arrhythmias were due to perpetuation of reentry. We considered an arrhythmia sustained if it consisted of more than two reentrant beats after the immediate postshock activation. Whereas this choice differs from criteria used in experimental studies (5), it was dictated by the stable reentrant pattern as well as by the large size of the model: longer times of simulation were prohibitive computationally. Shocks of strength directly above the ULV were often associated with a single beat (extra systole) after the immediate postshock activation. To better understand the mechanisms governing early postshock activity in ischemia, we also determined the highest shock strength above which this beat was no longer induced, upper limit of extra beat vulnerability (ULEB). Clearly, the ULEB was higher than the ULV.

The increment in shock strength used to scan the outcomes and determine the ULV and ULEB was 0.2A, whereas the CI time step was 20 ms. Finally, shock strength was the leading-edge value of the applied current.

Data analysis.

To analyze the electrical behavior of the myocardium before and after the shock, activation maps were constructed, and the action potential duration (APD) and effective refractory period (ERP) were measured. As per our implementation of global ischemia, APD and ERP under the conditions of pacing stimulation were the same at every node in the myocardium. APD was defined as the difference between the activation time and the time at which the action potential was 90% recovered (APD90). To measure the ERP, a pacing stimulus of twice diastolic threshold was used; the ERP was determined as the longest coupling interval that did not elicit propagation in the slice after the seventh paced beat. Activation times under pacing stimulation were calculated as the time interval between the onset of the last pacing stimulus and the time of maximum upstroke velocity of the action potential. After the shock, activation times were calculated as the time interval between the shock end and the time of maximum upstroke velocity of the action potential.

RESULTS

Action potential morphology and wavefront propagation during pacing.

Figure 2A presents activation maps after the seventh pacing stimulus in normoxia and in stages 1 and 2 of early acute ischemia. The activation patterns and thus conduction velocities are similar in normoxia and in ischemia stage 1, with the times of latest activation being 186 and 180 ms, respectively. This corresponds to only a 3% increase in conduction velocity from normoxia to ischemia stage 1. However, in ischemia stage 2 propagation is considerably slower than in the other two cases, and 250 ms after the seventh pacing stimulus the wave of excitation has not yet propagated through the entire slice. In ischemia stage 2, the latest activation time is 370 ms, and conduction velocity is reduced to 50.3% of its normal value.

Fig. 2.

A: activation maps in normoxia and in stages 1 and 2 of early acute ischemia after the seventh pacing stimulus. Isochrones are drawn each 27.8 ms from the onset of the seventh pacing stimulus (0 ms). Black color in the ischemia stage 2 maps represents activation times >250 ms. B: time course of the action potential at the node marked X in A. Vm, transmembrane potential.

Figure 2B depicts the action potential generated after the seventh pacing stimulus at the node marked with X in Fig. 2A. The main changes induced in the action potential by acute ischemia are the following: 1) depolarization of the resting potential from its normal value, −85.4 mV, to −74.1 and −66 mV in ischemia stages 1 and 2, respectively; 2) reduction of peak-to-peak action potential amplitude (APA) from 125.2 mV in normoxia to 105.6 and 87.7 mV, respectively; and 3) APD90 shortening from 138 ms in normoxia to 91 and 66 ms in stages 1 and 2 of early acute ischemia, respectively. Furthermore, the ERP also changes in acute ischemia compared with its normal value of 145 ms: in ischemia stage 1 it decreases to 115 ms, whereas in ischemia stage 2 it is prolonged to 180 ms. These results are consistent with previous experimental results (6, 44, 45).

ULV and ULEB.

Shocks of strength between 0.4 and 2.2A were applied at CIs from 40 to 260 ms to determine ULV and ULEB in normoxia and in stages 1 and 2 of early acute ischemia. For the normal myocardium, ULV and ULEB occur at CIULV of 220 ms and have values of 1.4A and 1.6A, respectively. Thus shocks of strength between 1.4 and 1.6A induce a single beat only, whereas sustained reentries are induced by shocks of strength below 1.4A. In ischemia stage 1, the ULV decreases to 0.8A, whereas the ULEB increases to 2.2A; both values are found at CIULV = 180 ms. In contrast, no arrhythmia, sustained or unsustained, is induced by any combination of shock strength and CI tested in ischemia stage 2.

To explore what mechanisms underlie the change in ULV due to acute ischemia, the present study compared, in normoxia and ischemia, the Vm distributions at the end of the shock, the shock-induced changes in action potential morphology, and the evolution of the postshock electrical activity in the myocardium for shocks below and above the ULV. The figures below present such comparisons for shocks delivered typically at the respective CIULV because these CIs correspond to the preshock state of the myocardium most susceptible to reentry induction at high shock strengths.

Vm distribution at shock end.

Figure 3A presents the Vm distribution at shock end on the top surface of the slice (Fig. 3A, left panel) and on a parallel plane through the middle of the slice (Fig. 3A, right panel) in normoxia and acute ischemia. The effect of the shock on the bottom surface of the slice is of opposite polarity (not shown). In the figure, the Vm distribution is shown for shocks of strength 0.8A and 1.4A. The shocks are applied at the respective CIULV in normoxia and ischemia stage 1 as to better compare the mechanisms resulting in ULV alteration. Because no ULV exists in ischemia stage 2, the shock outcome is independent of the CI. For comparison of postshock behavior, a CI of 240 ms was chosen in this case.

Fig. 3.

A: Vm distribution on the top surface of the slice (left) and on a parallel plane through the middle of the slice (right) at the end of shocks of strength 0.8 and 1.4A applied at the coupling interval (CI) at which the upper limit of vulnerability occurs (CIULV) in normoxia and ischemia stage 1 (220 and 180 ms, respectively) and at CI = 240 ms in ischemia stage 2. Color scale is saturated, i.e., Vms above +20 mV and below −90 mV appear red and blue, respectively. B: percentage of all myocardial nodes in the slice experiencing hyperpolarization below −90 mV (left) and depolarization above +20 mV (right) at shock end for the same shocks as in A.

The Vm distribution at shock end incorporates both the preshock state as well as the shock-induced VEP. It is the starting point of the postshock activity; differences in this distribution between normoxia and ischemia could manifest themselves as differences in shock outcomes, and thus as differences in electrical vulnerability. Figure 3A shows that, consistent with a previous study from our group (17), shock-end Vm on the surface (left) is stronger than that in the tissue depth (right). Furthermore, increasing shock strength increases the extent of the areas depolarized to more than +20 mV (red) or hyperpolarized below −90 mV (blue), particularly in the tissue depth (right panels). Finally, increasing the level of ischemia leads to a decrease in the extent of areas depolarized to more than +20 mV (red) and thus to an increase in the amount of tissue depolarized to intermediate Vm (colored in green and yellow). Differences in shock-end Vm distribution are quantified in Fig. 3B, which presents the percentage of all myocardial nodes in the slice experiencing depolarization above +20 mV and hyperpolarization below −90 mV as a function of shock strength and the level of acute ischemia. The histograms demonstrate that ischemic tissue exhibits smaller (in extent) shock-induced regions of positive polarization (Fig. 3B, right). These are regions of immediate postshock refractoriness and, hence, play a pivotal role in the outcome of the shock.

Shock-induced action potential morphology.

To further investigate the effect of acute ischemia on shock-induced changes in action potential morphology, the time course of Vm after the shock was examined at various locations in the slice. To illustrate our main findings, Fig. 4 presents the time course of Vm at the nodes marked a and b in Fig. 3A. Nodes a and b are representative of shock-induced positive and negative polarization, respectively. The time when shock occurred is marked by a thick black line.

Fig. 4.

Time course of Vm at nodes a and b within the tissue depth as shown in Fig. 3A in normoxia and in stages 1 and 2 of acute ischemia for the same shocks as in Fig. 3A. Time when shock was applied is indicated by the thick black line.

In normoxia, node a is at 32% repolarization when the shock is applied. Both shocks induce depolarization that prolongs APD90. Shock-induced changes in VmVm) and in APD90 (ΔAPD90) are 12 mV and 8 ms for the 0.8A shock, and 23 mV and 13 ms for the 1.4A shock. In ischemia stage 1, node a is at 27% repolarization; ΔVm is 10.7 and 21.5 V for the 0.8A and 1.4A shocks, respectively. However, despite the shock-induced depolarization, APD90 in this case is shortened to 89 ms for the 0.8A shock (ΔAPD90 = −2 ms) and does not change for the 1.4A shock (ΔAPD90 = 0 ms). In ischemia stage 2, the shock itself elicits an action potential at node a, which is of APD90 = 66 ms (same APD90 as the paced beat, i.e., ΔAPD90 = 0 ms) and APA = 79 mV for the 0.8A shock, and of APD90 = 62 ms (ΔAPD90 = −4 ms) and APA = 88.3 mV for the 1.4A shock. These data illustrate that shock-induced depolarization in normoxia extends the APD, whereas in acute ischemia, the APD is either shortened or unchanged. This is consistent with previous observations of shock effects in normoxia (40) and acute ischemia (23).

The right panel of Fig. 4 shows that node b is negatively polarized by the shock in all cases. The negative ΔVm induced by shocks of strength 0.8 and 1.4A is, respectively, 8.7 and 13.7 mV in normoxia, 6.8 and 10.8 mV in ischemia stage 1, and 6.6 and 9.8 mV in ischemia stage 2; and the more negative Vm reached during the 0.8 and 1.4A strength shocks is, respectively, −92 and −97 mV in normoxia, −79.5 and −83.5 mV in ischemia stage 1, and −70.8 and −74.2 mV in ischemia stage 2. APD90 is unchanged in all cases because the shock is applied while node b is at resting potential. These data illustrate our finding that shock-induced negative polarization increases with increasing shock strength in all three cases, but is reduced with increase in the level of ischemia for both shock strengths (due primarily to ischemia-induced resting potential elevation).

Postshock electrical activity in normoxia and ischemia stage 1.

Figure 5 illustrates the evolution of the postshock electrical activity in the normal myocardium (Fig. 5A) and in stage 1 of early acute ischemia (Fig. 5B) after weak shocks. Shown is an example of a 0.8A shock. Maps of Vm on the top surface and on a plane through the middle of the slice are presented. When the shock is turned off, wave front propagation starts at the boundaries between areas of opposite polarization. Therefore, 15 ms after the shock end, the majority of the slice is depolarized and only a small amount of tissue remains excitable (Fig. 5, 15-ms panels).

Fig. 5.

Evolution of postshock electrical activity in the normal myocardium (A) and in stage 1 of early acute ischemia (B) after a 0.8A shock applied at CIULV (220 and 180 ms, respectively). In each panel, maps of Vm distribution on the top surface of the slice (left) and on a plane through the middle of the slice (right) are presented. White arrows indicate direction of wavefront propagation, whereas black arrows in the 30, 60, and 200 ms panels refer to events discussed in the text. Color scale as in Fig. 3A.

Figure 5 shows that 30 ms after shock end in both normoxia and ischemia stage 1, all postshock activation waves are blocked, except for two that propagate through tissue in the depth: one at the posterior junction of left ventricular free wall and septum, and another in the right ventricular wall (arrows in 30-ms panels). These activations break through the surface, as indicated by the arrows in the 60-ms panels. Simultaneously, tissue in the areas depolarized by the shock repolarizes earlier in ischemia stage 1 than in normoxia (30 and 60 ms panels). The faster recovery of the tissue in ischemia stage 1 leads to propagation of these activations in all directions (60 and 120 ms panels in Fig. 5B, as well as isochrone map in Fig. 6, right), until all activity ceases following wavefront collision (Fig. 5B, 200 and 280 ms panels). In normoxia, however, the majority of tissue within the left ventricular free wall is still refractory 60 ms after shock end, and thus activity propagates only toward the posterior side of the slice (Fig. 5A, 60 ms panel, and isochrone map in Fig. 6, left). Because of this transient unidirectional block, a reentrant circuit is established (Fig. 5A, 200 and 280 ms panels, and Fig. 6, left) with direction of rotation shown by the curved black arrow in the 200 ms panel, Fig. 5A. Clearly, the 0.8A shock is below the ULV in normoxia and induces sustained reentry, whereas in ischemia stage 1 it is above the ULV and below the ULEB and thus it is followed by a single extra beat only.

Fig. 6.

Activation maps on a plane through the middle of the slice following a 0.8A shock applied at CIULV (220 and 180 ms, respectively) in normoxia (left) and in stage 1 of early acute ischemia (right). Isochrones are drawn every 20 ms from 30 to 150 ms and at 160 ms postshock.

Finally, the activation maps on a plane through the middle of the slice shown in Fig. 6 demonstrate that similar to the preshock, the postshock conduction velocity also remains largely unchanged in ischemia stage 1, as evidenced by the similar distance between isochrones in the two cases.

Figure 7 illustrates an example of the evolution of postshock electrical activity in the normal myocardium (Fig. 7A) and in stage 1 of early acute ischemia (Fig. 7B) after strong shocks. The shock strength here is 1.8A. Again, Vm maps on the top surface of the slice and on a plane through the middle of the slice are presented. As shown in Fig. 7A, when the shock is turned off, break excitation waves initiated at shock end quickly traverse the postshock excitable gap in the normal myocardium, and the entire slice becomes refractory (15 and 30 ms panels). Propagation is blocked and no arrhythmia is induced (60, 120, and 200 ms panels). This behavior is representative of shocks of strength above the ULEB.

Fig. 7.

Evolution of postshock electrical activity in the normal myocardium (A) and in stage 1 of early acute ischemia (B) after a 1.8A shock applied at CIULV (220 and 180 ms, respectively). In each panel, maps of Vm distribution on the top surface of the slice (left) and on a plane through the middle of the slice (right) are presented. Arrow in the 30-ms panel refers to events discussed in the text. Color scale is the same as in Fig. 3A.

In ischemia stage 1, break excitations traverse the postshock excitable gap after the shock (Fig. 7B, 15 ms panel). However, 30 ms after shock end, activity continues through the tissue in depth of the right ventricular wall (arrow in the 30 ms panel, and 60 ms panel), and breaks through the surface, resulting in an extra beat (120 and 200 ms panels).

Postshock electrical activity in stage 2 of early acute ischemia.

Figure 8A illustrates the evolution of the postshock activity in ischemia stage 2 for a 1.0A shock applied at CI = 240 ms. When the shock is turned off, break excitations ensue. However, 15 ms later, activation is halted and the tissue begins to recover (30 ms panel), with the entire slice reaching rest 70 ms after the shock is turned off. Similar behavior is observed in ischemia stage 2 after shocks of strength between 0.4A and 2.2A applied over a wide range of CIs.

Fig. 8.

A: evolution of postshock electrical activity in stage 2 of early acute ischemia after a 1.0A shock applied at CI = 240 ms. Maps of Vm distribution on the top surface of the slice (left) and on a plane through the middle of the slice (right) are presented. Arrows in the 4, 7, and 15 ms panels and circles in the 7 ms panel refer to mechanisms discussed in the text. Color scale as in Fig. 3A. B: time course of Vm in nodes 13 located within the small square in the 7 ms panel. Time when the shock was applied is indicated by the thick black line. The orange arrow represents direction of propagation from nodes 13.

Our simulations demonstrate that two mechanisms are responsible for the lack of propagation after 15 ms. First, in the tissue depth, where VEP is weaker than that on the surface, there are excitable gaps surrounded by depolarized tissue that never become activated (circles in the 7 ms panel of Fig. 8A). The reduced excitability of the tissue in ischemia stage 2 decreases the likelihood of break excitation after the shock. Second, Fig. 8A shows that what appear to be leading edges of the activations change color from red to yellow to green in the 4, 7, and 15 ms panels, respectively (e.g., wave front marked with arrows). This demonstrates decremental conduction; this behavior is highlighted in Fig. 8B where the time course of Vm is presented for nodes 13 located in the septum within the region outlined by the small square in the 7 ms panel of Fig. 8A. The orange arrow indicates direction of decremental conduction from node 1 to node 3. After the shock, the maximum Vm reaches 12, −10, and −58 mV for nodes 1, 2, and 3, respectively, indicating propagation failure.

DISCUSSION

The present study uses a realistic model of cardiac vulnerability to electric shocks in an attempt to provide mechanistic insight into the effect of 5 to 12 min of acute global ischemia on the ULV. This is the first time that a realistic 3-D model of stimulation/defibrillation involving anatomically accurate ventricular geometry and fiber architecture also incorporates membrane kinetics under the conditions of acute myocardial ischemia. In this study, acute ischemia is assumed to affect the membrane kinetics of every cell in the myocardium, thus simulating global ischemia. Our results show that if early acute ischemia is in its initial stage, ULV decreases by 42.8%, whereas ULEB increases by 37.5% compared with normoxia. In contrast, no arrhythmia is induced by any combination of shock strength and CI tested in stage 2 of early acute ischemia. To unravel the mechanisms underlying these changes in cardiac vulnerability, the electrophysiological behavior of the tissue before and after defibrillation shocks is compared in normoxia, and in stages 1 and 2 of early acute ischemia. In our simulations, as in previous studies (6), acute global ischemia modifies the electrophysiological properties of the tissue resulting in decreased APD90 and APA, increased postrepolarization refractoriness, and altered propagation velocity. Another fact important to vulnerability is that ischemia results in a different distribution of transmembrane potential at shock end, which in turn sets off a different postshock electrical activity in the tissue. In acute ischemia stage 1, shortened postshock refractoriness resulting mainly from 1) the ischemic electrophysiological substrate and 2) the decrease in the extent of areas positively polarized by the shock is the main factor responsible for the changes in both ULV and ULEB. In contrast, conduction failure underlies the lack of vulnerability in stage 2 of early acute ischemia. These mechanisms are discussed in detail in the sections to follow.

Action potential morphology and wavefront propagation.

In this study, acute ischemia is represented by the effects of its three major components, hyperkalemia, hypoxia, and acidosis, on [K+]o, IK,ATP activation, and INa and ICa,L inhibition. Previous studies (6) have demonstrated that these alterations in behavior at the ionic level account for the changes in action potential morphology caused by early acute ischemia, namely APD90 shortening, depolarization of resting potential, and decreased APA. In this study, these changes in ionic activity are represented in two degrees: after 10–12 min (stage 2 of ischemia) and 5–7 min (stage 1 of ischemia) of complete occlusion of blood flow (45, 47) (the latter can also be achieved after 10–15 min of partial reduction of coronary flow) (5).

Changes in ionic currents caused by acute ischemia have been shown to modify refractoriness and propagation velocity in cardiac tissue (6). Our simulations demonstrate that ERP changes from 145 ms in normoxia to 115 and 180 ms in stages 1 and 2 of ischemia, respectively, and is associated with APD90 shortening and increased postrepolarization refractoriness. APD90 shortening is due to IK,ATP activation (14, 45), whereas increased postrepolarization refractoriness is caused by the slow recovery of the Na+ inactivation gates under conditions of resting depolarization, primarily due to increased [K+]o (36). Depending on the degree of IK,ATP activation and [K+]o increase, ERP in acute ischemia can decrease, if the increase in postrepolarization refractoriness is smaller than the decrease in APD90 (ischemia stage 1), or increase when postrepolarization refractoriness is significantly prolonged (ischemia stage 2).

In our simulations, conduction velocity is 103 and 50.3% of its normal value in stages 1 and 2 of ischemia, respectively. Slower conduction in ischemia stage 2 is caused by decreased INa due, first, to slower recovery of the Na+ inactivation gates after resting depolarization of cell membrane and second, to decreased INa conductance caused by intracellular and extracellular acidosis (6, 36). A decrease in INa also occurs in ischemia stage 1, although to a smaller extent. However, its effect on conduction velocity is counteracted by the reduction in the potential difference between rest and threshold for activation (6). Therefore, propagation velocity in ischemia stage 1 remains similar to that in normoxia.

In summary, in ischemia stage 1, ERP is shortened whereas propagation velocity remains largely unchanged. In ischemia stage 2, refractoriness is prolonged whereas conduction velocity is significantly decreased. These electrophysiological changes play an important role in the postshock electrical behavior, and thus in the mechanisms of cardiac vulnerability to electric shocks, as discussed below.

Stage 1 of early acute ischemia: changes in immediate postshock electrical behavior.

The immediate postshock electrical behavior in the myocardium is different in ischemia stage 1 compared with normoxia. The ischemic conditions alter the transmembrane potential distribution at the end of the shock (as in Fig. 3A), as well as the shock-induced changes in action potential morphology (as in Fig. 4), mainly APD90. Indeed, as a result of ischemia stage 1, first, a smaller amount of tissue experiences shock-induced depolarization and thus APD extension, and second, the value of this APD extension, ΔAPD90, is reduced throughout the preparation. The latter finding is consistent with the experimental results of Knisley and Holley (23).

The changes in the impact of the shock as described above, combined with the intrinsic ischemia-related shortening in ERP (as discussed in the previous section), result in an overall faster postshock recovery of excitability in the ischemic tissue compared with the normal myocardium. On the other hand, postshock conduction velocity in ischemia stage 1 remains similar to the one in normoxia (see activation maps in Fig. 6). Therefore, ischemia stage 1 alters the immediate postshock activity in the myocardium by shortening postshock refractoriness. As a consequence, higher shock strengths are needed in the ischemic tissue to induce break excitations quickly enough so that these excitation waves can traverse the postshock excitable gaps before the adjacent areas recover from depolarization. This is the mechanism that underlies the increase in ULEB in ischemia stage 1.

Stage 1 of early acute ischemia: mechanisms underlying the decrease in ULV.

Whereas stronger shocks are needed to halt propagation immediately after the shock in ischemia stage 1, shock-induced reentry and thus sustained arrhythmias are more likely to occur in normoxia, as demonstrated by this investigation. Indeed, we found that ULV decreases from 1.4A in normoxia to 0.8A in ischemia stage 1. In normoxia, shock-induced extension of refractoriness provides the unidirectional block necessary for the establishment of a reentrant circuit, as determined in previous research (12, 37). In ischemia stage 1, the quicker recovery of the myocardium in the areas depolarized by the shock as well as the smaller extent of these areas decreases the likelihood of unidirectional block, resulting in a decrease in the ULV.

We are aware of only a single study that investigated the effect of acute global ischemia on ULV (5); the experiments were conducted after 10–15 min of partial obstruction of coronary flow in isolated Langendorff-perfused rabbit hearts. In the majority of these experiments (6 of 10), ULV significantly decreased; our results are consistent with this finding. However, the study by Behrens et al. (5) found that ULV slightly increased in three hearts and remained unchanged in one. In the period during which the measurements were performed (from 10 to 15 min after occlusion), APD90 decreased from 77 to 66% of its normal value, and activation times were, on average, ∼15% larger, indicating slowed propagation. Therefore, the electrophysiological conditions of the hearts could have been significantly different among experiments. This could explain the disparity of experimental results because different combinations of changes in refractoriness and propagation velocity could lead to differences in ULV.

Lack of vulnerability in stage 2 of early acute ischemia.

Our simulations demonstrate that both ULV and ULEB drop to zero under the conditions of ischemia stage 2: no arrhythmia or even a single extra systole could be induced, regardless of shock strength or CI. This is in agreement with the experimental findings of Babbs et al. (4), where a gradual decrease in DFT was observed with increasing [K+]o: DFT reached zero (spontaneous defibrillation) when [K+]o increased to 16.6 mmol/l. In our simulations, for all combinations of shock strength and CI tested, propagation initiated at shock end terminates shortly thereafter due to conduction failure. In acute ischemia, hyperkalemia and acidosis lower the excitability of cardiomyocytes (6, 34), whereas decreased APA reduces the current available for the stimulation of myocytes downstream. As a consequence, the safety factor for propagation decreases leading to conduction block under the conditions of severe depolarization of the resting potential (35).

Implications.

In the present study, ischemia-induced alterations in membrane kinetics take place in every cell in the myocardium. However, in diseased hearts, acute ischemia typically occurs only in the myocardial region supplied by the occluded blood vessel. In this case, the coexistence of both normal and ischemic myocardium leads to spatial heterogeneities in refractoriness and conduction velocity that have been proven to favor the establishment of reentrant circuits (10, 11, 19). Thus the ULV is expected to increase in regional compared with global ischemia, with the exact value depending first, on the degree of ischemia, and second, on how proarrhythmic is the spatial distribution of heterogeneities in each particular case of regional ischemia. In any case, if the postshock activations are not immediately halted, then even a single surviving wavefront, which is benign in global ischemia (such as the extra systoles observed in this study for shocks above the ULV but below the ULEB), can be expected to become reentrant after encountering spatial heterogeneities.

One can speculate, therefore, that in regional ischemia, for reentry not to be induced, shocks need to be above the highest ULEB in the tissue. That is, if the affected region had just become ischemic, then the shocks need to be above the ULEB for global ischemia stage 1 because, as determined here, it is higher than the ULEB in the normal myocardium. Thus it can be speculated that the ULV for regional ischemia in this case would be below the ULEB for global ischemia. In contrast, if ischemia in the affected region is more advanced (as in stage 2), then the shock needs to be above the normoxic ULEB to not induce reentry because as demonstrated here, postshock activations then fail to conduct. Thus the ULV for this degree of regional ischemia is expected to be below the ULEB for the normal heart. Consequently, stage 1 ULEB and the normoxic ULEB can be perceived as the upper limits of the ULV in regional ischemia stages 1 and 2, respectively. This speculation is made here with caution because the extent of the ischemic region is likely to also play a role in vulnerability to electric shocks: small ischemic regions may be of insufficient spatial extent to develop their own postshock activations, and thus the concept of a “regional” ULEB may not be applicable to them.

The model.

In this study, we use an anatomically accurate 3-D bidomain finite-element model based on canine ventricular anatomy to examine the acute ischemia-induced changes in cardiac vulnerability to electric shocks. In the model, not only are bidomain and membrane kinetics equations solved on a large 3-D domain, but the model also incorporates an anatomically precise representation of fiber architecture and canine ventricular geometry that guarantees realistic shock-induced VEP (17). The model has been previously used by our group to investigate the effect of electric shocks in the normal myocardium (17, 42). Here we extend the capabilities of the model by incorporating the changes in membrane kinetics associated with myocardial ischemia. Thus, we offer, for the first time, a tool to investigate postshock dynamics in the ischemic heart. Previous computer simulation of cardiac electrical behavior in ischemia have only implemented monodomain models of simplified tissue structure in one and two dimensions (14, 34) or considered solely hyperkalemia (1, 31). Thus our model presents a powerful new test bed to dissect the complex intrinsic electrophysiological interactions taking place in ischemic tissue.

Because of issues of computational tractability, the model used here represents only a slice of the canine ventricles and not the organ in its entirety. However, the validity of the mechanisms uncovered here is not limited by the geometry of the model. It is the changes in membrane dynamics due to acute myocardial ischemia that underlie the findings of this simulation study. Our simulations predict that global acute ischemia would cause the exact same effects in any preparation that could sustain reentry and that had geometry and fiber architecture creating significant VEP throughout the tissue.

Limitations.

The limitations of the model, and in particular, the fact that because of computational constraints the thickness of the canine ventricular slice cannot be increased beyond 4 mm, have been outlined in detail in previous publications (17, 42). Here we discuss only limitations specific to this study. Acute ischemia is represented by IK,ATP activation, increased [K+]o, and decreased maximum conductances of INa and ICa,L. These mechanisms are responsible for the majority of changes in action potential morphology and propagation in the first 10–12 min after coronary obstruction (6, 34). However, other changes at the ionic level occurring during ischemia, such as inhibition of the time-independent K+ current or transmural heterogeneity of ischemic alterations in membrane dynamics (6), may contribute to the mechanisms underlying the ULV. Furthermore, as discussed above, in this research we investigate the effects of global rather than regional ischemia and thus we do not take into account the spatial heterogeneities that are present under clinical conditions. The development of a model of regional ischemia in a stimulation/defibrillation model is of great complexity, and it is the subject of a forthcoming study. Finally, the electrode configuration was not changed in this study. On the basis of the study by Hillebrenner et al. (16), alternate electrode configuration is expected to only affect the value of the ULV and not the mechanisms underlying it. Despite these limitations, this study provides a wealth of information regarding the mechanisms that underlie the changes in electrical vulnerability during acute myocardial ischemia.

In conclusion, this research investigates the effect of altered membrane kinetics in stages 1 and 2 of early acute ischemia on the mechanisms underlying the ULV. Our results show that, under conditions of ischemia stage 1, shortened postshock refractoriness of the tissue plays a major role in both the decrease in ULV and the increase in ULEB. In ischemia stage 2, conduction failure is the main cause for the lack of vulnerability. The information provided by this study is pivotal to understanding how acute ischemia-induced changes at the ionic level modulate cardiac vulnerability to electric shocks. These findings could be significant to the clinical practice of defibrillation because the majority of patients that undergo defibrillation suffer from myocardial ischemia.

GRANTS

This work was supported by National Grant HL-063196 and American Established Investigator Award (to N. Trayanova), by grants from the Whitaker and the Keck Foundations (to J. C. Eason); and from Plan Nacional de Investigación Científica y Desarollo Tecnológico of the Ministerio de Ciencia y Tecnología (TIC 2001-2686) and Programa de Incentivo a la Investigación of Universidad Politécnica de Valencia (Spain), and by a postdoctoral fellowship (CTESPP/2002/69) from the Oficina de Ciencia y Tecnología of Generalitat Valenciana (Spain) (to B. Rodriguez).

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

View Abstract