Three Distinct Patterns of Ventricular Activation in Infarcted Human Hearts
An Intraoperative Cardiac Mapping Study During Sinus Rhythm
Background Comprehensive data based on single-beat analysis of the ventricular activation sequence during sinus rhythm in infarcted hearts are currently not available. It was the aim of our study (1) to measure and analyze these activation sequences on the epicardial surface of the right and left ventricles and on the left ventricular endocardial surface, and (2) to correlate specific activation patterns with the surface ECG.
Methods and Results Isochronal maps were computed from 127 endocardial and epicardial unipolar electrograms recorded simultaneously during sinus rhythm in 45 post–myocardial infarction patients operated on for recurrent ventricular tachycardia (age, 57±10 years [mean±SD], left ventricular ejection fraction, 29±9%). Patients with bundle-branch block, but not with intraventricular conduction defects, were excluded. Data such as the timing of initial and terminal activation, the number of breakthroughs, the total activation time, and the number of ventricular segments without activation were measured and analyzed according to location of the myocardial infarction. The global epicardial activation was characterized in all patients by a widespread initial breakthrough on the anterior right ventricle (16±8 milliseconds after QRS onset), which was followed by one or two other breakthroughs in 65% of patients. Subsequently, three characteristic epicardial patterns of the activation spread were found: (1) radial, from the right to the left ventricle, found in all patients with inferoposterior myocardial infarction; (2) counterclockwise rotation, in which posteroseptal crossing preceded the anteroseptal crossing, found in 38% of patients with anterior myocardial infarction; and (3) pincerlike encirclement, in which both septal crossings and/or breakthroughs occurred nearly simultaneously and merged at the left ventricular free wall (typical for apical involvement in anterior and combined myocardial infarction). The simultaneous presence of multiple major activation wave fronts typically found in patients with the pincerlike activation pattern was reflected on the surface ECG by multiphasic, notched QRS complexes. Activation delay was observed in 89% of patients, and terminal activation was topographically related to myocardial infarction in 94% of patients. Delayed activation exceeding the surface QRS was observed in 11% and 31% of cases on the endocardium and epicardium, respectively.
Conclusions These results offer a solid basis for a more precise interpretation of a wide range of electrophysiological data and provide a framework for future investigations of surface ECG reflections of endocardial and epicardial activation patterns recorded in patients with chronic myocardial infarction.
For more than 60 years, the ECG has been the basic clinical tool for the diagnosis of acute and chronic myocardial infarction (MI). However, our knowledge about the exact character of the sequence of ventricular activation in the presence of focal myocardial lesions remains relatively limited. The classic studies by Wilson et al1 2 in the mid-1930s, as well as the study by Durrer et al3 30 years later, focused mainly on the formal characteristics of epicardial and intramural electrograms in the regions overlying or adjacent to an MI scar. These studies constitute the basis of the electric window theory in electrocardiography. Daniel et al4 5 analyzed epicardial potential maps obtained during cardiac revascularization surgery, and these data shed more light on the time sequence of ventricular epicardial depolarization in infarcted human hearts. Our present understanding on this subject is closely linked to a more widespread use of endocardial catheter mapping techniques6 7 8 9 10 11 12 and perioperative direct endocardial and epicardial mapping.13 14 15 16 17 18 However, because of technical limitations at the time of these studies—which were done more than two decades ago—these data are related mainly to selected partial characteristics of the abnormal ventricular depolarization process (eg, location of first or last activity, presence of fragmentation). Furthermore, they do not provide simultaneous single-beat measurements, their spatial resolution is often limited, and the electrode positioning is variable because of the use of handheld probes.
The pioneer electrophysiological study by Durrer et al19 on reperfused normal hearts from accident victims substantially increased our understanding of normal ventricular depolarization. However, we still lack precise knowledge about the sequence of ventricular activation in infarcted hearts. Therefore, it was the aim of our study to trace the time sequence of ventricular depolarization both on the epicardial surface of the total heart and on the left ventricular endocardial surface in patients with chronic MI. Electrical mapping data from patients undergoing antiarrhythmic surgery for postinfarction recurrent ventricular tachycardia (VT) were used for this purpose. This study therefore focuses on the relation of infarction scar location on the epicardial and left ventricular endocardial depolarization sequence during sinus rhythm.
Perioperative mapping data obtained during spontaneous supraventricular heart rhythm from 76 consecutive patients who underwent antiarrhythmic surgery for refractory sustained postinfarction VT were retrospectively reviewed for this study. Data from patients with bundle-branch block, either already present preoperatively or emerging during the intraoperative mapping procedure, were excluded from further analysis. However, data from patients with fascicular blocks and nonspecific intraventricular conduction disturbances were included. The clinical characteristics of the 45 patients whose data were included in the study, together with the results of their preoperative evaluations, are summarized in Table 1⇓. There were 40 men and 5 women in the studied group, and their mean age (±SD) was 57±10 years. All had clinically documented VT that was reproducibly inducible during a preoperative electrophysiological study. At the time of the cardiac surgery, the mean time since MI was 62±54 months. Four patients had clinically documented multiple MIs.
Topographic Analysis of the Data Related to the Ventricular Activation
To enable a comparative topographical analysis of the various data related to the sequence of ventricular activation to be made, the left ventricular (LV) endocardial and total (ie, right ventricular [RV] and LV) epicardial surfaces of the heart were divided into 14 and 13 segments, respectively (Fig 1⇓). This schematic segmentation of the polar projection of the heart, which bears a resemblance to the bull’s-eye representation of the heart used in nuclear cardiac imaging techniques, is compatible with the basic pattern of the coronary ramification to the ventricular myocardium.
The location of the MI was determined visually on the epicardial surface and the LV endocardial surface by the surgeon who operated on all 76 patients (P.P.). On the epicardium, the borderlines of the infarction scar were determined while the epicardial sock electrode was in place, and the positions of electrodes were used as reference points. The patients were divided into three groups: anterior MI (AMI) (n=21), inferoposterior MI (IMI) (n=16), and combined MI (CMI) (ie, with both anterior and inferoposterior lesions) (n=8). Visual evaluation was also used for determining the presence of an aneurysm, based on the presence of two or more of the following criteria (modified according to Bolick et al20 ): (1) bulging of the expected epicardial contour of the concerned LV segment during systole, (2) LV wall thinning, or (3) the transmural extent of the lesion as determined macroscopically during surgical resection of the ventricular wall. The presence of an intramural thrombus was determined at the same time. LV aneurysm was present in 30 of 45 patients (67%), and thrombus adherent to the endocardial scar was found in 20% of aneurysmal hearts.
Multielectrode mapping of ventricular activation was performed during normothermic cardiopulmonary bypass while the hearts were in stable spontaneous rhythm. Sinus rhythm was present in 43 patients; atrial fibrillation was recorded in 1 of the 2 remaining patients and stable junctional escape rhythm with narrow QRS complexes was recorded in the other. In accordance with the previously described intraoperative mapping technique used at our institution,21 63 unipolar electrograms were recorded from the epicardial surface of the LV and RV with an elastic sock electrode array. The 64th recording channel was reserved for simultaneous limb lead recording. The positions of the epicardial course of the coronary arteries (left anterior descending artery and posterior descending artery) were noted, as was that of the acute margin of the heart. These reference points were transferred to the individual display of maps. Surface leads I, II, III, and V6 were recorded simultaneously on a chart recorder (Electronics for Medicine, VR16). Endocardial recordings were performed by means of an inflatable balloon with 64 electrodes, introduced into the intact LV by way of the mitral orifice through a left atrial incision. The endocardial recordings were not obtained in the patients operated on in the first stage of our series, and in some other patients they had to be rejected because of unacceptable quality. Thus, for 28 of the 45 patients, both epicardial and LV endocardial mapping data were available for analysis. In the majority of patients (22) with endocardial data, the unipolar recording technique was used, and in the 6 remaining patients, bipolar recordings were performed.
The mapping system is based on a MicroVAX II computer (Digital Equipment Corp) and runs custom-made integrated software (Institut de Génie Biomédical, Ecole Polytechnique).22 The system permits 127 epicardial and endocardial signals to be acquired simultaneously with one monitoring surface lead. The data acquisition module stores data from the last 24.5 seconds of the recording in a buffer memory and continuously updates them. This enables the operator to select a typical QRS complex during stable sinus rhythm and to further process the signal window comprising the chosen beat. After amplification with optimized gains and 0.05- to 200-Hz filtering, the signals are sampled at 500 Hz and converted to digital format.
The procedure of activation mapping comprises two principal steps: (1) identification of the local activation time at each electrode, and (2) generation of the isochronal map by means of an interpolation algorithm.
For the unipolar recordings, the local activation time was defined as the point at which the most rapid negative deflection (steepest negative slope) occurs within the intrinsic deflection. The software can automatically detect the local activation time in each electrogram on the basis of the calculation of the three-point Lagrange derivative; this is an accepted method comparable to other derivative calculations.23 24 The measured steepest slope negativity had to exceed −0.3 mV/ms to be accepted by the computer as local activation. This approach was previously used in our laboratory for sinus rhythm mapping in canine infarcted hearts.25 All signals were subsequently reviewed by one of us (R.H.) for correct positioning of the activation markers. The following additional criteria for the detection of the presence of local activation in unipolar recordings were applied:
1. Continuity of propagation in the neighboring leads.
2. Spatial filtering. Signal processing based on the Laplace operator was applied to assess the similitude of an electrogram to its neighbors. This calculation was done by subtracting the mean potential measured at the neighboring sites (weighted by their respective distance) from the analyzed electrogram. This approach turned out to be especially useful for determining whether an endocardial signal with QS morphology is due to breakthrough or MI. In the latter case, the spatial filter displays a flat signal without true deflections as a reflection of the fact that all neighboring electrodes recorded a similar electric hole. This flat signal indicates that the potential distribution does not contain any high spatial frequency components and that there is no spatial aliasing. Because of the possibility of spatial aliasing, spatial filtering was not considered when the output of the filter was not flat.
3. If a true QS configuration (ie, without even a minimal R wave) was present on the epicardium, it was considered to be an electrically dead region behaving as a window; therefore, no activation marker was set. This approach is consistent with recommendations of Laxer et al26 and was previously followed in our laboratory in experimental studies in sinus rhythm mapping in canine hearts.25 If QS configuration was encountered on the endocardium, the presence of local activation was judged according to the spatial derivation method (spatial filtering) described above.
The timing of the local activation was related to the zero time, which was defined as the onset of the QRS complex on the root mean square signal from the 64 endocardial leads. This was set visually by means of a cursor on an interactive display screen.
Local activation in bipolar recordings was determined by means of the maximal absolute amplitude of the electrogram. When a fragmented signal was present (duration >50 milliseconds) and the maximal absolute amplitude was <1 mV, no activation mark was set.
Display of Ventricular Depolarization Sequence
Ventricular depolarization was displayed in the form of isochronal maps in a polar projection of the LV endocardial surface and the LV and RV epicardial surface. These maps provide a circular representation of the heart as viewed from the apex. The apex is in the center of the circle and the ventriculoatrial junction is at its perimeter (Fig 1⇑).
Isochronal maps were computed using either linear or cubic B-spline interpolation algorithms and displayed as line maps or color maps on a high-resolution graphic monitor (CPD 1302, Sony Inc). Line maps were printed on a thermal hard-copy printer, and color maps were photographed directly from the screen. Fig 2⇓ is an example of an epicardial isochronal map recorded during sinus rhythm in a patient with anterior MI, along with some of the unipolar epicardial electrograms.
Definitions of Parameters and Related Terminology
The following parameters were defined and analyzed to characterize the ventricular depolarization:
Total activation time was the time between the start of ventricular depolarization (ie, onset of the endocardial QRS complex on the root mean square signal) and the last detected local activation. It was determined separately for the entire epicardium and for the LV endocardium. Thus, the total activation time corresponds to the time interval from the zero reference time to the latest observed local activation on either epicardium or LV endocardium. The simultaneous duration of the surface QRS complex was determined as the time interval from the zero reference time to the QRS offset in the limb lead monitored simultaneously with the mapping. The latest detected epicardial or endocardial activation was related to this time interval to detect post-QRS activity. However, the QRS complex duration on the body surface was measured from the preoperative 12-lead ECG. The data for these parameters are provided in Table 2⇓.
A breakthrough was defined as an epicardial or endocardial site where an activation wave front emerged and from where it spread radially. Thus, a breakthrough formed an island of activation surrounded by sites activated later. The number of breakthroughs observed on the epicardium and LV endocardium, their respective timing, and their location were analyzed. The first breakthrough appearing was designated the primary breakthrough.
Initial LV epicardial activity was defined as the timing and location of the initial activation at the LV epicardial surface. This activation could occur as either the emergence of a breakthrough or a tangential epicardial activation wave spread.
Terminal activation was the last local activation detected on the epicardium and LV endocardium. The following characteristics were studied: (1) timing with respect to the surface QRS complex offset (separately for the epicardium and LV endocardium) and (2) location of the terminal activity (within the last 10 milliseconds of the depolarization) and its relation to the MI scar (ie, whether related or not). One or several sites (neighboring or remote) of terminal activation might have been present in the same patient.
When electrograms without local activation were recorded, the epicardial and endocardial segments were designated as “dead.”
Local conduction delay was defined as slowing of conduction between any two neighboring electrodes so that the difference of their respective activation times was ≥30 milliseconds.
Anterior septal crossing and posterior septal crossing were defined as epicardial phenomena occurring when the activation wave front, spreading tangentially, crossed the anterior or posterior septum, respectively.
The definition recommended by the WHO/ISFC task force for left anterior fascicular block (LAFB) was adopted. The following criteria had to be fulfilled: (1) left axis deviation with a frontal plane QRS axis of −45° to −90°, (2) a qR pattern in aVL, (3) an R peak time in lead aVL ≥45 milliseconds, and (4) QRS duration <120 milliseconds.27
The definition recommended by the WHO/ISFC task force for left posterior fascicular block was adopted. The following criteria had to be fulfilled: (1) right axis deviation with a frontal QRS axis of +90° to +180°, (2) S1Q3 pattern of the QRS complex in the limb leads, and (3) QRS duration <120 milliseconds.27
Intraventricular block (nonspecific) was defined as a QRS complex of >120 milliseconds not meeting the criteria for either a left or a right bundle-branch block pattern.27
Surface ECG Analysis
Two parameters were analyzed in the 12-lead ECGs recorded within the 2 days before the heart surgery: (1) presence of diagnostic criteria for Q wave MI28 and (2) presence of multiphasic QRS complexes comprising abnormal notching of the waveform. Notches were defined as any departure in both slope and sign from the primary ventricular ECG curve exclusive of the fundamental directional changes of the QRS complex (ie, the nadirs of Q, R, and S deflection were not included).29
Data are expressed as mean±SD and were statistically analyzed by ANOVA. Differences at a level of P≤.05 were considered statistically significant.
Topography of MI Location
The MI location and extent were determined both macroscopically during the surgical procedure and electrically by means of intraoperative mapping. A topographic assessment was done for both the total epicardial surface of the heart and for the LV endocardial surface. These data are summarized in Table 1⇑. In further analysis, the intraoperative mapping data were examined. We note that the macroscopic evaluation revealed the involvement of the posterior apical segment, in addition to the anteroapical involvement, in 13 of 21 cases of AMI. There was septal involvement, according to the surgeon’s visual evaluation, in 19 of 20 patients with AMI who underwent ventriculotomy, in 10 of 16 patients with IMI, and in 6 of 8 patients with CMI.
The distribution of the infarcted segments in the three MI groups is shown in Fig 3⇓. The total number of segments without local activation was significantly different among the three groups (Table 2⇑). The RV was involved in the electrically dead zone in 1 patient with an inferoposterior scar. It is of interest that there was no absence of local activation on the epicardium in 2 of 21 cases of AMI, 1 of 16 cases of IMI, and 1 of 8 cases of CMI. In all patients, there was an electrically dead zone on the LV endocardium. However, an abnormal activation delay over the epicardial regions with a macroscopic scar was observed in all patients in whom no absence of local activation could be detected. In all these patients, a clearly heterogenous scar was already visually evident and was confirmed by histological examination in cases in which biopsy specimens were obtained.
Quantitative and Topographic Characteristics of Ventricular Depolarization
The quantitative parameters concerning the timing of the different activation events are shown for the three MI groups in Table 2⇑. The depolarization of the LV endocardium started from 1 or more breakthroughs. Its timing with respect to the onset of the QRS on the endocardial root mean square signal was 5±6 milliseconds in the AMI group, 2±3 milliseconds in the IMI group, and 5±7 milliseconds in the CMI group. The total number of endocardial breakthroughs varied between the MI groups: 1.9±0.6 in the AMI group, 1.3±0.4 in the IMI group, and 1.7±0.5 in the CMI group. None of the patients had more than 3 LV endocardial breakthroughs. Two or 3 LV endocardial breakthroughs were present in 10 of 13 patients with AMI, in 2 of 8 with IMI, and in 5 of 7 with CMI. All initial LV endocardial breakthroughs were located on either the septum or the anterior wall (Fig 4⇓). The LV endocardial activation terminated before the end of the surface QRS complex: 28±18 milliseconds in patients with AMI, 30±18 milliseconds in those with IMI, and 20±13 milliseconds in those with CMI. Local endocardial activation after the surface QRS offset (post-QRS activity) was observed in 3 of 28 patients (1 in each group) for whom corresponding LV endocardial data were available for computer analysis. The latest post-QRS LV endocardial activity exceeded the surface QRS by 10 milliseconds.
The primary epicardial breakthrough appeared at 16±7 milliseconds in the AMI group, 19±9 milliseconds in the IMI group, and 15±8 milliseconds in the CMI group. The distribution of the breakthrough sites within the groups is represented schematically in Fig 4⇑. The total number of epicardial breakthroughs was similar for all MI locations: 1.8±0.7 for AMI, 1.6±0.6 for IMI, and 2.0±0.5 for CMI. None of the patients had more than 3 epicardial breakthroughs. Two or more epicardial breakthroughs were observed in 14 of 21 AMI patients, 8 of 16 IMI patients, and 7 of 8 CMI patients. The most constant epicardial breakthrough feature was the site of the primary breakthrough, which was in all cases on the anterior RV epicardium, mainly in segments 12 and 13. These segments correspond to the anterior aspect of the RV between the septum and the acute margin. In 1 patient the breakthrough was relatively large and reached up to the posterior RV (segment 11).
Over the LV, initial LV epicardial activation started 27±14 milliseconds after depolarization onset in patients with AMI, 28±10 milliseconds in patients with IMI, and 29±13 milliseconds in patients with CMI. The location of the initially depolarized epicardial regions was confined mainly to anterior or posterior paraseptal segments in patients with AMI or CMI and to apical and anterior paraseptal segments in those with IMI.
Activation of the epicardium generally terminated later than that of the LV endocardium. It preceded the end of the surface QRS complex by 9±16 milliseconds in patients with AMI and by 6±23 milliseconds in patients with CMI. However, it exceeded the surface QRS complex by 3±21 milliseconds in patients with IMI. Post-QRS activity on the epicardium was observed in 14 of 45 patients, with a maximal delay of 60 milliseconds. Locations of the terminal activity in the studied population are presented in Fig 5⇓. It was almost invariably located in areas within the MI lesion, adjacent to the lesion, or both, although in 4 cases it was located on the epicardium. On the endocardium, it was mainly located on the LV free wall (segments 5 and 6) (16 of 28 patients). On the epicardium, it was confined predominantly to the LV; in 8 subjects it was also observed over the RV (segments 10 through 13).
Overall Pattern of the Sequence of Ventricular Depolarization
The above-mentioned quantitative and topographical parameters determine the global character of ventricular depolarization, ie, the macroscopic pattern of the activation wave progression on the epicardial and LV endocardial surfaces of the heart. To recognize patterns typical for different MI locations, we studied the overall depolarization sequence as detected on the epicardium and LV endocardium in the three MI groups. Not surprisingly, the location of the dead segments acting either as barriers or as areas of important conduction delay of the ventricular activation spread turned out to be the major determinant of the global depolarization pattern.
LV endocardial activation was characterized by a pattern of spread less tangential than that of activation of the epicardium, so visually the former displays a less fluent appearance. This is attributable to a delicate and highly individual interplay of excitation spread through the specialized conduction tissue and the working myocardium. As can be seen in Fig 4⇑, the breakthrough location was localized on the septum or the anterior wall in all patients for whom endocardial data were available. LV endocardial breakthrough occurred in AMI patients on the inferolateral, posterior, or septal endocardium in 8 of 13 cases. A second breakthrough occurred in 10 of these patients; it was located in 2 patients on the posterior endocardium, in 5 on the anterior endocardium, and in 2 on the midlateral endocardial surface. In the IMI group, the primary breakthrough site was anterior, anteroseptal, or both in 7 of 8 patients; 2 had secondary breakthroughs, located on the posteroseptal endocardial surface. In the CMI group, there was an approximately equal distribution of the primary breakthrough sites on the anterior and posterior endocardium (4 and 3 cases, respectively). Five of 7 patients had secondary breakthroughs. The spread of the activation wave front followed a rather uniform pattern, spreading radially from the breakthrough site towards the noninfarcted regions. If several breakthrough sites were present, a merging of the wave fronts occurred. There was significant local conduction delay in the LV endocardial peri-infarction zone, intrainfarction zone, or both in 24 of 28 patients (86%).
After the emergence of the RV epicardial breakthrough, the activation wave spread more or less radially over the RV and progressed in a right-to-left direction over the apex and the anterior and posterior septa. Spread of the activation wave over the anterior septum and the posterior septum was described by Wyndham et al30 as anterior and posterior septal crossing, respectively. In this way, the wave front ultimately reached the LV epicardial surface. However, an important activation contribution can also be derived from a distinct secondary LV epicardial breakthrough. This was observed in 29 of 45 patients (64%). Because of the partial or total apical involvement in all anterior lesions, radial transapical spread was not observed in patients with AMI. Consequently, in its further course the wave front spread lost its initially radial character and a circumnavigation of the scar by the activation front was observed. This resulted in a rotation-like spread pattern around the apex over the LV epicardium. In certain cases this rotation was preceded by a similar rotation on the endocardium (Fig 6⇓). The time sequence of the anterior and posterior crossing and/or breakthrough determined the main pattern of spread of the LV epicardial wave front. We observed the following characteristic LV epicardial activation patterns: (1) A radial pattern with both transapical and transseptal spread in a right-to-left direction, (2) counterclockwise rotation when posterior crossing or breakthrough precedes the anterior one, and (3) pincerlike activation when both of the septal crossings and/or breakthroughs occur nearly simultaneously and merge at the lateral LV wall.
The occurrence of these three epicardial activation patterns in relation to MI location is summarized in Table 3⇓. The radial pattern of epicardial activation is typical for IMI, and it was observed in all IMI patients. In addition, it was seen in 1 patient with CMI. The activation wave front omitted the infarcted region (typically of the inferoposterior wall), thereby leaving an electric hole (Fig 7⇓). The counterclockwise activation spread was found in 8 of 21 patients with AMI (Fig 8⇓). In the remaining 13 patients, a double-pronged pattern was observed. No patient with AMI had a radial transapical spread, a fact that can be explained by the apical involvement in all patients of this group (Fig 3⇑). In 5 of 8 patients with CMI, the pincerlike pattern was the most prevalent activation pattern (Fig 6⇑). A significant local peri-infarction conduction delay, an intrainfarction conduction delay, or both on the endocardium, epicardium, or both was observed in 41 of 45 patients (91%). Examples of typical patterns with infarction-related delayed conduction on the epicardium are presented in Figs 9⇓ and 10⇓. More particularly, the phenomenon of bridging of electrically dead regions by a ribbonlike conducting tissue was observed on the endocardium in 5 patients and on the epicardium in another 5. Typical examples are shown in Fig 10⇓.
We identified 5 patients fulfilling the criteria for LAFB according to the preoperative surface 12-lead ECG. The endocardial activation started in all patients on the posteroseptal endocardial surface. In 3 of these patients, no subsequent endocardial breakthroughs were observed. In the remaining 2, a second breakthrough was observed on the anterobasal endocardium (segment 3). However, it activated only a quite narrow basal ribbonlike area with further activation spread toward the basal lateral wall.
Right-axis deviation of ≥+90° was found on the preoperative ECG in 5 patients. However, in only 2 were the additional criteria for left posterior fascicular block (S1Q3 pattern) fulfilled. In both cases the endocardial activation pattern was characterized by breakthroughs on the anteroseptal endocardial wall and by lack of any breakthroughs in the posteroseptal region.
Surface ECG Analysis
In all but 1 patient (patient 34), the standard 12-lead ECG criteria for Q-wave MI were fulfilled. The patient without signs of old MI had a discrete posterior lesion on the posterobasal aspect of the LV (segment 9). The results of the analysis of QRS notching are summarized in Table 4⇓. Notching of the QRS complex was strongly correlated with the epicardial activation pattern (χ2, P<.001). Notching was found in all patients with pincerlike activation pattern (Fig 11⇓). Two of the 4 patients with a counterclockwise rotation pattern and QRS notching had a distinct LV epicardial breakthrough located in the lateral basal area (segment 6). Furthermore, 2 of the 8 patients with a radial activation pattern and multiphasic QRS complexes demonstrated prominent topographically unrelated areas of terminal activation.
Recent technical advances in intraoperative computer-assisted cardiac mapping allow for the collection of relatively extensive data on ventricular activation. However, the sequence of the ventricular activation in the infarcted heart during sinus rhythm has not been extensively studied until now. Our approach is based on a single-beat analysis of epicardial data of the entire heart and of the endocardial data from the LV. We believe that this method, compared with sequential mapping, yields more consistent and reliable results not biased by the risk of beat-to-beat activation variability during open-heart surgical procedures.
Proper interpretation of the electrode recordings is a crucial factor limiting the ultimate clinical usefulness of cardiac mapping.31 The discussion about the suitability of either the unipolar or the bipolar recording technique seems to be inherent to cardiac mapping.32 As pointed out by Pagé et al25 and Ideker et al,33 unipolar electrograms are more suitable than bipolar electrograms for detecting larger, organized, and more cohesive activation wave fronts. Therefore, we believe this technique is appropriate for the analysis of the overall activation sequence of the ventricles. However, the interpretation of unipolar signals is not without drawbacks. Unipolar ventricular signals with a smooth QS pattern (ie, without any other clear deflection or notch within the QS complex) constitute a specific problem related to the unipolar recording technique. On the basis of data from both canine and human infarcted hearts, it is generally accepted that this pattern, if observed on the epicardium, is indicative of transmural necrosis.1 2 3 However, the situation is more complicated on the endocardium and especially on the LV septum where QS complexes normally occur, and it might reflect the site of the endocardial breakthrough with a subsequent left-to-right transseptal activation wave front spread. To our knowledge, there are neither experimental nor human data available for clarification of this question. We have used the spatial filtering approach to solve this problem. This approach proved useful in distinguishing QS signals due to septal LV endocardial breakthrough in patients with MI not involving the septum.
Epicardial mapping during open-heart surgery creates a highly artificial environment for the beating heart, because a considerable part of its surface is exposed to relatively cold air. Although this technique has been used in both experimental and clinical settings for more than two decades, it was only recently demonstrated that data acquired in this manner are not significantly distorted.34 This is especially true for isochrone analysis, which is less susceptible to variations than is analysis of the absolute potentials on the epicardium.
Normal Sequence of Ventricular Depolarization as Determined by the Applied Mapping Technique
During the 8-year experience of our group with intraoperative mapping, we have seen only 1 patient free of structural heart disease for whom both epicardial and LV endocardial mapping data were available (the patient successfully underwent antiarrhythmic surgery for idiopathic LV tachycardia originating in the inferoposterior region). Corresponding isochronal maps are shown in Fig 12⇓. We observed three breakthroughs on the epicardium and three on LV endocardium. Although a substantially higher number of epicardial breakthroughs was recently observed in normal human hearts,35 the global activation pattern in this patient is consistent with the generally accepted sequence of normal ventricular depolarization in human hearts.13 19 36 From a methodological standpoint, this experience further points out that the mapping technique used is, despite its limitations (see “Limitations and Conclusions”), an adequate tool for the study of overall ventricular activation.
Endocardial Data From the Infarcted LV
The major purpose of previous catheter mapping studies was the search for arrhythmogenic regions on the endocardium.6 16 37 38 39 Some data from these studies can be compared with our data, and a summary of the results of previous studies is given in Table 5⇓. Vassallo et al9 studied a group of patients, clinically the most comparable to our group, with chronic MI and VT. There is good agreement in the sequence of regional activation and in the location of the breakthrough sites, with one always being present at the LV septal surface. We believe that the smaller number of breakthrough sites described by Vassallo et al is attributable to the lower spatial resolution of their approach. Also, the differences in the total activation times between the two approaches might be due to different definitions of local activation times used for bipolar and unipolar recordings as well as a different definition of the zero time. Furthermore, in Table 5⇓ it can be seen that some authors included in their studies patients with bundle-branch blocks, patients on antiarrhythmic medication, or both. Obviously, in these patients there were additional factors increasing the absolute duration of ventricular activation. The finding of Wiener et al12 that patterns of endocardial activation do not differ significantly in normal, hypokinetic, akinetic, and dyskinetic ventricles is at variance with our results. We have observed the absence of local activation or its important delay in both the intrainfarction zone and the peri-infarction zone.
Epicardial Data From Hearts With Old MI
The first data on the infarcted heart in humans were reported in the early 1970s by Daniel et al,5 who also recognized the need for further studies of the occurrence and distribution of Q waves in unipolar epicardial recordings from normal human hearts. The issue of normal occurrence of qR complexes on the LV epicardial surface—a topic of essential but underestimated importance for the interpretation of both epicardial and precordial electrograms—was earlier addressed by Roos et al.35 Epicardial activation data presented by Wiener et al18 from patients with various regional contraction abnormalities are comparable to data from our cohort for a part of the group (presence of akinesis or dyskinesis in 10 patients). There is good agreement concerning the location of the initial epicardial breakthrough over the anterior RV. However, the latest epicardial activation was directly related to the site of MI in 60% of cases.18 Our results show a higher occurrence of this phenomenon (91%). This difference can be explained by the limited spatial resolution of the handheld probe mapping, with which very localized areas of delayed activity could have been missed.
Impact of Spatial Resolution on the Interpretation of Epicardial Mapping
Multiple epicardial breakthroughs have been observed in high-resolution mapping studies. Using a sequential 1124-electrode recording technique in normal canine heart, Arisi et al39 40 observed 20 to 35 epicardial breakthroughs. Such an excellent spatial resolution with a mean interelectrode distance of 2 mm has not yet been used in any human in vivo study. Several factors might contribute to this high number of breakthroughs:
1. As suggested by Arisi et al,40 bulges in the wave fronts proceeding from endocardium to epicardium might play an important role in generating multiple epicardial breakthroughs. These bulges are due to the spatial arrangement of endocardium, which is not smooth and planar but a rather complex three-dimensional surface with multiple hollows and grooves.
2. A different definition of breakthrough was used in the present study: using isopotential (and not isochronal) maps, an island of negativity surrounded by positive peaks was considered a breakthrough. Such a definition is not identical to the classic definition of breakthrough in isochronal analysis, the presence of a site at which all surrounding tissue shows a later activation time.
3. There was a different depth of penetration of terminal Purkinje fibers into the ventricular wall in human and canine hearts.36
Nevertheless, in a recent study using single-beat simultaneous 117-electrode isochronal mapping in human hearts, Pieper and Pacifico41 demonstrated up to 12 epicardial breakthroughs. These breakthroughs were often small and did not exceed several millimeters, so neighboring microbreakthroughs merged into 1 major breakthrough within 1 to 2 milliseconds. These macrobreakthroughs subsequently determined the overall activation pattern. Although the microbreakthroughs appeared mainly on the distal two thirds of the epicardial surface, they were also observed on the basal epicardium close to the atrioventricular groove. Because appropriate definitions of the local activation time in unipolar and bipolar recordings do not differ significantly at appropriate sampling rates,42 43 superior spatial and time resolution in the study by Pieper and Pacifico is the most probable explanation for the observed differences in the total number of epicardial breakthroughs compared with our findings in the normal heart (Fig 12⇑).
On the other hand, Myerburg et al44 45 did not observe in canine preparations the presence of large numbers of endocardial breakthroughs. Although the very complex arborization pattern of the entire Purkinje system was precisely defined in that study, endocardial activation was shown to start simultaneously from numerous Purkinje endings concentrated in certain endocardial areas, so that no more than 3 true endocardial breakthroughs were observed. This might be mainly due to extended interconnections of the distal branches.
Our analysis, based on unipolar recordings, did not demonstrate the presence of more than 3 epicardial breakthroughs. Several factors of both methodological and biological character might contribute to this finding:
1. Spatial and time resolution: The use of 63 epicardial electrodes on abnormally hypertrophied or dilated hearts with aneurysm might lead to limited spatial resolution, which can be further impaired in regions where bulging occurs. This might be particularly true for the apical region frequently involved in the aneurysm. On the free wall, the spatial resolution is lower on the basal circumference because, with the design of our sock electrode, the interelectrode distance progressively increases in the basal direction. In fact, recent human and experimental data suggest that a minimum spatial resolution of 3.2 mm and a temporal resolution of 2 milliseconds are needed.43 46 Because we used a 500-Hz sampling rate and an electrode spacing of up to 20 mm, breakthroughs appearing relatively closely and within 2 milliseconds could not be detected.
2. Use of a unipolar recording technique focusing on larger cohesive wave fronts. This may have led to missing of microbreakthroughs, which generate far less electromotive force than larger breakthroughs. This might especially be the case in the presence of simultaneous activation in deeper layers of the ventricular wall.
3. Presence of substantially altered morphological substrate in infarcted hearts. This could be responsible for the differences between other studies and ours in two ways. First, impulses cannot excite muscle cells lateral to the orientation of the specialized conducting fibers unless Purkinje fibers–to–myocyte junctions are present.45 Although Purkinje fibers can survive even in large infarctions, significant damage to the electrical junctions between Purkinje fibers and ventricular muscle occurs.47 48 An electrophysiological correlate of this damage could be a Purkinje-myocyte transmission block, in which the impulse conducted down the specific conduction system does not excite ventricular myocardium in the infarcted area. Consequently, endocardial activation would be initiated on a limited number of sites and, thus, both the number of endocardial and epicardial breakthroughs would be reduced. Under such circumstances, myocyte-to-myocyte propagation would play a decisive role in determining the global ventricular activation pattern. This explanation appears to be particularly plausible because a substantial part of the Purkinje-fiber network is located in the apical and periapical subendocardial area, which was involved in the infarction in the majority of patients in our study (Fig 3⇑). Myocyte-to-myocyte activation wave propagation along the longitude of predominantly circumferentially oriented fibers in the basal parts of the LV49 are also consistent with the observed pincerlike activation pattern.
A second possible explanation is that intramural fibrosis of the infarcted heart further interferes with the normal transmural endocardial-to-epicardial activation spread. In normal myocardium, preferential spread of the activation wave along fibers exhibiting a gradual orientation shift was observed experimentally,50 and its possible role in accelerating conduction toward epicardium was suggested on the basis of simulation studies.51 However, in infarcted hearts, endocardial-to-epicardial conduction follows a circuitous route,52 and excessive slowing of conduction up to 30 mm/s has been observed.53 54 Consequently, the course of the activation wave front is considerably lengthened, and zigzag propagation occurs. These mechanisms might further limit the number of observed epicardial breakthroughs. Furthermore, the importance of M cells in accelerating intramural conduction toward epicardium is presently not well understood, but damage to them in infarcted hearts might also play a role.55 Further studies by means of high-resolution surface and intramural cardiac recordings are needed to elucidate the answers to these questions.
It has to be emphasized that one of the major goals of our study was to acquire reference data to which surface ECG recordings can be related. We have presumed that electrical microprocesses detected with bipolar electrograms, but not with unipolar electrograms, will not be of sufficient electromotive force to be reflected on the standard ECG. Therefore, the unipolar recording technique, despite its inherent drawbacks, may be the appropriate approach for achieving this goal. Our finding that the concomitant presence of several major epicardial activation wave fronts is required for them to be reflected by abnormal low-frequency notching of QRS complexes yields some indirect evidence supporting our unipolar approach. Our clinical experience with mapping-guided antiarrhythmic surgery further supports the hypothesis that unipolar recordings at the given spatial and time resolution offer valid data suitable for elucidation of global patterns of electrical activation in infarcted ventricles.21
Surface ECG Versus Epicardial Activation Pattern
Notching of the QRS complex discernible in the conventional ECG was designated low-frequency notching by Langner and Lauer56 and was frequently observed in patients with coronary heart disease and rarely in normal subjects. The presence of such notched, multiphasic QRS complexes may be related to the struggle between competing generator sites.29 This is the case when several epicardial wave fronts operate simultaneously. Thus, it should also apply to the pincerlike activation pattern, which comprises at least two spatially separated wave fronts. One could hypothesize that the normally occurring epicardial mosaic of short-lived microbreakthroughs does not contribute to low-frequency QRS notching. Our results seem to confirm this hypothesis. Low-frequency QRS complex notching was present in all patients with a pincerlike activation pattern. Furthermore, 4 of 15 patients with QRS notching and a different activation pattern demonstrated multiple competing activation wave fronts on the epicardium. These results are in accordance with a previous observational study describing the occurrence of RSR’ complexes (ie, the simplest form of the multiphasic QRS complex) not related to bundle-branch block in patients with old MI.57
Post-QRS Electrical Activity Recorded Directly From the Heart
The reports by Daniel et al4 5 were of direct in vivo observation of the epicardial activation sequence in patients with AMI showing delayed terminal activation over the anterolateral and apical LV. In another case of AMI, delayed epicardial activation was observed in area adjacent to the scar. Therefore, it was postulated that interference with the spread of excitation through the Purkinje network could have delayed the wave front that reached the infarcted region between the infarcted distal anterior wall and the anterior base with a substantial delay (activation between 85 to 135 milliseconds after QRS onset). We repeatedly observed such activation phenomena in the present study.
Electrical activity exceeding the duration of the surface QRS complex observed on the epicardium of patients undergoing surgery for VT was reported by Fontaine et al58 in their pioneer study using computer-assisted mapping. Their results for late activation and those of Wiener et al18 are comparable to ours. Both groups found late epicardial activity extending beyond the surface QRS in no more than 20% of patients with MI (with or without VT). Similarly, Josephson et al15 found epicardial activity exceeding the surface QRS complex by at least 10 milliseconds in 11.5% of patients operated on for VT due to coronary artery disease. Post-QRS epicardial activity was found in 14 of 45 patients in our study. The results of another study by Wiener et al17 suggested a higher incidence of late epicardial activity, which was detected in almost half of the patients with aneurysms. The cohort analyzed by Klein et al14 is comparable to our patients. The focus of their study was delayed activation (defined as such when it occurred >100 milliseconds after QRS onset). It was observed in 95% of patients with arrhythmias and in 12% without arrhythmias. If a similar criterion would be applied to our population, 36% of our patients showed delayed epicardial or endocardial activation. However, the timing of the mean latest epicardial activation is substantially shorter in our study (≈94±20 milliseconds, as opposed to 137±21 milliseconds in the study by Klein et al) (Table 5⇑).
Late LV endocardial activity exceeding the duration of the QRS complex has repeatedly been reported on the basis of catheter studies in patients with chronic MI and VT. Cassidy et al8 observed late endocardial activities in 20% of all recorded electrograms in 52 patients with MI and VT. In a similar cohort of 41 patients, Vassallo et al36 found endocardial LV electrograms exceeding the duration of the surface QRS complex in all patients. Wiener et al17 observed post-QRS endocardial activity in more than half of patients with LV aneurysms mapped during cardiac surgery. In another intraoperative mapping study, Kienzle et al16 detected late electrograms on the LV endocardium in 4 of 13 patients with MI and VT. On the other hand, Wiener et al12 did not observe endocardial electrograms exceeding the duration of the surface QRS complex in any of 16 mapped patients. In a recent study, Hood et al59 did not observe any post-QRS activity in transmural ventricular recordings from 8 patients with old MI and VT.
It has to be emphasized that all the studies cited above differ considerably from the present one in using (1) bipolar recordings in all studies; (2) catheter techniques in the majority of studies; (3) various definitions of late activity, which was usually defined as the occurrence of any electrical activity manifested as deviation from baseline7 17 (in only one study was a definition used that was comparable to ours, based on distinct local activity15 ); and, (4) in some studies, patients with bundle-branch blocks, those on antiarrhythmic medication, or both were included. One has to bear in mind that these important differences in methodology may account for considerable differences in all comparisons of our data on post-QRS electrical activity (mainly on the endocardium) with data from the other studies mentioned.
The unipolar recording technique was used in a previous report from our laboratory.60 In a relatively small cohort, post-QRS activity (endocardial, epicardial, or both) was found in 83% of patients without bundle-branch block. However, the post-QRS activity was defined as “any depolarization extending beyond the unfiltered surface QRS complex, including the waveforms occurring after the intrinsic deflection.” Therefore, in some electrograms displaying post-QRS activity, the local activation per se was set before the offset of thoracic QRS.
Chronic MI With and Without VT
All our patients underwent surgery because of recurrent VT. Therefore, no data are available for comparison of ventricular activation sequences in patients with and without VT. Nevertheless, this issue is important for judging whether our results on the global activation sequence derived from patients with VT can be extrapolated to infarcted hearts without VT. Currently there is no clear answer to this question. The existing data suggest that the structural differences between infarcted hearts with and without VT are more quantitative than qualitative.20 This would, in turn, mean that the global activation data can be extrapolated to nonarrhythmogenic infarcted hearts. On the other hand, it has to be considered that, compared with non-VT MIs, infarcts with VT are characterized by a larger absolute amount of patchy myocardial necrosis and increased areas of irregularly spared subendocardium.20 Recording electrodes localized directly above this structurally changed ventricular muscle may detect some kind of distinct local activity, delayed or not. According to the criteria used in our study, such signals would not be classified as “dead.” Thus, the mapping correlates of these morphological differences would be mainly characterized by the occurrence of regions displaying local activation despite their macroscopically infarcted appearance. In fact, this phenomenon was observed in our study quite often, and the area considered electrically dead tended to be smaller than the visually estimated infarction. Furthermore, in 9% of our patients, the expected electrically dead areas were absent. These data could be the electrophysiological counterpart to the arteriography findings of Saxon et al,61 who observed persistence of a significant and clearly identifiable blood supply to the infarcted region in 84% of MI patients with sustained VT.
Pagé et al25 observed in experimental canine MI with inducible VT that epicardial lesions characterized by a missing local activation were bridged by a tissue band with recordable local activation typically displaying a distinct rs deflection after the cavity QS complex. Such bridging was not observed in non-VT infarcts, in which tissue displaying delayed local activation formed “lagoons” in the otherwise electrically silent infarction. Bundles of viable myocardial fibers embedded in fibrous tissue are a common finding in chronic MI. Continuous myocardial bundles traversing the MI scar have also been confirmed in human infarcted hearts prone to VT.52 Although the surviving myocytes were mostly observed in the subendocardium, they were found also intramurally and subepicardially.
The electrophysiological correlates of these bridges in the experiments were very relevant: they served as a common pathway for the figure-of-8 reentrant circuit during VT. However, a similar phenomenon was not a ubiquitous feature in our study and was observed on the epicardium in only 5 patients (11%) and on the LV endocardium in 5 patients (see Fig 10⇑). Well-recognized significant differences in human compared with canine MI anatomy or still suboptimal spatial resolution of our recording technique (64 electrodes, as in the study by Pagé et al,25 but used on larger human hearts) might account for this discrepancy. The latter factor was considered an important reason that presumably subepicardially or subendocardially localized reentrant circuits during VT could not be mapped completely in a previous report from our laboratory.21 Furthermore, we cannot exclude the presence of three-dimensional bridging detectable only by intramural recordings.
LAFB and Left-Axis Deviation
Cardiac mapping data on the precise pattern of ventricular activation sequence in patients with LAFB and left-axis deviation are scarce. However, such data are of fundamental importance for the comprehensive understanding of surface ECGs in this condition. In vivo mapping studies are the only potential source of this knowledge, because intraventricular conduction disturbances are not subject to absolute pathological confirmation, and animal data have limited validity for humans. In contrast to previously reported data,9 we found a typical pattern of LV endocardial activation in patients with LAFB. This was characterized in most patients by an initial breakthrough site on the posterior septum with a subsequent counterclockwise activation spread via the inferior wall towards the lateral and anterior LV endocardial surface. Wyndham et al62 studied 4 patients with LAFB by means of intraoperative epicardial mapping. However, epicardial events are only the epiphenomena of a primary different endocardial activation pattern in LAFB. To our knowledge, our report is the first direct observation of a precise endocardial activation sequence in humans with LAFB. A lack of epicardial breakthrough on the anterolateral LV, as described previously,62 was observed in 3 patients; in 2 there was a clear breakthrough in epicardial segment 3 (anterobasal region). In accordance with the above-mentioned study62 and previous experimental data,63 we always observed an area on the anterolateral epicardium activated within the last 10 milliseconds of the ventricular activation. The basic way in which LAFB patients differed from a group without conduction disturbances studied previously by Wyndham et al13 was that an epicardial breakthrough normally present on the basal anterolateral LV wall did not appear; on the contrary, this region was activated the latest. Our results give further support to the observation that LAFB in patients with coronary artery disease comprises a delayed activation of the basal anterolateral LV epicardium.62
Limitations and Conclusions
The limitations of our study are inherent to most of the clinically used mapping systems. The major limitation for a comprehensive characterization of the ventricular activation sequence is the lack of endocardial data from the RV. This problem has been already solved technically, and data should be available in the near future from patients undergoing antiarrhythmic surgery in whom both left and right endocardial mapping is clinically indicated. Further improvement in the form of high-resolution intramural recordings (comparable to the in vitro data on normal ventricular activation presented by Durrer et al3 ) is very difficult (though not unfeasible59 ) in intraoperative settings. The mapping system used is based on recordings from a total of 127 electrodes for the entire epicardial and LV endocardial surface. The resulting spatial resolution of the system—though superior to the results of approaches taken in previously published studies on activation of infarcted hearts in sinus rhythm—might still not be sufficient for detection of some subtle phenomena. Furthermore, because of the time resolution of 2 milliseconds (ie, a sampling rate of 500 Hz) used, short-lived electrical processes could not be detected.41 64 A system supporting twice the number of electrodes with 1-kHz sampling is currently being implemented in our laboratory. Our study does not include a control group of normal subjects; the technique used precludes collection of data from normal hearts on a larger scale.
Nevertheless, we believe that the data presented offer a solid basis for a more precise interpretation of a wide range of electrocardiographic and electrophysiological data in patients with old MI. In our opinion, the data on the characteristic activation patterns can be generalized and applied as a framework to infarcts that are of similar extent but without complicating VT. They should therefore provide a better insight into the exact nature of disturbances of the ventricular activation sequence in this common and clinically very relevant pathological condition.
This work was supported by the Medical Research Council of Canada. We gratefully acknowledge the meticulous technical and administrative assistance of Martine Germain and the expert secretarial help of Diane Abastado.
Presented, in part, at the 65th Scientific Sessions of the American Heart Association, New Orleans, La, November 16-19, 1992, and published in abstract form in Circulation. 1992;90(pt 2):I-584.
- Received June 16, 1994.
- Revision received August 24, 1994.
- Accepted September 5, 1994.
- Copyright © 1995 by American Heart Association
Wilson FA, Johnston FD, Hill IGW. The form of the electrocardiogram in experimental myocardial infarction, IV: additional observations on the later effects produced by ligation of the anterior descending branch of the left coronary artery. Am Heart J. 1935;10: 1025-1034.
Durrer D, Van Lier AAW, Bueller J. Epicardial and intramural excitation in chronic myocardial infarction. Am Heart J. 1964;6: 765-776.
Daniel TM, Boineau JP, Sabiston DC Jr. Comparison of human ventricular activation with a canine model in chronic myocardial infarction. Circulation. 1971;44:74-89.
Cassidy DM, Vassallo JA, Buxton AE, Doherty JU, Marchlinski FE, Josephson ME. The value of catheter mapping during sinus rhythm to localize site of origin of ventricular tachycardia. Circulation. 1984;69:1103-1110.
Cassidy DM, Vassallo JA, Marchlinski FE, Buxton AE, Untereker WJ, Josephson ME. Endocardial mapping in humans in sinus rhythm with normal left ventricles: activation patterns and characteristics of electrograms. Circulation. 1984;70:37-42.
Wyndham CR, Meeran MK, Smith T, et al. Epicardial activation of the intact human heart without conduction defect. Circulation. 1979;59:161-168.
Klein H, Karp RB, Kouchoukos NT, Zorn GL Jr, James TN, Waldo AL. Intraoperative electrophysiologic mapping of the ventricles during sinus rhythm in patients with a previous myocardial infarction: identification of the electrophysiologic substrate of ventricular arrhythmias. Circulation. 1982;66:847-853.
Josephson ME, Simson MB, Harken AH, Horowitz LN, Falcone RA. The incidence and clinical significance of epicardial late potentials in patients with recurrent sustained ventricular tachycardia and coronary artery disease. Circulation. 1982;66:1199-1204.
Kienzle MG, Miller J, Falcone RA, Harken A, Josephson ME. Intraoperative endocardial mapping during sinus rhythm: relationship to site of origin of ventricular tachycardia. Circulation. 1984;70:957-965.
Wiener I, Mindich B, Pitchon R. Determinants of ventricular tachycardia in patients with ventricular aneurysms: results of intraoperative epicardial and endocardial mapping. Circulation. 1982;65: 856-861.
Wiener I, Mindich B, Pitchon R, Pichard A, Kupersmith J, Estioko M, Jurado R, Camunas J, Litwak R. Epicardial activation in patients with coronary artery disease: effects of regional contraction abnormalities. Circulation. 1982;65:154-160.
Durrer D, Dam RT, Freud GE, Janse MJ, Meijler FL, Arzbaecher RC. Total excitation of the isolated human heart. Circulation. 1970;41:899-912.
Bolick DR, Hackel DB, Reimer KA, Ideker RE. Quantitative analysis of myocardial infarct structure in patients with ventricular tachycardia. Circulation. 1986;74:1266-1279.
Kaltenbrunner W, Cardinal R, Dubuc M, et al. Epicardial and endocardial mapping of ventricular tachycardia in patients with myocardial infarction: is the origin of the tachycardia always subendocardially localized? Circulation. 1991;84:1058-1071.
Damiano RJ, Blanchard SM, Asano T. The effects of distant potentials on unipolar electrograms in an animal model utilizing the right ventricular isolation procedure. J Am Coll Cardiol. 1988;11: 1100-1109.
Willems JL, Robles de Medina EO, Bernard R, et al. Criteria for intraventricular conduction disturbances and pre-excitation: World Health Organization/International Society and Federation for Cardiology Task Force Ad Hoc. J Am Coll Cardiol. 1985;5: 1261-1275.
Chou TC. Electrocardiography in Clinical Practice. Orlando, Fla: Grune & Stratton; 1992:147-225.
Flowers NC, Horan LG, Thomas JR, Tolleson WJ. The anatomic basis for high-frequency components in the electrocardiogram. Circulation. 1969;39:531-539.
Wyndham CR, Smith T, Meeran MK, Mammana R, Levitsky S, Rosen KM. Epicardial activation in patients with left bundle branch block. Circulation. 1980;61:696-703.
Berbari EJ, Lander P, Scherlag BJ, Lazzara R, Geselowitz DB. Ambiguities of epicardial mapping. J Electrocardiol. 1992;24(suppl): 16-20.
Ideker RE, Tang AS, Daubert JP. On the trail of ventricular tachycardia or the adventure of the unspeckled band. PACE Pacing Clin Electrophysiol. 1988;11: 650-655.
Green LS, Taccardi B, Ershler PR, Lux RL. Epicardial potential mapping: effects of conducting media on isopotential and isochrone distributions. Circulation. 1991;84:2513-2521.
Roos JP, Dam RT, Durrer D. Epicardial and intramural excitation of normal heart in six patients 50 years of age and older. Br Heart J. 1968;30:630-637.
Arisi G, Stilli D, Spaggiari S, Baruffi S, Musso E, Macchi E, Taccardi B. Potential distribution on the dog’s ventricular surface as revealed by 1912 epicardial leads. In: Macfarlane PW, ed. Progress in Electrocardiology. Tunbridge Wells, UK: Pitman Medical Publishing Co Ltd; 1979:59-63.
Arisi G, Macchi E, Baruffi S, Spaggiari S, Taccardi B. Potential fields on the ventricular surface of the exposed dog heart during normal excitation. Circ Res. 1983;52:706-715.
Myerburg RJ, Nilsson K, Gelband H. Physiology of canine intraventricular conduction and endocardial excitation. Circ Res. 1972;30:217-243.
Myerburg RJ, Gelband H, Castellanos A, Nilsson K, Sung RJ, Bassett AL. Electrophysiology of endocardial intraventricular conduction: the role and function of the specialized conducting system. In: Wellens HJJ, Lie KI, Janse MJ, eds. The Conduction System of the Heart: Structure, Function and Implications. The Hague, The Netherlands: Martinus Nijhoff Medical Division; 1978:336-359.
Friedman PL, Fenoglio JJ, Wit AL. Time course for reversal of electrophysiological and ultrastructural abnormalities in subendocardial Purkinje fibers surviving extensive myocardial infarction in dogs. Circ Res. 1975;36:127-144.
Scherlag BJ, el-Sherif N, Hope R, Lazzara R. Characterization and localization of ventricular arrhythmias resulting from myocardial ischemia and infarction. Circ Res. 1974;35:372-383.
Armour JA, Randall WC. Structural basis for cardiac function. Am J Physiol. 1970;218:1517-1523.
Watabe S, Taccardi B, Lux RL, Ershler PR. Effect of nontransmural necrosis on epicardial potential fields: correlation with fiber direction. Circulation. 1990;82:2115-2127.
Pollard AE, Burgess MJ, Spitzer KW. Computer simulations of three-dimensional propagation in ventricular myocardium: effects of intramural fiber rotation and inhomogeneous conductivity on epicardial activation. Circ Res. 1993;72:744-756.
Spear JF, Michelson EL, Moore EN. Reduced space constant in slowly conducting regions of chronically infarcted canine myocardium. Circ Res. 1983;53:176-185.
de Bakker JM, van Capelle FJ, Janse MJ, et al. Slow conduction in the infarcted human heart: ‘zigzag’ course of activation. Circulation. 1993;88:915-926.
Antzelevitch C, Sicouri S, Litovsky SH, et al. Heterogeneity within the ventricular wall: electrophysiology and pharmacology of epicardial, endocardial, and M cells. Circulation. 1991;69:1427-1449.
Fontaine G, Guiraudon G, Frank R, Vedel J, Grosgogeat Y, Cabrol C, Facquet J. Stimulation studies and epicardial mapping in ventricular tachycardia: study of mechanisms and selection for surgery. In: Kulbertus HE, ed. Reentrant Arrhythmias: Mechanisms and Treatment. Baltimore, Md: University Park Press: 1977:334-350.
Hood MA, Pogwizd SM, Peirick J, Cain ME. Contribution of myocardium responsible for ventricular tachycardia to abnormalities detected by analysis of signal-averaged ECGs. Circulation. 1992;86:1888-1901.
Lacroix D, Savard P, Shenasa M, et al. Spatial domain analysis of late ventricular potentials: intraoperative and thoracic correlations. Circ Res. 1990;66:55-68.
Gallagher JJ, Ticzon AR, Wallace AG, Kasell J. Activation studies following experimental hemiblocks in the dog. Circ Res. 1974;35: 752-763.
Arisi G, Macchi E, Corradi C, Lux RL, Taccardi B. Epicardial excitation during ventricular pacing: relative independence of breakthrough sites from excitation sequence in canine right ventricle. Circ Res. 1992;71:840-849.