Author + information
- Received October 3, 2000
- Revision received December 28, 2000
- Accepted January 24, 2001
- Published online May 1, 2001.
- Tamir Wolf, PhD∗,* (, )
- Lior Gepstein, MD, PhD∗,
- Uzi Dror, BSc∗,
- Gal Hayam, BSc∗,
- Rona Shofti, DVM∗,
- Asaph Zaretzky, DVM∗,
- Gideon Uretzky, MD†,
- Uri Oron, PhD‡ and
- Shlomo A Ben-Haim, MD, DSc∗
- ↵*Reprint requests and correspondence: Dr. Tamir Wolf, Cardiovascular System Laboratory, The Bruce Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, Efron Street, POB 9649, Haifa 31096, Israel
This study delineates between infarcts varying in transmurality by using endocardial electrophysiologic information obtained during catheter-based mapping.
The degree of infarct transmurality extent has previously been linked to patient prognosis and may have significant impact on therapeutic strategies. Catheter-based endocardial mapping may accurately delineate between infarcts differing in the transmural extent of necrotic tissue.
Electromechanical mapping was performed in 13 dogs four weeks after left anterior descending coronary artery ligation, enabling three-dimensional reconstruction of the left ventricular chamber. A concomitant reduction in bipolar electrogram amplitude (BEA) and local shortening indicated the infarcted region. In addition, impedance, unipolar electrogram amplitude (UEA) and slew rate (SR) were quantified. Subsequently, the hearts were excised, stained with 2,3,5-triphenyltetrazolium chloride and sliced transversely. The mean transmurality of the necrotic tissue in each slice was determined, and infarcts were divided into <30%, 31% to 60% and 61% to 100% transmurality subtypes to be correlated with the corresponding electrical data.
From the three-dimensional reconstructions, a total of 263 endocardial points were entered for correlation with the degree of transmurality (4.6 ± 2.4 points from each section). All four indices delineated infarcted tissue. However, BEA (1.9 ± 0.7 mV, 1.4 ± 0.7 mV, 0.8 ± 0.4 mV in the three groups respectively, p < 0.05 between each group) proved superior to SR, which could not differentiate between the second (31% to 60%) and third (61% to 100%) transmurality subgroups, and to UEA and impedance, which could not differentiate between the first (<30%) and second transmurality subgroups.
The degree of infarct transmurality extent can be derived from the electrical properties of the endocardium obtained via detailed catheter-based mapping in this animal model.
The ability to discern the degree of myocardial infarction (MI) is considered a cornerstone in diagnostic cardiology. Although certain modalities investigate the size of the infarct (1–6), the extent of transmurality is also of prognostic importance (7,8). Numerous studies (9–11)have attempted to correlate this anatomical characteristic with bedside assessment of patients with MI, mainly by electrocardiographic assessment upon patient arrival, yielding varying results. Other methods used for infarct transmurality assessments have sparse data and are limited by lack of appropriate in vivo assessment capacity and technical constraints (12–14).
In recent years, there have been a growing number of patients in which reperfusion of dysfunctional regions results in improved cardiac performance. Nevertheless, such recovery is also dependent upon complex interactions among tissue perfusion, the amount of viable myocardium and the extent of transmural necrosis (8). This further emphasizes the role that quantification of the degree of infarct transmurality may have and its effect on reperfusion strategy.
One new approach for myocardial viability assessment in the catheterization laboratory has been mapping the electromechanical properties of the endocardium. This approach has enabled determination of the presence, precise location and endocardial extent of infarcted tissue (15–17). However, the ability to distinguish between various transmurality subtypes using this technique has not yet been evaluated.
The aim of the present study was to test the hypothesis that the characteristics of intracardiac electrical signals sampled from the endocardial surface may differentiate between infarcts extending to various degrees throughout the left ventricular (LV) wall, namely transmural versus subendocardial infarcts.
The study included 13 mongrel dogs (weight 20 to 32 kg) subjected to chronic coronary ligation. The investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the U.S. National Institutes of Health (NIH Publication No. 82-23, revised 1996). Anesthesia was induced with ketamine 10 mg/kg intravenously (IV) and diazepam 1 mg/kg IV, and after intubation was maintained by isofluorane 1% and fentanyl 0.25 mg/kg/min IV. Ventilation was kept constant with a veterinary anesthesia ventilator (Hallowell EMC, Model 2000, Pittsfield, Massachusetts). Left thoracotomy was performed, after which the left anterior descending (LAD) coronary artery was ligated distal to the first diagonal branch. Following surgery, the animals were treated with analgesics and antibiotics and allowed to recover for four weeks; afterwards LV mapping was performed on each of the dogs.
Nonfluoroscopic electromechanical mapping system
The nonfluoroscopic electromechanical mapping system has been described previously (18). Briefly, the system (NOGA, Biosense Webster, Tirat-Hacarmel, Israel) utilizes ultralow magnetic fields generated by three external magnetic field emitters that are placed beneath the operating table. Additional location sensors are integrated into 7F standard electrophysiologic deflectable tip catheters (NOGA-STAR, Biosense Webster), one serving as a reference catheter and the other as the mapping catheter. The data acquired by the sensors are sent to a processing unit, enabling real-time accurate tracking of the catheter tip within the LV cavity with minimal use of fluoroscopy.
A reference catheter was fixed to the back of the animal and a mapping catheter was introduced into the LV cavity under fluoroscopic guidance. The location of the navigated mapping catheter was recorded relative to the fixed reference catheter’s position, compensating for animal movement during the procedure. The real-time location and orientation of the mapping catheter tip were continuously displayed on a Silicon Graphics workstation, enabling accurate three-dimensional (3D) navigation of the catheter. By sequentially sampling the location and local electrogram of multiple endocardial sites, a 3D reconstruction of the LV chamber was created.
The quality of each sampled endocardial point is important for the accurate analysis of the acquired data and consequent 3D maps. The exclusion criteria utilized have been described elsewhere and include premature beat detection, variations in electrogram morphology, low location, loop and cycle length stabilities (16). Catheter pressure on the endocardium, as indicated by marked ST segment elevation (>0.1 mV above baseline) in the local unipolar electrogram was also evaluated.
Maps were obtained using a catheter with a 2-mm tip electrode and a closely spaced (0.5 mm) ring electrode (1 mm). Thus, both local unipolar and bipolar electrograms (filtered at 0.5 to 400 Hz and 30 to 400 Hz, respectively) were recorded at each of the sampled sites. In addition to the peak-to-peak unipolar and bipolar electrogram amplitudes (UEA and BEA, respectively), we studied unipolar signal slew rate (SR, defined as the maximum absolute dV/dt of the negative intrinsic deflection of the unipolar electrogram signal). In addition, local endocardial impedance was measured using a generator with a stabilized output amplitude producing a cine signal of 1 μA at 50 kHz as described previously (19). All electrical information obtained was color-coded and superimposed on the geometrical reconstruction of the LV chamber and was observed online during the mapping procedure, enabling real-time analysis of findings.
Linear local shortening maps
Regional mechanical function of the myocardium was assessed with a previously described algorithm (16)that calculates the fractional shortening of regions of endocardium at end-systole, termed local endocardial shortening (LS). Briefly, LS was calculated as the difference between the distance of each endocardial site and all its neighbors at end-diastole and end-systole (normalized to end-diastole). A positive LS ratio is given to a site in which the distance between two neighboring sites decreases during systole (i.e., physiological contraction), whereas a negative ratio depicts a site in which the distance between two neighboring sites increases during systole (abnormal myocardial motion).
Pathological verification of MI
At the end of each experiment the animal was sacrificed by administration of an intravenous KCl overdose and the heart was excised. The coronary arteries were then perfused with 300 ml of TTC (2,3,5-triphenyltetrazolium chloride, 5 g/250 ml normal saline) and the hearts fixed in 4% formaldehyde solution. Infarcted regions were identified as regions not stained by TTC, the presence of a fibrous scar, and myocardial thinning. The hearts were then sliced transversely into 5-mm to 7-mm sections and scanned on both sides. The method employed in determining the mean transmural extent in each slice has been described by Kemper and colleagues (12)and is shown in Figure 1. Using morphometric software, two radii drawn from the center of mass (LV chamber) to the maximal circumferential extent of the infarct defined both the infarcted area and the noninfarcted area. The total transmurality area was defined as the sum of the infarcted and noninfarcted areas. Subsequently, the mean transmurality extent (% transmurality) was defined as (infarct area/total transmurality area) × 100. Because a gradient of necrosis exists between the apical and basal sides of each slice, calculation of mean transmurality extent was performed for each side and the mean value presented as the final value of transmurality.
Finally, samples (6 μm sections) of the infarct tissue (mid-anterior wall) as well as samples of healthy tissue (lateral wall) were obtained from the various sections, embedded in paraffin and stained with hematoxilin-eosin and with Masson’s trichrome. Histologic examination and verification of the infarcted tissue were performed using conventional light microscopy.
Correlation between endocardial indices and transmurality
The geometrical reconstructions of the LV were similarly divided into transverse sections. Comparison was made possible because of the high spatial resolution of the mapping system (i.e., a location error of <1 mm) (18). By applying one or two radiofrequency ablations (500 kHz RF generator [RFG-3C, Radionics, Burlington, Massachusetts] in a temperature-controlled mode [70°C for up to 60 s]) to a known site (along the anterior wall) subsequent alignment of the endocardial depiction of the LV and the actual LV chamber was enabled. The precise region of necrotic tissue was located using the electromechanical assessment allocated (i.e., a concomitant reduction in LS and BEA). Subsequently, each reconstructed chamber was divided into transverse sections in accordance with the pathological slices, using both width and number of slices as indices to guarantee identification. In each reconstructed section, all points from the center of the infarct (+0.5 cm to either side) as portrayed in the BEA maps (specifying a color range of 0.7 mV to 7 mV) observed during the postmapping phase were selected and correlated with the extent of pathological transmurality.
Data are presented as the mean ± 1 standard deviation. Electrogram characteristics (amplitude and SR), impedance and LS were assessed in both infarcted and noninfarcted regions within the various groups of transmurality extent, and compared using one-way analysis of variance and subsequent Bonferroni multiple comparisons analysis. Correlation between the electrical indices and mechanical function was performed using Pearson bivariate correlation. A value of p < 0.05 was considered statistically significant.
Electromechanical properties of infarcted tissue
Myocardial infarction resulted in reductions in the electrical and mechanical endocardial properties of the tissue (Fig. 2). The UEA and BEA, SR and endocardial impedance were significantly reduced (p < 0.05) in infarcted tissue (12.5 ± 1.7 mV, 1.6 ± 0.8 mV, 1.1 ± 0.7 mV/ms, 169.9 ± 21.0 Ω, respectively) compared with noninfarcted tissue (16.7 ± 1.4 mV, 6.1 ± 0.9 mV, 2.9 ± 0.3 mV/ms, 218.1 ± 24.7 Ω, respectively). Concomitantly, as previously noted, a significant reduction in LS values when comparing viable with nonviable tissue was also observed (9.4 ± 0.9% vs. −2.4 ± 1.7%, p < 0.05).
Correlation of electrical and mechanical data indicated a significant linear correlation (p < 0.05) between LS and impedance (r = 0.90), BEA (r = 0.68) and SR (r = 0.86), but not to UEA (r = 0.37, p = NS).
Pathologic transmurality extent
Postmortem pathological assessment revealed all infarcts to be of a continuous nature (i.e., no patchy infarcts were obtained), thus enabling precise quantification of the infarcted region. The full spectrum of transmurality extent was recognized in the hearts studied. A total of 67 slices (5.1 ± 1.4 slices from each LV) in which pathological and histological verification of infarction was made were entered into the study. The mean transmurality extent observed was 45.2 ± 19.0%. Three groups of transmurality extent were compared: 1) <30%, 17 slices; 2) 31% to 60%, 31 slices; and 3) 61% to 100%, 19 slices.
A gradient of transmurality extent was observed in all hearts; the mass of infarcted tissue, the most extensive transmural percentage, was observed in the middle slices, while either apical or basal slices, or both, revealed nontransmural infarct extensions. Furthermore, of the 13 hearts examined, eight exhibited a nontransmural nature in all slices (i.e., infarcts <60% transmurality) and the remaining five were found to be of transmural extension, according to the pattern described above. Thus, in addition to the ability of electromechanical mapping to distinguish between infarcts of varying nature in different infarcted hearts, the ability of catheter-based mapping to differentiate between infarcts of various degrees of transmurality within the same chamber was an inherent part of the assessment performed.
The results of the histological examination of all 67 slices corresponded to the TTC staining results and verified the presence of infarcted tissue. Specifically, loss of myocytes and a dominance of collagen fibers as well as inflammatory cells and macrophages were observed in all infarcted hearts.
Comparison of pathology and endocardial electrical properties
A total of 142 ± 16 points were acquired in all maps (total mapping time: 61 ± 14 min), with 116 ± 11 points remaining after editing. From these maps, all points found within 0.5 cm of the center of the infarcted region were included in the analysis. Thus, a total of 263 points were selected for analysis (an average of 4.6 ± 2.4 points from each transverse section; range 3 to 12 points in each section). The results are summarized in Table 1and typical findings are shown in Figure 3. All indices were found to be significantly lower than those found within noninfarcted tissue. Bipolar electrogram amplitude (1.9 ± 0.7 mV, 1.4 ± 0.7 mV, 0.8 ± 0.4 mV in the three groups, respectively, p < 0.05 between each group) was the most effective parameter in its ability to delineate among all three infarct subgroups, despite the overlap in values between transmurality subgroups (Fig. 4). Unipolar SR could not differentiate between the intermediate (31% to 60%) and transmural (61% to 100%) infarct subgroups, whereas UEA and impedance values proved inferior in that they could not differentiate between the low (<30%) and intermediate transmurality subgroups.
This study presents data indicating that the extent of infarct transmurality is accurately assessed in vivo using electrical parameters derived from the endocardial surface of the LV chamber. Of these parameters, we have demonstrated that BEA is the most sensitive tool for differentiating between subendocardial and transmural infarcts.
One of the most important variables in determining the prognosis of patients with MI is the extent of LV dysfunction and infarct size (2,20–24).
Necrotic myocardial tissue is known to be associated with abnormal contraction, low-amplitude and fractionated endocardial electrograms (25). Similarly, endocardial impedance values have been shown to be reduced in infarcted tissue (19,26). Furthermore, our group (15,19)and others (16,17)have demonstrated that by spatially associating these endocardial indices, infarcted tissue could be identified and localized and its extent on the endocardial surface sized.
Nonetheless, to date such endocardial information has not been associated with the degree of infarct transmurality. In order to test the hypothesis that there is such correlation, electromechanical mapping was used to associate the relevant electrophysiological parameters with the precise pathological location from which they were obtained. Subsequently, we were able to demonstrate that the reduction in viable myocyte density in the infarcted region caused diminished electrogram amplitude, reduced slew rate and lowered impedance values. Although all of these values could delineate between subendocardial and more transmural infarcts, BEA could distinguish between the widest spectrum of transmurality in this study. Furthermore, this finding was statistically significant despite the observed overlap of BEA values between transmurality subgroups.
The most plausible explanation for lack of UEA sensitivity is the nature of the recording. Unipolar signals are affected by global activity, whereas the bipolar recording is believed to be a more accurate index of local electrical activity (27)and therefore is better in dintinguishing between the subendocardial subtypes.
Miller et al. (28)performed endocardial electrogram recordings before and after removal of subendocardial scar tissue in patients with anterior wall MI. Miller et al. (28)observed an increase in the mean amplitude of bipolar signals recorded, from 0.5 ± 0.8 mV to 1.0 ± 1.6 mV. Bipolar amplitude increased by >0.2 mV in 49% and <0.2 mV in 39% after resection of the recordings made. However, 12% of the recordings demonstrated a decrease >0.2 mV in amplitude.
These results indicate that the endocardial surface alone is a major contributor to both amplitude and characteristics of endocardial signals recorded from scar tissue. Despite the differences in methodology between the study by Miller et al. (28)and the one described here, the differences in BEA recordings from the various transmurality subgroups were >0.2 mV. This finding, and the observed decrease in bipolar amplitude following endocardial scar resection (28), indicate that endocardial electrical signal recordings may be generated by cells from deeper myocardial layers. If this is indeed the case, endocardial mapping is a valid method of infarct transmurality characterization.
Importance of transmural identification
Transmural infarcts have been associated with a higher frequency of coronary thrombi and plaque rupture, large infarct sizes and a high number of ventricular wall ruptures (29–31). Thus, the substantial amount of tissue damage occurring within this infarct subtype causes a general demise in cardiac function, leading to cardiogenic shock and left heart failure (29,32,33).
Patients with nontransmural infarcts have a higher degree of sudden cardiac death, which is believed to result from reperfusion and recurrent ischemic events bringing forth electrical instability, for example the presence of microreentrant pathways between the islands of necrotic tissue (34,35).
Recovery of myocardial function following revascularization has also been associated with the extent of transmural necrosis, as viability of the outer myocardial layers may prevent infarct expansion and attenuate LV remodeling, promote electrical stability in the areas adjacent to the infarct and decrease the incidence of ischemic events and mortality (7,8).
The belief in the prognostic importance of knowledge of infarct extent within the LV wall prompted development of imaging modalities to accurately define and assess infarct transmurality extent in vivo. The resolution of conventional two-dimensional echocardiography is limited and cannot estimate the size of any infarct >20% in thickness (3). Kemper et al. (12), by using contrast echocardiography, were able to predict the extent of transmurality by analysis of delayed-contrast enhancement (15–20 diastoles) 2 h after acute coronary occlusion. However, beyond the limitations that are inherent to all echocardiographic studies, use of a contrast medium is subjected to artifacts due to acoustic shadowing, lateral wall defects and low contrast-border resolution (12).
Assessment with magnetic resonance imaging uses criteria such as end-diastolic wall thickness and lack of systolic wall thickening to identify scar tissue (13,36). However, like all other imaging modalities, which depend on mechanical function alone to assess myocardial tissue viability, overestimation occurs because of inclusion of regions subject to other myocardial pathologies such as chronic ischemia or stunned and hibernating tissues.
This limitation may not hold for the electromechanical assessment performed throughout this study because it is proposed that the degree of tissue damage may be ascertained by coupling the electrical and mechanical information obtained (15–17). Hence, this technique differs from others in that concomitant electromechanical assessment is used to differentiate between necrosis and ischemic, hibernating or stunned tissue in which mechanical dysfunction occurs but electrical disturbances are less pronounced.
The results of the current study indicate a concomitant impairment of both electrical and mechanical function in infarcted tissue. Interestingly, SR, BEA and endocardial impedance all correlated significantly with the reduction in LS.
Importance and clinical relevance of findings
Knowledge of the degree to which the necrotic wavefront extends within the LV wall is of importance as there are significant prognostic implications in the ability to predict the extent of disease and the benefit of revascularization procedures (7,8). The unique property of such a mapping technique to register information while enabling real-time interpretation and application of therapeutic strategies increases the significance of our findings. Currently, much attention is being directed at derivation of localized therapeutic procedures. Although determination of the beneficial effects of such an approach is in the initial stages, knowledge of the transmural nature of the infarcted tissue may aid in determining exactly where such applications as angiogenesis (37)are to be applied.
This study was performed using a dedicated animal model of LAD ligation in the canine producing a uniform substrate of infarction. Thus, the first limitation of this study would be the lack of variability, which is observed in the clinical setting. Moreover, morphologic subendocardial and transmural infarcts are of discrete pathogenesis and transmural infarcts could not be produced in the model employed. Thus, endocardial signals in human patients must be analyzed to fully assess the manner by which infarct extent and pathogenesis are responsible for observed variations.
This study was aimed at defining local electrical criteria for assessing infarct transmurality. Therefore, averaging of point values was performed at the core of the infarct within each section. The number of sites sampled within a given area is critical to the accuracy of such an analysis. Thus, mapping was performed in a detailed fashion, enabling the analysis of 263 points within the center of the infarct. However, points were not evenly distributed throughout the sections analyzed. This is a potential in BEA analysis, as these recordings may be influenced by electrode/fiber orientation. Nevertheless, in most sections examined, a relatively high number of points (range 3 to 12) were selected for analysis, “averaging out” the potential for such false positive recordings. In addition, with infarction the affected endocardium is frequently smooth because of loss of viable myocytes, thereby reducing the possibility for electrode/fiber orientation effects on bipolar electrogram morphology.
Finally, the measurements performed throughout this study assume that the central region of the infarcted zone is not affected by the electrical currents that still reside within the border and healthy regions. Future studies assessing this complex relationship are needed to fully understand the relation between the electrical properties of infarcted and adjacent healthy tissues. Furthermore, the correlation of readings obtained during electromechanical mapping with the widely used conventional body surface electrocardiogram may shed more light into theories linking conventional electrocardiograms with the extent of transmurality of infarcted tissues.
From our research to correlate endocardial potentials with infarct transmurality, we conclude that interpretation of the electrophysiologic properties of the endocardium may be used in vivo to assess the depth of tissue necrosis. These results may have important research and clinical implications for patients with ischemic heart disease.
The authors wish to acknowledge Raymond Coleman, Edith Cohen and Armin Schwartzman for their valuable assistance throughout the study. We also wish to thank Deborah Shapiro for her editorial assistance.
☆ This work was supported by a grant from Biosense Webster (Israel) Ltd.
- bipolar electrogram amplitude
- left anterior descending
- local shortening
- left ventricle
- myocardial infarction
- slew rate
- 2,3,5-triphenyltetrazolium chloride
- unipolar electrogram amplitude
- Received October 3, 2000.
- Revision received December 28, 2000.
- Accepted January 24, 2001.
- American College of Cardiology
- Sobel B.E,
- Breshnahan G.F,
- Shell W.E,
- Yoder R.D
- Lieberman A.N,
- Weiss J.L,
- Jugdutt B.I,
- et al.
- Delbeke D,
- Lorenz C.H,
- Votaw J.R,
- et al.
- Kaul S
- ↵Agati L, Autore C, Iacoboni C, et al. The complex relation between myocardial viability and functional recovery in chronic left ventricular dysfunction. Am J Cardiol 1998;81:33–5G.
- Phibbs B
- Spodick D.H
- Kemper A.J,
- Force T,
- Perkins L,
- Gilfoil M,
- Parisi A.F
- Geskin G,
- Kramer C.M,
- Rogers W,
- et al.
- Kornowski R,
- Hong M.K,
- Gepstein L,
- et al.
- Gepstein L,
- Goldin A,
- Lessick J,
- et al.
- Callans D.J,
- Ren J.F,
- Michele J,
- Marchlinski F.E,
- Dillon S.M
- Gepstein L,
- Hayam G,
- Ben-Haim S.A
- Wolf T,
- Gepstein L,
- Hayam G,
- et al.
- Gardner P.I,
- Ursell P.C,
- Fenoglio J.J,
- Wit A.L
- Miller J.M,
- Tyson G.S,
- Hargrove W.C,
- et al.
- Erhardt L.R
- Alonso D.R,
- Sceidt S,
- Post M,
- Killip T
- Kornowski R,
- Leon M.B,
- Fuchs S,
- et al.