- Email: [email protected]
Mechanisms in T-wave alternans caused by intraventricular block
Journal of Electrocardiology Vol. 33 No. 4 2000
Mechanisms in T-Wave Alternans Caused By Intraventricular Block
J. A. Abildskov, MD, and Robert L. Lux, MD
Abstract: It is recognized that 2:1 intraventricular (IV) block can result in T-wave alternans but is usually assumed that it would also affect QRS waveform. Block in a local region is not, however, varied activation sequence of the same muscle mass because the blocked region is not activated and is not part of the mass that is activated in cycles without block. Also, the block region may have electrocardiogram (ECG) effects when its state differs from other regions. In view of those considerations, the ECG effects of IV block were evaluated by using a computer model of excitation and recovery. ECGs were calculated from differences between the excited state and various degrees of recovery. Results provided evidence that boundaries associated with regions of block rather than regions having varied activation sequence were the major factors in T-wave alternans caused by IV block. Effects of the boundaries included cancellation of the effects of IV block on QRS complexes. Findings suggest that IV block cannot be excluded as a mechanism of T-wave alternans in the absence of QRS alternans. Key words: Activation sequence, QRS alternans, computer model.
One of the possible mechanisms of T-wave alternans is 2:1 block of excitation in a portion of the ventricular muscle (1,2). It has usually been assumed that such block would affect QRS as well as T waveform and that the absence of QRS differences in alternate complexes was evidence that this mechanism was not responsible for T-wave alternans (3–7). If the mechanism was simply that of different activation sequence in alternate com-
plexes, QRS as well as T-wave alternans would be obligatory. However, conditions during 2:1 block of a localized region are not simply those of varied activation sequence. With such block, there is alternating presence and absence of excitation in the region of the block rather than varied activation sequence of the same muscle mass. Also, when a blocked region is not excited, it is still present in a refractory state and may have electrocardiogram (ECG) effects when it is adjacent to excitable muscle or when it becomes excitable and is adjacent to refractory muscle. These considerations suggest that there is not an obligatory relation between QRS and T-wave effects of localized intraventricular (IV) block and that mechanisms other than activation sequence variation may be responsible for the ECG effects of a conduction defect (CD) on ECG waveform. In this study, a computer model of propagated
From the Nora Eccles Harrison Cardiovascular Research and Training Institute, and the Division of Cardiology, University of Utah, Salt Lake City, UT. Supported by awards from the Nora Eccles Treadwell Foundation and the Richard A. and Nora Eccles Harrison Fund for Cardiovascular Research; and grant HL52338 from the National Institutes of Health (Heart, Lung and Blood Institute). Reprint requests: J. A. Abildskov, MD, University of Utah, CVRTI, 95 South 2000 East, Salt Lake City, UT 84112-5000. Copyright © 2000 by Churchill Livingstone® 0022-0736/00/3304-0002$10.00/0 doi: 10.1054/jelc.2000.18108
311
312 Journal of Electrocardiology Vol. 33 No. 4 October 2000 excitation was used to elucidate the mechanism of ECG effects caused by CDs. Findings suggest that boundaries of potential difference at the site of CDs are the major factor. They also suggest that T-wave alternans caused by CDs can occur with only slight variation of QRS waveform and the mechanism cannot be excluded by failure to detect QRS alternans.
Materials and Methods The computer model has been used in previous studies of reentrant arrhythmias and ECG waveform (8 –18). It consisted of a matrix of 25 ⫻ 25 individual units each having 6 neighbors to which excitation could be propagated. Time was represented as steps with a single time step (TS) being the time required to transfer excitation from an excited unit to a fully excitable neighbor. After excitation, units became refractory for a period defined by the relation K公CL, in which K was a property assigned to each unit, CL was the preceding cycle length in the unit, and the square root was an approximation of the CL–refractory period (RP) relation in cardiac muscle. After the RP, units became excitable but exhibited slow propagation represented by the requirement of 4, 3, and 2 TSs for transfer of excitation to units in successive stages of 4-TS duration each. ECGs were calculated by determining instantaneous source vectors for each matrix unit based on state differences. For the QRS complex, vectors from excited units in the direction of excitable neighbors were defined and summed to yield the source vector for each unit during the TS in which that unit was excited. Source vectors at all other TSs were defined from state differences and directed from less toward more fully recovered units. During recovery, 3 states, each of which represented one third of the difference between inexcitable and fully recovered units, were considered. Potentials at each of 2 sites located near the matrix were calculated by superposing potentials from all source vectors and assuming an infinite homogeneous medium. Potential differences at the 2 sites were taken to yield a bipolar ECG lead. Importantly, ECGs calculated only from units excited during the preceding TS and ECGs calculated from state differences during all other TSs were displayed both individually and in combination. This allowed differentiation of ECG effects caused exclusively by activation from those caused by activation in the presence of a CD during the time that it was refractory. ECGs were also calculated with the CD region absent from the
matrix. These showed QRS features when activation in the CD was blocked and T-wave features secondary to that activation. For the purposes of this study, the matrix was considered to represent either the epicardial or endocardial ventricular surface. Conduction defects in these layers resulted in unclosed boundaries of potential difference when they were in a different state than the surrounding matrix. These boundaries resulted in ECG deflections that occur as a consequence of activation or recovery boundaries that intersect cardiac surfaces (collision). Conduction defects at varied locations were created by K values that resulted in 2:1 propagation in the defects for the assigned rate of stimulation. In all cases, the initial stimulus was applied when the entire matrix, including the CD, was in a fully excitable state after an assigned prior CL. K values were selected such that regular stimuli at a CL shorter than the initial assigned one resulted in alternate block and conduction in the CD with propagation at full velocity in the remainder of the matrix. K values outside the CD were either uniform or arranged in layers such that the sequence of recovery was opposite that of excitation and QRS and T deflections had the same polarity. The matrix with stimulus site, ECG electrode sites, and CD locations are shown by the diagram in Figure 1. The CDs consisted of 21 units in a 3 ⫻ 5 array at the location shown in the diagram. The diagram shows CD location relative to stimulus and ECG electrode sites, but not the relative size of CDs and the total matrix. The ECG electrode sites were located at a distance equal to the dimensions of 6 units above the matrix.
Results The principal finding was evidence that a boundary of potential difference between a CD and the surrounding matrix resulted in T-wave alternans with only slight variation of QRS waveform. As with other sources of ECG deflections, effects of the boundary depended on its dimension, location, and polarity, as well as the ECG lead in which effects were observed. CD-Matrix Boundary CD-matrix boundaries occurred as a consequence of differences between states of the CD and surrounding matrix such as a refractory CD and recovered matrix. A boundary of opposite polarity ex-
IV Block T-Wave Alternans •
Fig. 1. Diagram of the matrix used in the study. Electrode sites from which an ECG lead was calculated are shown at X and Y. The site of stimulation is indicated by S. Conduction defects are shown at A near the origin of excitation, B midway between ECG electrode sites, and C distant from the origin of excitation.
Fig. 2. Conditions simulated in the study. The cardiac surface represented by the matrix is shown with CD indicated by the dotted lines. The poles of an ECG lead are shown as X and Y. When the state of the CD and surrounding matrix differed (ie, 1 excitable and the other refractory) the CD was a source of ECG deflection related to the relative magnitude of the angles shown.
Abildskov and Lux 313
isted when the CD was in a recovered state while the surrounding matrix was refractory, and the boundary disappeared when the CD and matrix were in the same state. As with other boundaries of potential difference that intersect a cardiac surface, the CD-matrix boundaries resulted in ECG deflections. The conditions simulated in this study are shown in Figure 2. The cardiac surface shown on edge is represented by the solid line shown with a CD represented by the dotted line at various locations. An ECG lead with its poles (X, Y) at a distance from the cardiac surface is shown. When the CD was refractory and the surrounding matrix excitable, or the CD excitable and matrix refractory, the CDs were the source of ECG deflections. In the bipolar lead shown, deflections were related to the relative magnitude of the angles at the poles. As shown in Figure 2A, the angle at X was larger than that at Y, resulting in an ECG deflection. Figure 2B shows a CD midway between the ECG poles so angles at the poles were equal and no deflection occurs. Figure 2C shows a CD located near the lower ECG pole (Y), resulting in unequal angles at X and Y, and an ECG deflection of opposite polarity to that resulting from conditions shown in Figure 2A. In addition to the CD location, the polarity of ECG deflections caused by a CD-matrix boundary depended on the relative state of the CD and surrounding matrix (ie, which was recovered or refractory) and on the polarity of the ECG lead. ECG deflections caused by a CD-matrix boundary are shown by the diagram in Figure 3 together with diagrammatic QRST complexes to show the timing
314 Journal of Electrocardiology Vol. 33 No. 4 October 2000
Fig. 3. Diagram of ECG deflections caused by a CD-matrix boundary with the CD located near the origin of excitation. The diagrammatic QRS and T waves show the time of occurrence of deflections caused by the CD-matrix boundary. The deflection at (A) was caused by a refractory CD and surrounding matrix that had recovered. At (B), the matrix adjacent to 1 edge of the CD was excited and became refractory, at (C) the other edge of the CD was excited and became refractory. At (D), the CD recovered whereas the matrix remained refractory, and at (E), the matrix at 1 edge of the CD recovered. At (F), the matrix at both edges of the CD have recovered and no CD-matrix boundary exists.
of deflections caused by the boundary. In this example, it has been assumed that recovery properties outside the CD were uniform so the sequence of repolarization was the same as that of depolarization and QRS and T deflections have opposite polarity. It has also been assumed that the CD was located near the site of origin of excitation. The first QRST complex represents one in which excitation occurred in the entire matrix and the second complex is one in which excitation in the CD was blocked. The deflection with its onset at (A) was caused by a boundary between the refractory CD and recovered matrix adjacent to the CD. The deflection at (B) occurred when the matrix adjacent to 1 edge of the CD was excited and became refractory, leaving the boundary at the other edge unopposed. At (C), excitation reached the other edge of the CD so both the CD and surrounding matrix were refractory and no CD-matrix boundary was present. At (D), the CD recovered whereas the matrix remained refractory resulting in a boundary of opposite polarity to the initial one. At (E), the matrix at 1 edge of the CD has recovered and the deflection is the effect of the remaining edge of the recovered CD-refractory matrix boundary. At (F), the matrix beyond the CD has recovered and no CD-matrix boundary exists. Deflections caused by the CD-matrix boundary were superposed on QRST complexes with alternate conduction and block in CDs. The detailed form of deflection caused by the boundaries and their time phase with QRST deflection depended on a variety of factors that will be shown in subsequent examples.
Figure 4 shows a computer simulation of the condition assumed for the diagram in Figure 3. The assigned prior CL was 200 TS and stimuli were applied at 100-TS intervals, which resulted in activation of the CD in alternate cycles. During the first QRS complex (Figure 4A), the entire matrix including the CD was activated. Because the CD was located near the stimulus site, it was activated during the early portion of the complex. Recovery properties outside the CD were uniform so the matrix surrounding the CD recovered early in the first T wave and the resulting CD-matrix boundary affected the remainder of the T wave. The boundary resulted in an upward deflection, which superposed on the T wave decreased the amplitude of the negative peak. The upward deflection alone is evident after the first T wave. During the second QRS complex, the CD continued to be refractory and propagation there was blocked. When excitation reached 1 edge of the CD, however, the boundary between refractory and recovered regions included the CD as well as the neighboring matrix. This was similar to the boundary when the CD was excited and the resulting QRS was similar to that of the first complex. After the second QRS complex, the CD recovered whereas the surrounding matrix was refractory, resulting in a CD-matrix boundary of opposite polarity to the initial one. That boundary was altered, then removed as the matrix surrounding the CD recovered so the second T wave was not affected by a CD-matrix boundary. In subsequent complexes there was alternating conduction and block in the CD and repetition of the events de-
IV Block T-Wave Alternans •
Abildskov and Lux 315
Figure 4B–D. Figure 4B shows QRS complexes calculated from effects of activation only without effects of the CD-matrix boundary. Alternating waveform caused by alternating conduction and block in the CD was present. Figure 4C shows deflections calculated from events exclusive of excitation. These include T waves and also effects of the CD-matrix boundary during activation in the complexes with block in the CD. Those deflections superposed on the QRS complexes shown in Figure 4B resulted in complexes similar to those without block. Figure 4D shows QRS and T deflections calculated from a matrix in which the region of the CD was absent. In this record, activation and recovery of the CD region and a CD-matrix boundary do not contribute to QRS or T waveform. CD Locations
Fig. 4. Calculated ECG deflections from a matrix with 2:1 propagation in a CD located near the origin of excitation. (A) shows QRST deflections with T-wave alternans and minor variation of QRS complexes. (B) shows QRS complexes calculated from effects of activation only with alternating QRS waveform caused by alternating failure to activate the CD. (C) shows deflections calculated from events other than excitation showing T-wave alternans and effects of the CD-matrix boundary during activation. (D) shows QRST deflections calculated from a matrix in which the region of the CD was absent.
scribed in the first 2 complexes. The CD-matrix boundary affected T-wave amplitude in complexes in which the CD was activated but not in complexes with block of the CD. It should be noted that if the CD had not recovered before the T wave of complexes with block, T-wave amplitude of those complexes as well as those with conduction in the CD would be affected by a CD-matrix boundary and T-wave alternans would not be present. The mechanisms by which a CD-matrix boundary affected ECG waveform are further shown by
The location of CDs affected the time of onset and duration of deflections caused by CD-matrix boundaries and the relation of the boundaries to ECG leads. When CDs were equidistant from the poles of a lead, the boundary had equal effects on the poles and did not produce deflections resulting in T-wave alternans. During the complexes with block of the CD, however, the CD-matrix boundary still contributed to activation and recovery boundaries outside the CD, so QRS and T deflections were only slightly affected by the block. These conditions are shown in Figure 5. Figure 5A shows QRST complexes without T-wave alternans and with similar QRS complexes with and without CD block. Figure 5B shows QRS complexes from activation alone showing effects of CD block in alternate cycles. Figure 5C shows deflections caused by events other than excitation including the contribution of the CD-matrix boundary to the QRS complex in cycles with block of the CD. Figure 5D shows deflections when the region of the CD was absent from the matrix and shows both QRS and T waveform without effects of the CD or CD-matrix boundary. Figure 6 shows effects of a CD distant from the origin of excitation. In that condition, the CD was not excited until late in the QRS in cycles without block of the CD. In the example shown, recovery properties outside the CD were uniform so the matrix surrounding the CD did not recover until late in the first T wave. The amplitude of that wave was, therefore, not affected by a CD-matrix boundary. After the onset of the boundary late in the first T wave, it continued to exist until late in the second QRS when the matrix surrounding the CD was
316 Journal of Electrocardiology Vol. 33 No. 4 October 2000 Graded Recovery Properties The distribution of recovery properties affected the time phase of deflections caused by a CD-matrix boundary and those caused by other events. Examples of the effect of CDs at various locations in a matrix with graded recovery properties are shown in Figure 7. In these examples, 5 layers of K values ranging from 2 to 6 with the highest value near the stimulus site were assigned. With those conditions, the sequence of recovery was opposite that of excitation and QRS and T deflections had the same polarity as they had with normal ventricular excitation. Figure 7A shows effects of a CD near the
Fig. 5. Calculated ECG deflections from a matrix with 2:1 propagation in a CD located midway between the poles of the ECG lead. (A) shows QRST deflections with no alternans of either QRS or T. (B) shows QRS complexes calculated from events during activation only and showing effects of CD block in alternate cycles. (C) shows deflections caused by events exclusive of excitation including effects of the CD-matrix boundary during excitation. (D) shows QRST deflections when the region of the CD is absent from the matrix.
excited and became refractory. Earlier portions of that QRS complex are, therefore, affected by the CD-matrix boundary, resulting in QRS alternans. After that QRS, the CD recovered whereas the surrounding matrix was still refractory so a boundary of opposite polarity to the previous one was present. This affected the amplitude of the second T wave and resulted in T-wave alternans. During both QRS and T deflection in the cycles with CD block, the CD-matrix boundary combined with the activation and recovery boundaries as described in previous examples.
Fig. 6. Calculated QRST deflections from a matrix with 2:1 propagation in a CD located distant from the origin of excitation. (A) shows QRST deflections with alternans of both QRS and T waves. (B) shows QRS complexes calculated from effects of activation only, and (C) shows deflections caused by other events. (D) shows QRST deflections when the region of the CD was inexcitable.
IV Block T-Wave Alternans •
Abildskov and Lux 317
activated until late in the first QRS but the matrix near the CD recovered early in the first T wave. The resulting boundary between recovered matrix and refractory CD produced a downward deflection that decreased the amplitude of the first T wave. The negative deflection persisted until near the end of the second QRS when the matrix near the CD was excited and became refractory so the amplitude of the complex preceding that point was also decreased. At the end of the second QRS, both matrix and CD were refractory and early in the second T wave the CD and surrounding matrix both recovered so T-wave amplitude was not affected by a CD-matrix boundary. Because the preceding T wave was affected by such a boundary, T-wave alternans was present. Duration of Refractoriness in the CD
Fig. 7. Calculated QRST deflections from matrices with 2:1 propagation in CDs and graded recovery properties resulting in QRS and T deflections of the same polarity. (A) shows effects of a CD near the origin of excitation with T-wave alternans. (B) shows effects of a CD equidistant from the ECG poles in which no alternans is present. (C) shows effects of a CD distant from the origin of excitation with alternans of both QRS and T.
origin of excitation. In the first cycle, the CD was activated near the onset of the QRS but a CD-matrix boundary caused by recovery adjacent to the CD was not present until near the end of the T wave. Neither QRS or T wave amplitude was affected by a CD-matrix boundary in this cycle. After the first T wave an upward deflection caused by a boundary between the refractory CD and recovered matrix was present and resulted in an upward deflection. That boundary disappeared early in the second QRS when the matrix near the CD became refractory so that QRS was not affected by a CD-matrix boundary. After the second QRS, the CD recovered forming a boundary with the refractory matrix. The resulting downward deflection affected the amplitude of the second T wave and produced T-wave alternans. Figure 7B shows a calculated ECG with a CD equidistant from the poles of the lead. In this condition, the CD had equal effects on the poles and did not result in T-wave alternans. Figure 7C shows a calculated ECG with a CD distant from the origin of excitation. In this condition, the CD was not
In the preceding examples, the RP in CDs was sufficiently long to result in 2:1 block but recovery occurred before the T wave of complexes with block. With those conditions, effects of the CDmatrix boundary were present in alternate complexes and T-wave alternans occurred. When the RP in a CD extended beyond the T wave of the complex with block, the CD-matrix boundary had the same effect or lack of effect in complexes with and without block and T-wave alternans was absent. An example is shown in Figure 8. Figure 8A shows T-wave alternans caused by a CD near the origin of excitation. The duration of refractoriness in the CD was such that a CD-matrix boundary affected T waves in the complexes with conduction but not those with block of the defect. Figure 8B shows decreased magnitude of T-wave alternans with increased duration of refractoriness in the CD and Figure 8C shows elimination of alternans with further increase of the duration of CD refractoriness.
Discussion This study provided evidence that 2:1 block in a localized ventricular region could result in QRS and/or T-wave alternans as the result of a boundary between the region of block and surrounding myocardium. It suggested that alternans caused by such block was not the result of different activation patterns in cycles with and without block as is usually assumed. The electrocardiographic effects of failure to activate a refractory region were largely eliminated by activation of the surrounding area
318 Journal of Electrocardiology Vol. 33 No. 4 October 2000
Fig. 8. Calculated QRST deflections from matrices with 2:1 propagation in a CD near the origin of excitation and with varied duration of CD refractoriness. (A) shows T-wave alternans when CD refractoriness did not extend to the T wave of complexes with block. (B) shows decreased magnitude of T-wave alternans with longer CD refractoriness. (C) shows no T-wave alternans with further prolongation of CD refractoriness.
because the refractory region contributed to the activation boundary. This finding has the important implication that 2:1 intraventricular block cannot be excluded as the mechanism of T-wave alternans because it is not associated with QRS alternans. T-wave features secondary to the activation sequence were not the reason for alternans because they were largely eliminated by the contribution of the CD region to T waveform. Rather than activation sequence, the mechanism of alternans was presence of the boundary between a CD and the surrounding region when these were in different states of polarization. The boundary resulted in ECG deflections that were superposed on QRST deflections caused by other events. The pattern of deflections produced by the boundary depended on the location of CDs, excitation and recovery sequences, and the particular ECG lead. Alternans
occurred when refractoriness in CDs was sufficiently long to result in block in alternate cycles but did not extend into the T wave in cycles with block. When the duration of refractoriness did extend into that T wave, the CD-matrix boundary affected that wave and T-wave alternans was absent or decreased in magnitude. This is of special interest because recent publications have reported decreased incidence or magnitude of T-wave alternans after administration of agents that prolong refractoriness. Kavesh et al. (19) reported deflected decreased magnitude after procainamide and Groh et al. (20) found that amiodarone decreased the prevalence of T-wave alternans. Findings in this study suggest those effects are compatible with 2:1 intraventricular block as the mechanism of the alternans. The main significance of findings in this study is that they suggest 2:1 intraventricular block may be a more frequent mechanism of T-wave alternans than previously supposed. The findings do not exclude other mechanisms but it should be noted that alternation of action potential duration, which has been considered evidence for a cellular mechanism, may be a manifestation of IV block. The presence of a localized refractory region in alternate cycles is likely to have electrotonic effects on action potentials in the surrounding area so alternating action potential duration may neither prove an intrinsic cellular mechanism or exclude IV block as the mechanism of T-wave alternans. The findings do not provide evidence for particular mechanisms in various clinical conditions. Finally, the mechanism of 2:1 intraventricular block is compatible with the established relation between T-wave alternans and ventricular arrhythmias (3–7,21–26). Unidirectional block and slow propagation in potential reentry paths are basic requirements for reentrant excitation, which is the probable mechanism of cardiac fibrillation and other ventricular tachyarrhythmias. Decreased conduction velocity and conduction block are wellknown effects of myocardial ischemia in which an association of T-wave alternans and susceptibility to ventricular tachycardia and fibrillation has been noted (4 –7). The computer model used in this study has also been used in studies of various arrhythmias and its limitations for that purpose are described in previous publications (8 –18). In the present study of ECG waveform, the limitations of the model mainly concern cardiac size, geometry, ECG lead characteristics, and the lack of electrotonic interactions during repolarization. The latter factor would be expected to reduce, but not eliminate, effects of
IV Block T-Wave Alternans •
boundaries associated with regions of block. The 2-dimensional matrix did not allow simulation of the spatial features of ventricular activation and recovery and ECG leads were calculated with the assumption of an infinite homogeneous conducting medium. To specify quantitative relations of the model to the heart it is necessary to assume either the spatial or temporal dimensions of the model. Calculated characteristics of the model based on such assumptions are a reasonable approximation of those in ventricular muscle (27). Despite its limitations, it is likely that the model correctly reflected the qualitative effects of CDs on ECG waveform. Only boundaries of potential difference that intersect epicardial or endocardial surfaces have ECG effects and boundaries on the matrix simulated such intersections. There is no apparent reason why CDs that intersect either of these surfaces do not result in boundaries of potential difference with ECG effects that include T-wave alternans.
11.
12.
13.
14.
15.
16.
17.
References 1. Hellerstein HK, Liebow IM: Electrical alternans in experimental coronary artery occlusion. Am J Physiol 160:366, 1950 2. Smith JM, Cohen RJ: Simple finite-element model accounts for wide range of ventricular arrhythmias. Proc Natl Acad Sci U S A 81:233, 1984 3. Verrier RL, Nearing BD: Electrophysiologic basis for T wave alternans as an index of vulnerability to ventricular fibrillation. J Cardiovasc Electrophysiol 5:445, 1994 4. Carron DL, Cardinal R, Savard P, et al: Characterisation of unipolar waveform alternation in acutely ischiaemic porcine myocardium. Circ Res 20:521, 1986 5. Surawicz B, Fisch C: Cardiac alternans: Diverse mechanisms and clinical manifestations. J Am Coll Cardiol 20:483, 1992 6. Sutton PMI, Taggart P, Lab M, et al: Alternation of epicardial repolarization as a localized phenomenon in man. Eur Heart J 12:70, 1991 7. Puletti M, Curione M, Rigbretta G, et al: Alternans of the ST segment and T wave in acute myocardial infarction. J Electrocardiol 13:297, 1980 8. Moe GK, Rheinboldt WC, Abildskov JA: A computer model of atrial fibrillation. Am Heart J 67:200, 1964 9. Abildskov JA, Steinhaus BM: Vulnerability to fibrillation in a mathematical model (Ueda Lecture). Jpn J Electrocardiol 10:3, 1990 10. Abildskov JA: Mechanism of the vulnerable period in
18.
19.
20.
21.
22.
23.
24.
25.
26. 27.
Abildskov and Lux 319
a model of cardiac fibrillation. J Cardiovasc Electrophysiol 1:303, 1990 Abildskov JA, Lux RL: Local events in a computer model of cardiac excitation. J Cardiovasc Electrophysiol 2:45, 1991 Abildskov JA, Lux RL: The mechanism of simulated torsade de pointes in a computer model of propagated excitation. J Cardiovasc Electrophysiol 2:224, 1991 Abildskov JA, Lux RL: Effects of nonuniform slowing of conduction on vulnerability to fibrillation in a computer model. J Cardiovasc Electrophysiol 3:48, 1992 Abildskov JA, Lux RL: Effects of premature excitation and tachycardia on the spatial distribution of refractoriness and propagation of excitation in a computer model. J Electrocardiol 25:287, 1992 Abildskov JA, Lux RL: Mechanisms in simulated torsade de pointes. J Cardiovasc Electrophysiol 4:547, 1993 Abildskov JA, Lux RL: Cycle length effects on the initiation of simulated torsade de pointes. J Electrocardiol 27:1, 1994 Abildskov JA, Lux RL: Spiral waves in a computer model of cardiac excitation. Pacing Clin Electrophysiol 17:1, 1994 Abildskov JA: Additions to the wavelet hypothesis of cardiac fibrillation. J Cardiovasc Electrophysiol 5:553, 1994 Kavesh NG, Shorofsky SA, Sarong SE, et al: The effect of procainamide on T wave alternans. J Cardiovasc Electrophysiol 10:649, 1999 Groh WJ, Shinn TS, Engelstein EE, et al: Amiodarone reduces the prevalence of T wave alternans in a population with ventricular tachyarrhythmias. J Cardiovasc Electrophysiol 10:1335, 1999 Smith JM, Clancy EA, Valeri CR, et al: Electrical alternans and cardiac electrical instability. Circulation 77:110, 1988 Rosenbaum DS, Jackson LE, Smith JM, et al: Electrical alternans and vulnerability to ventricular arrhythmias. N Engl J Med 330:235, 1994 Adam DR, Smith JM, Akselrod SM, et al: Fluctuations in T wave morphology and susceptibility to ventricular fibrillation. J Electrocardiol 17:209, 1984 Nearing B, Huang AH, Verrier RL: Dynamic tracing of cardiac vulnerability by complex demodulation of the T wave. Science 252:437, 1991 Schwartz PJ, Malliani A: Electrical alternation of the T wave: Clinical and experimental evidence of its relationship with the sympathetic nervous system and with long Q-T syndrome. Am Heart J 89:45, 1975 Wellens HJJ: Isolated alternans of the T wave. Chest 64:319, 1972 Abildskov JA, Lux RL: Simulated torsade de pointes—The role of conduction defects and mechanism of QRS rotation. J Electrocardiol 33:55, 2000
Mechanisms in T-Wave Alternans Caused By Intraventricular Block
J. A. Abildskov, MD, and Robert L. Lux, MD
Abstract: It is recognized that 2:1 intraventricular (IV) block can result in T-wave alternans but is usually assumed that it would also affect QRS waveform. Block in a local region is not, however, varied activation sequence of the same muscle mass because the blocked region is not activated and is not part of the mass that is activated in cycles without block. Also, the block region may have electrocardiogram (ECG) effects when its state differs from other regions. In view of those considerations, the ECG effects of IV block were evaluated by using a computer model of excitation and recovery. ECGs were calculated from differences between the excited state and various degrees of recovery. Results provided evidence that boundaries associated with regions of block rather than regions having varied activation sequence were the major factors in T-wave alternans caused by IV block. Effects of the boundaries included cancellation of the effects of IV block on QRS complexes. Findings suggest that IV block cannot be excluded as a mechanism of T-wave alternans in the absence of QRS alternans. Key words: Activation sequence, QRS alternans, computer model.
One of the possible mechanisms of T-wave alternans is 2:1 block of excitation in a portion of the ventricular muscle (1,2). It has usually been assumed that such block would affect QRS as well as T waveform and that the absence of QRS differences in alternate complexes was evidence that this mechanism was not responsible for T-wave alternans (3–7). If the mechanism was simply that of different activation sequence in alternate com-
plexes, QRS as well as T-wave alternans would be obligatory. However, conditions during 2:1 block of a localized region are not simply those of varied activation sequence. With such block, there is alternating presence and absence of excitation in the region of the block rather than varied activation sequence of the same muscle mass. Also, when a blocked region is not excited, it is still present in a refractory state and may have electrocardiogram (ECG) effects when it is adjacent to excitable muscle or when it becomes excitable and is adjacent to refractory muscle. These considerations suggest that there is not an obligatory relation between QRS and T-wave effects of localized intraventricular (IV) block and that mechanisms other than activation sequence variation may be responsible for the ECG effects of a conduction defect (CD) on ECG waveform. In this study, a computer model of propagated
From the Nora Eccles Harrison Cardiovascular Research and Training Institute, and the Division of Cardiology, University of Utah, Salt Lake City, UT. Supported by awards from the Nora Eccles Treadwell Foundation and the Richard A. and Nora Eccles Harrison Fund for Cardiovascular Research; and grant HL52338 from the National Institutes of Health (Heart, Lung and Blood Institute). Reprint requests: J. A. Abildskov, MD, University of Utah, CVRTI, 95 South 2000 East, Salt Lake City, UT 84112-5000. Copyright © 2000 by Churchill Livingstone® 0022-0736/00/3304-0002$10.00/0 doi: 10.1054/jelc.2000.18108
311
312 Journal of Electrocardiology Vol. 33 No. 4 October 2000 excitation was used to elucidate the mechanism of ECG effects caused by CDs. Findings suggest that boundaries of potential difference at the site of CDs are the major factor. They also suggest that T-wave alternans caused by CDs can occur with only slight variation of QRS waveform and the mechanism cannot be excluded by failure to detect QRS alternans.
Materials and Methods The computer model has been used in previous studies of reentrant arrhythmias and ECG waveform (8 –18). It consisted of a matrix of 25 ⫻ 25 individual units each having 6 neighbors to which excitation could be propagated. Time was represented as steps with a single time step (TS) being the time required to transfer excitation from an excited unit to a fully excitable neighbor. After excitation, units became refractory for a period defined by the relation K公CL, in which K was a property assigned to each unit, CL was the preceding cycle length in the unit, and the square root was an approximation of the CL–refractory period (RP) relation in cardiac muscle. After the RP, units became excitable but exhibited slow propagation represented by the requirement of 4, 3, and 2 TSs for transfer of excitation to units in successive stages of 4-TS duration each. ECGs were calculated by determining instantaneous source vectors for each matrix unit based on state differences. For the QRS complex, vectors from excited units in the direction of excitable neighbors were defined and summed to yield the source vector for each unit during the TS in which that unit was excited. Source vectors at all other TSs were defined from state differences and directed from less toward more fully recovered units. During recovery, 3 states, each of which represented one third of the difference between inexcitable and fully recovered units, were considered. Potentials at each of 2 sites located near the matrix were calculated by superposing potentials from all source vectors and assuming an infinite homogeneous medium. Potential differences at the 2 sites were taken to yield a bipolar ECG lead. Importantly, ECGs calculated only from units excited during the preceding TS and ECGs calculated from state differences during all other TSs were displayed both individually and in combination. This allowed differentiation of ECG effects caused exclusively by activation from those caused by activation in the presence of a CD during the time that it was refractory. ECGs were also calculated with the CD region absent from the
matrix. These showed QRS features when activation in the CD was blocked and T-wave features secondary to that activation. For the purposes of this study, the matrix was considered to represent either the epicardial or endocardial ventricular surface. Conduction defects in these layers resulted in unclosed boundaries of potential difference when they were in a different state than the surrounding matrix. These boundaries resulted in ECG deflections that occur as a consequence of activation or recovery boundaries that intersect cardiac surfaces (collision). Conduction defects at varied locations were created by K values that resulted in 2:1 propagation in the defects for the assigned rate of stimulation. In all cases, the initial stimulus was applied when the entire matrix, including the CD, was in a fully excitable state after an assigned prior CL. K values were selected such that regular stimuli at a CL shorter than the initial assigned one resulted in alternate block and conduction in the CD with propagation at full velocity in the remainder of the matrix. K values outside the CD were either uniform or arranged in layers such that the sequence of recovery was opposite that of excitation and QRS and T deflections had the same polarity. The matrix with stimulus site, ECG electrode sites, and CD locations are shown by the diagram in Figure 1. The CDs consisted of 21 units in a 3 ⫻ 5 array at the location shown in the diagram. The diagram shows CD location relative to stimulus and ECG electrode sites, but not the relative size of CDs and the total matrix. The ECG electrode sites were located at a distance equal to the dimensions of 6 units above the matrix.
Results The principal finding was evidence that a boundary of potential difference between a CD and the surrounding matrix resulted in T-wave alternans with only slight variation of QRS waveform. As with other sources of ECG deflections, effects of the boundary depended on its dimension, location, and polarity, as well as the ECG lead in which effects were observed. CD-Matrix Boundary CD-matrix boundaries occurred as a consequence of differences between states of the CD and surrounding matrix such as a refractory CD and recovered matrix. A boundary of opposite polarity ex-
IV Block T-Wave Alternans •
Fig. 1. Diagram of the matrix used in the study. Electrode sites from which an ECG lead was calculated are shown at X and Y. The site of stimulation is indicated by S. Conduction defects are shown at A near the origin of excitation, B midway between ECG electrode sites, and C distant from the origin of excitation.
Fig. 2. Conditions simulated in the study. The cardiac surface represented by the matrix is shown with CD indicated by the dotted lines. The poles of an ECG lead are shown as X and Y. When the state of the CD and surrounding matrix differed (ie, 1 excitable and the other refractory) the CD was a source of ECG deflection related to the relative magnitude of the angles shown.
Abildskov and Lux 313
isted when the CD was in a recovered state while the surrounding matrix was refractory, and the boundary disappeared when the CD and matrix were in the same state. As with other boundaries of potential difference that intersect a cardiac surface, the CD-matrix boundaries resulted in ECG deflections. The conditions simulated in this study are shown in Figure 2. The cardiac surface shown on edge is represented by the solid line shown with a CD represented by the dotted line at various locations. An ECG lead with its poles (X, Y) at a distance from the cardiac surface is shown. When the CD was refractory and the surrounding matrix excitable, or the CD excitable and matrix refractory, the CDs were the source of ECG deflections. In the bipolar lead shown, deflections were related to the relative magnitude of the angles at the poles. As shown in Figure 2A, the angle at X was larger than that at Y, resulting in an ECG deflection. Figure 2B shows a CD midway between the ECG poles so angles at the poles were equal and no deflection occurs. Figure 2C shows a CD located near the lower ECG pole (Y), resulting in unequal angles at X and Y, and an ECG deflection of opposite polarity to that resulting from conditions shown in Figure 2A. In addition to the CD location, the polarity of ECG deflections caused by a CD-matrix boundary depended on the relative state of the CD and surrounding matrix (ie, which was recovered or refractory) and on the polarity of the ECG lead. ECG deflections caused by a CD-matrix boundary are shown by the diagram in Figure 3 together with diagrammatic QRST complexes to show the timing
314 Journal of Electrocardiology Vol. 33 No. 4 October 2000
Fig. 3. Diagram of ECG deflections caused by a CD-matrix boundary with the CD located near the origin of excitation. The diagrammatic QRS and T waves show the time of occurrence of deflections caused by the CD-matrix boundary. The deflection at (A) was caused by a refractory CD and surrounding matrix that had recovered. At (B), the matrix adjacent to 1 edge of the CD was excited and became refractory, at (C) the other edge of the CD was excited and became refractory. At (D), the CD recovered whereas the matrix remained refractory, and at (E), the matrix at 1 edge of the CD recovered. At (F), the matrix at both edges of the CD have recovered and no CD-matrix boundary exists.
of deflections caused by the boundary. In this example, it has been assumed that recovery properties outside the CD were uniform so the sequence of repolarization was the same as that of depolarization and QRS and T deflections have opposite polarity. It has also been assumed that the CD was located near the site of origin of excitation. The first QRST complex represents one in which excitation occurred in the entire matrix and the second complex is one in which excitation in the CD was blocked. The deflection with its onset at (A) was caused by a boundary between the refractory CD and recovered matrix adjacent to the CD. The deflection at (B) occurred when the matrix adjacent to 1 edge of the CD was excited and became refractory, leaving the boundary at the other edge unopposed. At (C), excitation reached the other edge of the CD so both the CD and surrounding matrix were refractory and no CD-matrix boundary was present. At (D), the CD recovered whereas the matrix remained refractory resulting in a boundary of opposite polarity to the initial one. At (E), the matrix at 1 edge of the CD has recovered and the deflection is the effect of the remaining edge of the recovered CD-refractory matrix boundary. At (F), the matrix beyond the CD has recovered and no CD-matrix boundary exists. Deflections caused by the CD-matrix boundary were superposed on QRST complexes with alternate conduction and block in CDs. The detailed form of deflection caused by the boundaries and their time phase with QRST deflection depended on a variety of factors that will be shown in subsequent examples.
Figure 4 shows a computer simulation of the condition assumed for the diagram in Figure 3. The assigned prior CL was 200 TS and stimuli were applied at 100-TS intervals, which resulted in activation of the CD in alternate cycles. During the first QRS complex (Figure 4A), the entire matrix including the CD was activated. Because the CD was located near the stimulus site, it was activated during the early portion of the complex. Recovery properties outside the CD were uniform so the matrix surrounding the CD recovered early in the first T wave and the resulting CD-matrix boundary affected the remainder of the T wave. The boundary resulted in an upward deflection, which superposed on the T wave decreased the amplitude of the negative peak. The upward deflection alone is evident after the first T wave. During the second QRS complex, the CD continued to be refractory and propagation there was blocked. When excitation reached 1 edge of the CD, however, the boundary between refractory and recovered regions included the CD as well as the neighboring matrix. This was similar to the boundary when the CD was excited and the resulting QRS was similar to that of the first complex. After the second QRS complex, the CD recovered whereas the surrounding matrix was refractory, resulting in a CD-matrix boundary of opposite polarity to the initial one. That boundary was altered, then removed as the matrix surrounding the CD recovered so the second T wave was not affected by a CD-matrix boundary. In subsequent complexes there was alternating conduction and block in the CD and repetition of the events de-
IV Block T-Wave Alternans •
Abildskov and Lux 315
Figure 4B–D. Figure 4B shows QRS complexes calculated from effects of activation only without effects of the CD-matrix boundary. Alternating waveform caused by alternating conduction and block in the CD was present. Figure 4C shows deflections calculated from events exclusive of excitation. These include T waves and also effects of the CD-matrix boundary during activation in the complexes with block in the CD. Those deflections superposed on the QRS complexes shown in Figure 4B resulted in complexes similar to those without block. Figure 4D shows QRS and T deflections calculated from a matrix in which the region of the CD was absent. In this record, activation and recovery of the CD region and a CD-matrix boundary do not contribute to QRS or T waveform. CD Locations
Fig. 4. Calculated ECG deflections from a matrix with 2:1 propagation in a CD located near the origin of excitation. (A) shows QRST deflections with T-wave alternans and minor variation of QRS complexes. (B) shows QRS complexes calculated from effects of activation only with alternating QRS waveform caused by alternating failure to activate the CD. (C) shows deflections calculated from events other than excitation showing T-wave alternans and effects of the CD-matrix boundary during activation. (D) shows QRST deflections calculated from a matrix in which the region of the CD was absent.
scribed in the first 2 complexes. The CD-matrix boundary affected T-wave amplitude in complexes in which the CD was activated but not in complexes with block of the CD. It should be noted that if the CD had not recovered before the T wave of complexes with block, T-wave amplitude of those complexes as well as those with conduction in the CD would be affected by a CD-matrix boundary and T-wave alternans would not be present. The mechanisms by which a CD-matrix boundary affected ECG waveform are further shown by
The location of CDs affected the time of onset and duration of deflections caused by CD-matrix boundaries and the relation of the boundaries to ECG leads. When CDs were equidistant from the poles of a lead, the boundary had equal effects on the poles and did not produce deflections resulting in T-wave alternans. During the complexes with block of the CD, however, the CD-matrix boundary still contributed to activation and recovery boundaries outside the CD, so QRS and T deflections were only slightly affected by the block. These conditions are shown in Figure 5. Figure 5A shows QRST complexes without T-wave alternans and with similar QRS complexes with and without CD block. Figure 5B shows QRS complexes from activation alone showing effects of CD block in alternate cycles. Figure 5C shows deflections caused by events other than excitation including the contribution of the CD-matrix boundary to the QRS complex in cycles with block of the CD. Figure 5D shows deflections when the region of the CD was absent from the matrix and shows both QRS and T waveform without effects of the CD or CD-matrix boundary. Figure 6 shows effects of a CD distant from the origin of excitation. In that condition, the CD was not excited until late in the QRS in cycles without block of the CD. In the example shown, recovery properties outside the CD were uniform so the matrix surrounding the CD did not recover until late in the first T wave. The amplitude of that wave was, therefore, not affected by a CD-matrix boundary. After the onset of the boundary late in the first T wave, it continued to exist until late in the second QRS when the matrix surrounding the CD was
316 Journal of Electrocardiology Vol. 33 No. 4 October 2000 Graded Recovery Properties The distribution of recovery properties affected the time phase of deflections caused by a CD-matrix boundary and those caused by other events. Examples of the effect of CDs at various locations in a matrix with graded recovery properties are shown in Figure 7. In these examples, 5 layers of K values ranging from 2 to 6 with the highest value near the stimulus site were assigned. With those conditions, the sequence of recovery was opposite that of excitation and QRS and T deflections had the same polarity as they had with normal ventricular excitation. Figure 7A shows effects of a CD near the
Fig. 5. Calculated ECG deflections from a matrix with 2:1 propagation in a CD located midway between the poles of the ECG lead. (A) shows QRST deflections with no alternans of either QRS or T. (B) shows QRS complexes calculated from events during activation only and showing effects of CD block in alternate cycles. (C) shows deflections caused by events exclusive of excitation including effects of the CD-matrix boundary during excitation. (D) shows QRST deflections when the region of the CD is absent from the matrix.
excited and became refractory. Earlier portions of that QRS complex are, therefore, affected by the CD-matrix boundary, resulting in QRS alternans. After that QRS, the CD recovered whereas the surrounding matrix was still refractory so a boundary of opposite polarity to the previous one was present. This affected the amplitude of the second T wave and resulted in T-wave alternans. During both QRS and T deflection in the cycles with CD block, the CD-matrix boundary combined with the activation and recovery boundaries as described in previous examples.
Fig. 6. Calculated QRST deflections from a matrix with 2:1 propagation in a CD located distant from the origin of excitation. (A) shows QRST deflections with alternans of both QRS and T waves. (B) shows QRS complexes calculated from effects of activation only, and (C) shows deflections caused by other events. (D) shows QRST deflections when the region of the CD was inexcitable.
IV Block T-Wave Alternans •
Abildskov and Lux 317
activated until late in the first QRS but the matrix near the CD recovered early in the first T wave. The resulting boundary between recovered matrix and refractory CD produced a downward deflection that decreased the amplitude of the first T wave. The negative deflection persisted until near the end of the second QRS when the matrix near the CD was excited and became refractory so the amplitude of the complex preceding that point was also decreased. At the end of the second QRS, both matrix and CD were refractory and early in the second T wave the CD and surrounding matrix both recovered so T-wave amplitude was not affected by a CD-matrix boundary. Because the preceding T wave was affected by such a boundary, T-wave alternans was present. Duration of Refractoriness in the CD
Fig. 7. Calculated QRST deflections from matrices with 2:1 propagation in CDs and graded recovery properties resulting in QRS and T deflections of the same polarity. (A) shows effects of a CD near the origin of excitation with T-wave alternans. (B) shows effects of a CD equidistant from the ECG poles in which no alternans is present. (C) shows effects of a CD distant from the origin of excitation with alternans of both QRS and T.
origin of excitation. In the first cycle, the CD was activated near the onset of the QRS but a CD-matrix boundary caused by recovery adjacent to the CD was not present until near the end of the T wave. Neither QRS or T wave amplitude was affected by a CD-matrix boundary in this cycle. After the first T wave an upward deflection caused by a boundary between the refractory CD and recovered matrix was present and resulted in an upward deflection. That boundary disappeared early in the second QRS when the matrix near the CD became refractory so that QRS was not affected by a CD-matrix boundary. After the second QRS, the CD recovered forming a boundary with the refractory matrix. The resulting downward deflection affected the amplitude of the second T wave and produced T-wave alternans. Figure 7B shows a calculated ECG with a CD equidistant from the poles of the lead. In this condition, the CD had equal effects on the poles and did not result in T-wave alternans. Figure 7C shows a calculated ECG with a CD distant from the origin of excitation. In this condition, the CD was not
In the preceding examples, the RP in CDs was sufficiently long to result in 2:1 block but recovery occurred before the T wave of complexes with block. With those conditions, effects of the CDmatrix boundary were present in alternate complexes and T-wave alternans occurred. When the RP in a CD extended beyond the T wave of the complex with block, the CD-matrix boundary had the same effect or lack of effect in complexes with and without block and T-wave alternans was absent. An example is shown in Figure 8. Figure 8A shows T-wave alternans caused by a CD near the origin of excitation. The duration of refractoriness in the CD was such that a CD-matrix boundary affected T waves in the complexes with conduction but not those with block of the defect. Figure 8B shows decreased magnitude of T-wave alternans with increased duration of refractoriness in the CD and Figure 8C shows elimination of alternans with further increase of the duration of CD refractoriness.
Discussion This study provided evidence that 2:1 block in a localized ventricular region could result in QRS and/or T-wave alternans as the result of a boundary between the region of block and surrounding myocardium. It suggested that alternans caused by such block was not the result of different activation patterns in cycles with and without block as is usually assumed. The electrocardiographic effects of failure to activate a refractory region were largely eliminated by activation of the surrounding area
318 Journal of Electrocardiology Vol. 33 No. 4 October 2000
Fig. 8. Calculated QRST deflections from matrices with 2:1 propagation in a CD near the origin of excitation and with varied duration of CD refractoriness. (A) shows T-wave alternans when CD refractoriness did not extend to the T wave of complexes with block. (B) shows decreased magnitude of T-wave alternans with longer CD refractoriness. (C) shows no T-wave alternans with further prolongation of CD refractoriness.
because the refractory region contributed to the activation boundary. This finding has the important implication that 2:1 intraventricular block cannot be excluded as the mechanism of T-wave alternans because it is not associated with QRS alternans. T-wave features secondary to the activation sequence were not the reason for alternans because they were largely eliminated by the contribution of the CD region to T waveform. Rather than activation sequence, the mechanism of alternans was presence of the boundary between a CD and the surrounding region when these were in different states of polarization. The boundary resulted in ECG deflections that were superposed on QRST deflections caused by other events. The pattern of deflections produced by the boundary depended on the location of CDs, excitation and recovery sequences, and the particular ECG lead. Alternans
occurred when refractoriness in CDs was sufficiently long to result in block in alternate cycles but did not extend into the T wave in cycles with block. When the duration of refractoriness did extend into that T wave, the CD-matrix boundary affected that wave and T-wave alternans was absent or decreased in magnitude. This is of special interest because recent publications have reported decreased incidence or magnitude of T-wave alternans after administration of agents that prolong refractoriness. Kavesh et al. (19) reported deflected decreased magnitude after procainamide and Groh et al. (20) found that amiodarone decreased the prevalence of T-wave alternans. Findings in this study suggest those effects are compatible with 2:1 intraventricular block as the mechanism of the alternans. The main significance of findings in this study is that they suggest 2:1 intraventricular block may be a more frequent mechanism of T-wave alternans than previously supposed. The findings do not exclude other mechanisms but it should be noted that alternation of action potential duration, which has been considered evidence for a cellular mechanism, may be a manifestation of IV block. The presence of a localized refractory region in alternate cycles is likely to have electrotonic effects on action potentials in the surrounding area so alternating action potential duration may neither prove an intrinsic cellular mechanism or exclude IV block as the mechanism of T-wave alternans. The findings do not provide evidence for particular mechanisms in various clinical conditions. Finally, the mechanism of 2:1 intraventricular block is compatible with the established relation between T-wave alternans and ventricular arrhythmias (3–7,21–26). Unidirectional block and slow propagation in potential reentry paths are basic requirements for reentrant excitation, which is the probable mechanism of cardiac fibrillation and other ventricular tachyarrhythmias. Decreased conduction velocity and conduction block are wellknown effects of myocardial ischemia in which an association of T-wave alternans and susceptibility to ventricular tachycardia and fibrillation has been noted (4 –7). The computer model used in this study has also been used in studies of various arrhythmias and its limitations for that purpose are described in previous publications (8 –18). In the present study of ECG waveform, the limitations of the model mainly concern cardiac size, geometry, ECG lead characteristics, and the lack of electrotonic interactions during repolarization. The latter factor would be expected to reduce, but not eliminate, effects of
IV Block T-Wave Alternans •
boundaries associated with regions of block. The 2-dimensional matrix did not allow simulation of the spatial features of ventricular activation and recovery and ECG leads were calculated with the assumption of an infinite homogeneous conducting medium. To specify quantitative relations of the model to the heart it is necessary to assume either the spatial or temporal dimensions of the model. Calculated characteristics of the model based on such assumptions are a reasonable approximation of those in ventricular muscle (27). Despite its limitations, it is likely that the model correctly reflected the qualitative effects of CDs on ECG waveform. Only boundaries of potential difference that intersect epicardial or endocardial surfaces have ECG effects and boundaries on the matrix simulated such intersections. There is no apparent reason why CDs that intersect either of these surfaces do not result in boundaries of potential difference with ECG effects that include T-wave alternans.
11.
12.
13.
14.
15.
16.
17.
References 1. Hellerstein HK, Liebow IM: Electrical alternans in experimental coronary artery occlusion. Am J Physiol 160:366, 1950 2. Smith JM, Cohen RJ: Simple finite-element model accounts for wide range of ventricular arrhythmias. Proc Natl Acad Sci U S A 81:233, 1984 3. Verrier RL, Nearing BD: Electrophysiologic basis for T wave alternans as an index of vulnerability to ventricular fibrillation. J Cardiovasc Electrophysiol 5:445, 1994 4. Carron DL, Cardinal R, Savard P, et al: Characterisation of unipolar waveform alternation in acutely ischiaemic porcine myocardium. Circ Res 20:521, 1986 5. Surawicz B, Fisch C: Cardiac alternans: Diverse mechanisms and clinical manifestations. J Am Coll Cardiol 20:483, 1992 6. Sutton PMI, Taggart P, Lab M, et al: Alternation of epicardial repolarization as a localized phenomenon in man. Eur Heart J 12:70, 1991 7. Puletti M, Curione M, Rigbretta G, et al: Alternans of the ST segment and T wave in acute myocardial infarction. J Electrocardiol 13:297, 1980 8. Moe GK, Rheinboldt WC, Abildskov JA: A computer model of atrial fibrillation. Am Heart J 67:200, 1964 9. Abildskov JA, Steinhaus BM: Vulnerability to fibrillation in a mathematical model (Ueda Lecture). Jpn J Electrocardiol 10:3, 1990 10. Abildskov JA: Mechanism of the vulnerable period in
18.
19.
20.
21.
22.
23.
24.
25.
26. 27.
Abildskov and Lux 319
a model of cardiac fibrillation. J Cardiovasc Electrophysiol 1:303, 1990 Abildskov JA, Lux RL: Local events in a computer model of cardiac excitation. J Cardiovasc Electrophysiol 2:45, 1991 Abildskov JA, Lux RL: The mechanism of simulated torsade de pointes in a computer model of propagated excitation. J Cardiovasc Electrophysiol 2:224, 1991 Abildskov JA, Lux RL: Effects of nonuniform slowing of conduction on vulnerability to fibrillation in a computer model. J Cardiovasc Electrophysiol 3:48, 1992 Abildskov JA, Lux RL: Effects of premature excitation and tachycardia on the spatial distribution of refractoriness and propagation of excitation in a computer model. J Electrocardiol 25:287, 1992 Abildskov JA, Lux RL: Mechanisms in simulated torsade de pointes. J Cardiovasc Electrophysiol 4:547, 1993 Abildskov JA, Lux RL: Cycle length effects on the initiation of simulated torsade de pointes. J Electrocardiol 27:1, 1994 Abildskov JA, Lux RL: Spiral waves in a computer model of cardiac excitation. Pacing Clin Electrophysiol 17:1, 1994 Abildskov JA: Additions to the wavelet hypothesis of cardiac fibrillation. J Cardiovasc Electrophysiol 5:553, 1994 Kavesh NG, Shorofsky SA, Sarong SE, et al: The effect of procainamide on T wave alternans. J Cardiovasc Electrophysiol 10:649, 1999 Groh WJ, Shinn TS, Engelstein EE, et al: Amiodarone reduces the prevalence of T wave alternans in a population with ventricular tachyarrhythmias. J Cardiovasc Electrophysiol 10:1335, 1999 Smith JM, Clancy EA, Valeri CR, et al: Electrical alternans and cardiac electrical instability. Circulation 77:110, 1988 Rosenbaum DS, Jackson LE, Smith JM, et al: Electrical alternans and vulnerability to ventricular arrhythmias. N Engl J Med 330:235, 1994 Adam DR, Smith JM, Akselrod SM, et al: Fluctuations in T wave morphology and susceptibility to ventricular fibrillation. J Electrocardiol 17:209, 1984 Nearing B, Huang AH, Verrier RL: Dynamic tracing of cardiac vulnerability by complex demodulation of the T wave. Science 252:437, 1991 Schwartz PJ, Malliani A: Electrical alternation of the T wave: Clinical and experimental evidence of its relationship with the sympathetic nervous system and with long Q-T syndrome. Am Heart J 89:45, 1975 Wellens HJJ: Isolated alternans of the T wave. Chest 64:319, 1972 Abildskov JA, Lux RL: Simulated torsade de pointes—The role of conduction defects and mechanism of QRS rotation. J Electrocardiol 33:55, 2000