- Email: [email protected]
Mechanisms of ischemia-induced ST-segment changes
Journal of Electrocardiology 38 (2005) 8 – 13 www.elsevier.com/locate/jelectrocard
Mechanisms of ischemia-induced ST-segment changes Robert S. MacLeoda,b,T, Shibaji Shomea,b, Jeroen Stinstraa,b, Bonnie B. Punskea,b, Bruce Hopenfeldc a Bioengineering Department, University of Utah, Salt Lake City, Utah 84112-5000, USA Nora Eccles Harrison Cardiovascular Research and Training Institute, University of Utah, Salt Lake City, Utah 84112-5000, USA c National Heart, Lung, and Blood Institute, National Institutes of Health, USA Received 10 June 2005; accepted 10 June 2005
b
Abstract
Many aspects of ischemia-induced changes in the electrocardiogram lack solid biophysical underpinnings although variations in ST segments form the predominant basis for diagnostic and monitoring of patients. This incomplete knowledge certainly plays a role in the poor performance of some forms of electrocardiogram-based detection and characterization of ischemia, especially when it is limited to the subendocardium. The focus of our recent studies has been to develop a comprehensive mechanistic model of the electrocardiographic effects of ischemia. The computational component of this model is based on highly realistic heart geometry with anisotropic fiber structure and allows us to assign ischemic action potentials to contiguous regions that can span a prescribed thickness of the ventricles. A separate, high-resolution model of myocardial tissue provides us with a means of setting electrical characteristics of the heart, including the status of gap junctional coupling between cells. The experimental counterpart of this model consists of dog hearts, either in situ or isolated and perfused with blood, in which we control coronary blood flow by means of a cannula and blood pump. By reducing blood flow through the cannula for various durations, we can replicate any phase of ischemia from hyper acute to early infarction. Based on the results of these models, there is emerging a mechanism of the electrocardiographic response to ischemia that depends strongly on the anisotropic conductivity of the myocardium. Ischemic injury currents flow across the boundary between healthy and ischemic tissue, but it is their interaction with local fiber orientation and the associated conductivity that generates secondary currents that determine epicardial ST-segment potentials. Results from experiments support qualitatively the findings of the simulations and underscore the role of myocardial anisotropy in electrocardiography. D 2005 Elsevier Inc. All rights reserved.
Keywords:
Myocardial ischemia; Simulation; Computer modeling
1. Introduction We report a combination of computer simulations and experiments, the overall goal of which is to understand the origins and consequences of injury currents that arise during myocardial ischemia. The entire project consisted of 3 major components: a model of myocardial tissue with which we can calculate anisotropic electrical conductivity under a range of conditions that arise in ischemia; a whole heart
T Corresponding author. Nora Eccles Harrison CVRTI, University of Utah, Salt Lake City, Utah 84112-5000, USA. Tel.: +1 801 587 9511; fax: +1 801 587 9511. E-mail address: [email protected] (R.S. MacLeod). 0022-0736/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jelectrocard.2005.06.095
model in which we can place ischemic regions of any size and location, and compute cardiac potentials; and elaborate experimental preparations that permit us to measure cardiac potentials during controlled episodes of ischemia of variable extent and duration under a range of pacing and perfusion conditions. We have reported on each component separately in the past [1-6] and here focus on what the combination of simulation and experimentation suggests may be [1] a mechanism for subendocardial ischemia over a range of conditions. The motivation for such a study has many origins, starting with the pioneering studies of Samson and Scher [7], Holland and Brooks [8,9], and Holland and Arnsdorf [10] into the biophysical source of changes in electrocardiogram
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
morphology during ischemia. These investigators sought to explain the occurrence of currents of injury that many others had observed decades before [11,12], and their findings addressed concerns by investigators such as Binetti et al [13], who suggested that epicardial potentials could underestimate infarction size and thus mislead investigators in the study of pharmacologic interventions for the ability to reduce infarct size in dogs [13]. Conflicting interpretations of ST-segment changes as a marker of extent of ischemia in humans raged during the 1970s [14,15], and in the following decades, there has been an acceptance that the ST segment measured in the standard electrocardiogram was not a suitable quantitative marker of the extent or degree of ischemia [16,17]. Studies such as those by Kle´ber et al [18,19] and Smith et al [20,21], making use of both cellular and direct cardiac mapping, have greatly expanded understanding of the electrocardiographic consequence of acute ischemia. Kle´ber and his coworkers provided many details regarding the mechanisms for changes in action potential morphology and for the generation of intramyocardial gradients of potential that ultimately drive the generation of injury currents. Smith and his colleagues not only provided detailed maps of epicardial potentials over regions of heart in which they induced variable amounts of ischemia but also created quantitative models for predicting epicardial potentials. By comparing predictions from earlier models based on solid angle formulations by Holland and Brooks [9] with their own models and measured potentials, they demonstrated a model that generated results consistent with experimental data. Essential to the model was a volumetric description of the potential distribution, in contrast to the surface-based description of Holland and Brooks [9]. Smith et al [20,21] also relaxed the assumption that potential was uniform across the ischemic zone. The use of body surface mapping and inverse solution approaches opened anew the discussion of using STsegment potentials to monitor ischemia under the assumption that, with comprehensive or at least additional information, it may be possible to detect, localize, and even quantify the ischemic region that arises in acute ischemia. The focus also shifted from ischemia that was fully transmural to the more common case of subendocardial ischemia. Specific diagnosis and localization of ischemia that involves only a portion of the ventricular wall has proven particularly challenging and is based on ST-segment depression rather than elevation near the affected region. A major study that made use of a geometrically realistic model was by Li et al [22], which described a mathematical model and associated experiments using the sheep heart in which the investigators produced subendocardial ischemia and measured and computed the associated epicardial potentials. Their results suggested that subendocardial ischemia arising from occlusions of both the left anterior descending (LAD) and circumflex arteries produced ambiguous epicardial
9
ST-segment depressions, although endocardial potentials were distinct for the 2 cases. They found in both experiments and simulations ST-segment depression in the epicardium over the border zone between healthy and ischemic tissue. The models that Li et al [22] used did not, however, include anisotropic conductivity and thus lacked an important component of actual heart tissue. Subsequent studies by Johnston and Kilpatrick [23] have showed that anisotropic conductivity does, indeed, alter the epicardial potentials resulting from subendocardial ischemia, although their model used simplified geometry representing a segment of the left ventricle and a rectilinear slab. Thus, the stage was set for a study that incorporated highly detailed, anisotropic tissue and heart models, and supporting experiments during ischemic episodes. The goals were not just to examine the findings of Li et al [22] but to develop a more basic understanding of the nature of ischemia injury currents and the potentials they produce during the ST segment, as Smith et al [20,21] did from their studies. We felt that tissue anisotropy might play an important role in the underlying mechanisms, given the results of Johnston and Kilpatrick [23]. In addition, however, we felt that it was imperative to have a highrealistic model of the whole heart in which to carry out detailed simulations using advanced numerical approaches. We also wished to investigate the response of the heart to the full range of conditions that follow ischemia, including the collapse of capillaries, the shift in interstitial water, and the eventual closure of gap junctions, all of which change tissue electrical characteristics and thus, potentially, the resulting electrocardiographic fields. Our studies, which are still ongoing, have revealed that anisotropic conductivity not only does, indeed, play a role in ST-segment potentials but also suggests a fundamental mechanism for the role of injury currents that arises only under anisotropic conditions. The mechanism we postulate leads to ST-segment potentials that originate not directly from the currents that cross the boundary between ischemic and healthy tissue but because of potential differences that arise when these currents are oriented differently with respect to local fiber direction at different edges of the ischemic zone. We have also studied in experiments and simulations the effect of changes in cell-to-cell coupling that are known to follow prolonged exposure to ischemia as gap junctions begin to close.
2. Methods 2.1. Myocardial model The basis of the simulation model was a fully isotropic model of the heart based on the Auckland canine geometry [24] that we resampled to facilitate the inclusion of regions of ischemia of variable location and transmural extent, as we have published previously [2]. Briefly, the model generation
10
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Fig. 1. Experimental findings and simulations of increasing transmural extent of ischemia. Upper panel contains ST-segment potential maps from an experiment with progressive reduction in flow through the LAD cannula. The lower panel contains the (approximately) equivalent simulation results with increasing transmural extent of the ischemic zone.
began with fitting to the nodes of the epicardial and endocardial surfaces analytic surfaces based on spherical harmonic basis functions, which allowed a reorganization of the geometry into concentric shells. A straightforward rule–based meshing scheme then allowed us to generate specific geometric models at any level of spatial density; however, the resulting model volume elements had curved surfaces, which required some additional complexity in the model description. The other significant advantage of this scheme was that it allowed very simple characterization of an ischemic region of prescribed circumferential extent that could be very simply extended from the endocardial surface, the innermost shell, to the epicardium, the outermost shell, or any transmural extent between. The most accurate and detailed approach to simulating electric potentials in the heart that is still computationally tractable is the bidomain [25], and for this, one needs not only a geometric model of volume elements but also values for the electrical conductivity both along and across the local fiber direction in the intracellular and extracellular space, 4 separate values. Values in the literature of myocardial conductivity vary quite widely [26-29], and to resolve these differences, we created a high-resolution model of myocardium that included realistic cell shape and connectivity, including a network of capillaries [1,3]. The model provided a means of computing conductivity of tissue under ischemic conditions of reduced capillary flow and varying levels of gap junction opening. 2.2. Simulations The bidomain approach allows simulation of cardiac potentials but is computationally expensive when used to calculate complete activation sequences over full heart geometries. To reduce the computational cost, we simulated cardiac potentials under the simplifying conditions of a fixed
potential difference between healthy and ischemic tissue. Smith and others [20] have suggested that this need not be the case, but there do not exist robust data describing more realistic distributions. Moreover, it is not likely that the fundamental findings from our simulations would differ significantly. We were able to take advantage of the increased efficiency of this approach to simulate a wide array of ischemic conditions and thus evaluate the effect of different transmural extents of ischemia, as well as the changes that arise under varying degrees of cell-to-cell coupling. 2.3. Experiments Experimental data for these studies came from 2 different preparations of canine hearts that we have described elsewhere in detail [30-32]. In the in situ case, the heart remained in the animal, exposed through a midsternal opening, and suspended in a pericardial cradle. In the isolated heart, the preparation consisted of an isolated dog heart suspended in a torso-shaped electrolytic tank. In both in situ and isolated heart cases, we controlled coronary perfusion either by means of a snare encircling the LAD artery or by cannulating the LAD artery and perfusing it by means of a calibrating precision pump. We alternated episodes of normal coronary flow with various levels of reduced flow, either stepwise or complete, to simulate a range of different forms of acute ischemia. Regular monitoring of blood gases ensured that perfusion was adequate and pH was normal between interventions. In each of preparations, we recorded epicardial potentials simultaneously from 490 epicardial sock electrodes contained in a flexible array attached to a nylon stocking fitted over the ventricles. We recorded all channels from the sock electrode using our custom built acquisition system, which permitted simultaneous recording of up to 1024
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
11
Fig. 2. Experimental findings and simulations of increasing duration of ischemia. Upper panel contains ST-segment potential maps from an experiment with prolonged ischemia under 10% of normal flow rates in the cannulated LAD artery. Values in minutes indicate the duration of the reduced flow from onset. The lower panel contains the simulation results with different values of gap junction resistance to mimic early (left map) and late (right map) durations of ischemia.
channels at a sampling rate of 1000 Hz with adjustable gain and each channel buffered by sample-and-hold circuitry, as described previously [32,33]. 3. Results We present here 2 representative results for which we have qualitative agreement through experiments. In the first, in Fig. 1, we show the effect of transmural extent of ischemia. In the experiment, this results from progressively reducing the flow through the cannula from 50 of normal flow to 25 and then 1, with 6 minutes of time between each of the measurements shown in the figure. The equivalent simulation result, shown in the lower panel, was to compute the epicardial potentials for subendocardial ischemia that ranged from 40 through 70 to 90 of the transmural extent. Note the rotation of the axis drawn in each panel that joins the peripheral minima through the central elevation of
Fig. 3. Proposed mechanism of epicardial potentials from subendocardial ischemia. The diagram shows schematically the ischemic and healthy zones, and the extracellular potential difference that arises at the border between the two. The size of plus and minus signs indicates the relative amplitude of the potential difference, which differs at lateral and transmural boundaries. Also shown is the assumed local fiber orientation.
potential. We will outline the significance of this finding in the Discussion section. Fig. 2 contains results from an experiment in which we performed a prolonged reduction in LAD arterial flow and approximately equivalent simulations. The upper panel of the figure presents an overview the ST-segment potentials after reduction in LAD flow to 10 of normal rates. One can clearly see the transition from control levels through a peak level of ST-segment elevation after 9 minutes of occlusion, followed by a gradual reduction of elevation almost to control levels after 24 minutes. The lower panel contains 2 ST-segment potential maps simulated at 70 transmural subendocardial ischemia with 2 different values of gap junctional resistance to also mimic early and late phases of ischemia. 4. Discussion The results of these simulations and experiments are still preliminary, yet they already suggest both qualitative agreement between the computational model and experiments, as well as mechanisms for ischemic injury current. By studying the potentials computed throughout the heart in the simulations, we have developed a mechanistic model of injury current that can explain both our findings and those of others in previous studies, the details of which are available elsewhere [2,6]. The key component of this proposed mechanism, shown in Fig. 3, is that although the intracellular potential difference across the border zone can be uniform, we assume 1 transmembrane potential for all the healthy cells and a second, uniform transmembrane potential for the entire ischemic zone; the resulting extracellular potentials depend on local tissue conductivity. Thus, as the local conductivity changes, as we know it does in an anisotropic model of the myocardium, then so too must the extracellular
12
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Fig. 4. Postmortem cardiac imaging of experimental hearts. The left-hand panel contains a photo of the heart used for the MRI and the right-hand panel a volume-rendered reconstruction of the anatomy and the markers based on volume rendering the set of magnetic resonance images. The photo shows as well the site of occlusion in the LAD artery.
potential. In the specific case in Fig. 3, we have assumed circumferentially oriented fibers so that the relationship between fiber orientation and ischemic border orientation is quite different between the lateral borders and the transmural border. As a result, so too are there potential differences at the respective boundaries. With these potential differences arising between lateral and transmural ischemic borders, there will be current flow within the heart, which will lead to potentials on the epicardial surface, as indicated in Fig. 3. In the configuration shown, the epicardial potential over the ischemic zone would appear positive and that over the lateral border zones would be relatively negative. On the body surface, one would then see ST-segment elevation and depression, which would appear at locations dictated by the locations of the heart within the torso and the ischemic region within the heart. Probably, the greatest challenge in performing ischemia experiments is establishing the degree of perfusion in the tissues to identity the regions most likely to experience ischemia. By reducing blood flow to a major coronary artery and observing reproducible and characteristic changes in epicardial potentials, we feel confident that we are, indeed, creating ischemic regions; their location and transmural extent, on the other hand, are still relatively poorly documentable. A related source of uncertainty is that the geometric models for the simulations came from a single geometry, the Auckland canine model [24], and thus can
agree only qualitatively with any data we obtained from actual experiments. There is every reason to expect that changes in cardiac geometry, even with the same ischemic regions included, will lead to different epicardial potentials. We have therefore begun to use magnetic resonance imaging (MRI) to acquire high-resolution scans of the hearts in our studies, with the goal of creating realistic geometric models, ideally in which we have also marked the perfusion bed of the cannulated vessel. Fig. 4 shows a photograph (left-hand panel) and a volume rendering (right-hand panel) of the MRI data from the same sample heart. Also shown in both panels are the reconstructions of the MRI contrast markers that we used to align geometries; we will attempt to use the same agent to mark the perfusion beds of the LAD in future experiments. At present, we are creating geometric models from this heart/torso tank geometry with which to carry out further studies. In future experiments, we will infuse magnetic resonance contrast agents into the coronary circulation and attempt to image them and thus the perfusion bed for a particular heart. Enhanced authenticity of the geometric model can only help in comparing computations with experiments and thus in refining our simulations and the mechanisms such simulations suggest. Acknowledgments Support for this research has come from awards from the Nora Eccles Treadwell Foundation, the Richard A and Nora
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Eccles Harrison Fund for Cardiovascular Research (San Leandro, CA), and the National Institutes of Health NCRR National Center for Research Resources for the Center for Bioelectric Field Modeling, Simulation, and Visualization (HL P41RR12553). The authors express their gratitude to Matt Allison, Yonild Lian, Patti Larrabee, and Jayne Davis for expert technical assistance in the experimental studies, to Ted Dustman for development of analysis software, and to Phil Ershler and Bruce Steadman for the acquisition system. We also thank Drs Elliot McVeigh and Dan Ennis from the National Heart, Lung, and Blood Institute and Stanford University, respectively, for their generous assistance in generating magnetic resonance images of the heart.
References [1] Stinstra JG, Hopenfeld B, MacLeod RS. A model for passive cardiac conductivity. Int J Bioelectromagnetism 2003;5(1):185. [2] Hopenfeld B, Stinstra JG, MacLeod RS. Mechanism for ST depression associated with contiguous subendocardial ischemia. J Cardiovasc Electrophysiol 2004;15(10):1200. [3] Stinstra JG, Hopenfeld B, MacLeod RS. Using models of the passive cardiac conductivity and full heart anisotropic bidomain to study the epicardial potentials in ischemia. Proceedings of the IEEE Engineering in Medicine and Biology Society 26th Annual International Conference: IEEE Press; 2004. [4] Shome S, Stinstra JG, Hopenfeld B, et al. A study of the dynamics of cardiac ischemia using experimental and modeling approaches. Proceedings of the IEEE Engineering in Medicine and Biology Society 26th Annual International Conference: IEEE Press; 2004. [5] Shome S, Rossi S, Punske B, et al. Anisotropic epicardial ST potential patterns reveal location and depth of ventricular necroses. Proceedings of the BMES 2004 Annual Fall Meeting: BMES; 2004. [6] Hopenfeld B, Stinstra JG, MacLeod RS. The effect of conductivity on ST segment epicardial potentials arising from subendocardial ischemia. Annal Biomed Eng 2005;33(6):751. [7] Samson WE, Scher AM. Mechanism of ST-segment alteration during acute myocardial injury. Circ Res 1960;8:780. [8] Holland RP, Brooks H. The QRS complex during myocardial ischemia: an experimental analysis in the porcine heart. J Clin Invest 1976;57:541. [9] Holland RP, Brooks H. TQ-ST segment mapping: critical review and analysis of current concepts. Am J Cardiol 1977;4:110. [10] Holland RP, Arnsdorf MF. Solid angle theory and the electrocardiogram: physiologic and quantitative interpretations. Prog Cardiovasc Dis 1977;19(6):431. [11] Eyster JAE, Meek WJ, Goldberg H, et al. Potential changes in an injured region of the cardiac heart. Am J Physiol 1938;717:717. [12] Nahum LH, Hamilton WF, Heff HE. Injury current in the electrocardiogram. Am J Physiol 1943;139:202.
13
[13] Binetti G, Markiewicz W, Billingham M, et al. Problems in assessing infarction size by epicardial mapping: preliminary studies with quinidine. Eur J Cardiol 1976;4(4):429. [14] Fozzard HA, DasGupta DS. ST-segment potentials and mapping: theory and experiments. Circulation 1976;54:533. [15] Moroko PR, Libby P, Covell JW, et al. Precordial ST-segment elevation mapping: an atraumatic method for assessing alterations in the extent of myocardial ischemic injury. Am J Cardiol 1972;2:223. [16] Fozzard HA. ST-segment mapping is not a clinical tool. In: Rapport, RS, editor. Current controversies in cardiovascular disease. Saunders; 1980. p. 281. [17] Fozzard HA, Makielski JC. The electrophysiology of acute myocardial ischemia. Annu Rev Med 1985;3:275. [18] Kle´ber AG. Extracellular potassium accumulation in acute myocardial ischemia. J Mol Cell Cardiol 1984;16:389. [19] Kle´ber AG, Janse MJ, Wilms-Schopmann FJG, et al. Changes in conduction velocity during acute ischemia in ventricular myocardium of the isolated porcine heart. Circulation 1986;73:189. [20] Smith GT, Geary G, Ruf W, et al. Epicardial mapping and electrocardiographic models of myocardial ischemic injury. Circ Res 1979;60:930. [21] Smith GT, Geary GG, Blanchard W, et al. An electrocardiographic model of myocardial ischemia injury. J Electrocardiol 1983;16(3):223. [22] Li D, Li CY, Yong AC, et al. Source of electrocardiographic ST changes in subendocardial ischemia. Circ Res 1998;82:957. [23] Johnston PR, Kilpatrick D. The effect of conductivity values on ST segment shift in subendocardial ischaemia. IEEE Trans Biomed Eng 2003;50(2):150. [24] Nielsen PMF, Grice IJL, Smaill BH, et al. Mathematical model of geometry and fibrous structure of the heart. Am J Physiol 1991; 260:H1365. [25] Henriquez CS. Simulating the electrical behavior of cardiac tissue using the bidomain model. Crit Rev Biomed Eng 1993;21(1):1. [26] Clerc L. Directional differences of impulse spread in trabecular muscle from mammalian heart. J Physiol 1976;255:335. [27] Kle´ber AG, Janse MJ, Capelle FJLv, et al. Mechanism and time course of ST- and TQ-segment changes during acute regional myocardial ischemia in the pig heart determined by extracellular and intracellular recordings. Circ Res 1978;42:603. [28] Roberts DE, Hersch LT, Scher AM. Influence of cardiac fiber orientation on wavefront voltage, conduction velocity and tissue resistivity. Circ Res 1979;44:701. [29] Roberts DE, Scher AM. Effect of tissue anisotropy on extracellular potential fields in canine myocardium in situ. Circ Res 1982;50:342. [30] MacLeod RS, Punske B, Shome S, et al. The role of heart rate and coronary flow during myocardial ischemia. J Electrocardiol 2001:43. [31] MacLeod RS, Miller RM, Gardner MJ, et al. Application of an electrocardiographic inverse solution to localize myocardial ischemia during percutaneous transluminal coronary angioplasty. J Cardiovasc Electrophysiol 1995;6:2. [32] MacLeod RS, Ni Q, Punske B, et al. Effects of heart position on the body-surface ECG. J Electrocardiol 2000;33(Suppl):229. [33] MacLeod RS, Lux RL, Taccardi B. A possible mechanism for electrocardiographically silent changes in cardiac repolarization. J Electrocardiol 1997;30(Suppl):114.
Mechanisms of ischemia-induced ST-segment changes Robert S. MacLeoda,b,T, Shibaji Shomea,b, Jeroen Stinstraa,b, Bonnie B. Punskea,b, Bruce Hopenfeldc a Bioengineering Department, University of Utah, Salt Lake City, Utah 84112-5000, USA Nora Eccles Harrison Cardiovascular Research and Training Institute, University of Utah, Salt Lake City, Utah 84112-5000, USA c National Heart, Lung, and Blood Institute, National Institutes of Health, USA Received 10 June 2005; accepted 10 June 2005
b
Abstract
Many aspects of ischemia-induced changes in the electrocardiogram lack solid biophysical underpinnings although variations in ST segments form the predominant basis for diagnostic and monitoring of patients. This incomplete knowledge certainly plays a role in the poor performance of some forms of electrocardiogram-based detection and characterization of ischemia, especially when it is limited to the subendocardium. The focus of our recent studies has been to develop a comprehensive mechanistic model of the electrocardiographic effects of ischemia. The computational component of this model is based on highly realistic heart geometry with anisotropic fiber structure and allows us to assign ischemic action potentials to contiguous regions that can span a prescribed thickness of the ventricles. A separate, high-resolution model of myocardial tissue provides us with a means of setting electrical characteristics of the heart, including the status of gap junctional coupling between cells. The experimental counterpart of this model consists of dog hearts, either in situ or isolated and perfused with blood, in which we control coronary blood flow by means of a cannula and blood pump. By reducing blood flow through the cannula for various durations, we can replicate any phase of ischemia from hyper acute to early infarction. Based on the results of these models, there is emerging a mechanism of the electrocardiographic response to ischemia that depends strongly on the anisotropic conductivity of the myocardium. Ischemic injury currents flow across the boundary between healthy and ischemic tissue, but it is their interaction with local fiber orientation and the associated conductivity that generates secondary currents that determine epicardial ST-segment potentials. Results from experiments support qualitatively the findings of the simulations and underscore the role of myocardial anisotropy in electrocardiography. D 2005 Elsevier Inc. All rights reserved.
Keywords:
Myocardial ischemia; Simulation; Computer modeling
1. Introduction We report a combination of computer simulations and experiments, the overall goal of which is to understand the origins and consequences of injury currents that arise during myocardial ischemia. The entire project consisted of 3 major components: a model of myocardial tissue with which we can calculate anisotropic electrical conductivity under a range of conditions that arise in ischemia; a whole heart
T Corresponding author. Nora Eccles Harrison CVRTI, University of Utah, Salt Lake City, Utah 84112-5000, USA. Tel.: +1 801 587 9511; fax: +1 801 587 9511. E-mail address: [email protected] (R.S. MacLeod). 0022-0736/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jelectrocard.2005.06.095
model in which we can place ischemic regions of any size and location, and compute cardiac potentials; and elaborate experimental preparations that permit us to measure cardiac potentials during controlled episodes of ischemia of variable extent and duration under a range of pacing and perfusion conditions. We have reported on each component separately in the past [1-6] and here focus on what the combination of simulation and experimentation suggests may be [1] a mechanism for subendocardial ischemia over a range of conditions. The motivation for such a study has many origins, starting with the pioneering studies of Samson and Scher [7], Holland and Brooks [8,9], and Holland and Arnsdorf [10] into the biophysical source of changes in electrocardiogram
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
morphology during ischemia. These investigators sought to explain the occurrence of currents of injury that many others had observed decades before [11,12], and their findings addressed concerns by investigators such as Binetti et al [13], who suggested that epicardial potentials could underestimate infarction size and thus mislead investigators in the study of pharmacologic interventions for the ability to reduce infarct size in dogs [13]. Conflicting interpretations of ST-segment changes as a marker of extent of ischemia in humans raged during the 1970s [14,15], and in the following decades, there has been an acceptance that the ST segment measured in the standard electrocardiogram was not a suitable quantitative marker of the extent or degree of ischemia [16,17]. Studies such as those by Kle´ber et al [18,19] and Smith et al [20,21], making use of both cellular and direct cardiac mapping, have greatly expanded understanding of the electrocardiographic consequence of acute ischemia. Kle´ber and his coworkers provided many details regarding the mechanisms for changes in action potential morphology and for the generation of intramyocardial gradients of potential that ultimately drive the generation of injury currents. Smith and his colleagues not only provided detailed maps of epicardial potentials over regions of heart in which they induced variable amounts of ischemia but also created quantitative models for predicting epicardial potentials. By comparing predictions from earlier models based on solid angle formulations by Holland and Brooks [9] with their own models and measured potentials, they demonstrated a model that generated results consistent with experimental data. Essential to the model was a volumetric description of the potential distribution, in contrast to the surface-based description of Holland and Brooks [9]. Smith et al [20,21] also relaxed the assumption that potential was uniform across the ischemic zone. The use of body surface mapping and inverse solution approaches opened anew the discussion of using STsegment potentials to monitor ischemia under the assumption that, with comprehensive or at least additional information, it may be possible to detect, localize, and even quantify the ischemic region that arises in acute ischemia. The focus also shifted from ischemia that was fully transmural to the more common case of subendocardial ischemia. Specific diagnosis and localization of ischemia that involves only a portion of the ventricular wall has proven particularly challenging and is based on ST-segment depression rather than elevation near the affected region. A major study that made use of a geometrically realistic model was by Li et al [22], which described a mathematical model and associated experiments using the sheep heart in which the investigators produced subendocardial ischemia and measured and computed the associated epicardial potentials. Their results suggested that subendocardial ischemia arising from occlusions of both the left anterior descending (LAD) and circumflex arteries produced ambiguous epicardial
9
ST-segment depressions, although endocardial potentials were distinct for the 2 cases. They found in both experiments and simulations ST-segment depression in the epicardium over the border zone between healthy and ischemic tissue. The models that Li et al [22] used did not, however, include anisotropic conductivity and thus lacked an important component of actual heart tissue. Subsequent studies by Johnston and Kilpatrick [23] have showed that anisotropic conductivity does, indeed, alter the epicardial potentials resulting from subendocardial ischemia, although their model used simplified geometry representing a segment of the left ventricle and a rectilinear slab. Thus, the stage was set for a study that incorporated highly detailed, anisotropic tissue and heart models, and supporting experiments during ischemic episodes. The goals were not just to examine the findings of Li et al [22] but to develop a more basic understanding of the nature of ischemia injury currents and the potentials they produce during the ST segment, as Smith et al [20,21] did from their studies. We felt that tissue anisotropy might play an important role in the underlying mechanisms, given the results of Johnston and Kilpatrick [23]. In addition, however, we felt that it was imperative to have a highrealistic model of the whole heart in which to carry out detailed simulations using advanced numerical approaches. We also wished to investigate the response of the heart to the full range of conditions that follow ischemia, including the collapse of capillaries, the shift in interstitial water, and the eventual closure of gap junctions, all of which change tissue electrical characteristics and thus, potentially, the resulting electrocardiographic fields. Our studies, which are still ongoing, have revealed that anisotropic conductivity not only does, indeed, play a role in ST-segment potentials but also suggests a fundamental mechanism for the role of injury currents that arises only under anisotropic conditions. The mechanism we postulate leads to ST-segment potentials that originate not directly from the currents that cross the boundary between ischemic and healthy tissue but because of potential differences that arise when these currents are oriented differently with respect to local fiber direction at different edges of the ischemic zone. We have also studied in experiments and simulations the effect of changes in cell-to-cell coupling that are known to follow prolonged exposure to ischemia as gap junctions begin to close.
2. Methods 2.1. Myocardial model The basis of the simulation model was a fully isotropic model of the heart based on the Auckland canine geometry [24] that we resampled to facilitate the inclusion of regions of ischemia of variable location and transmural extent, as we have published previously [2]. Briefly, the model generation
10
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Fig. 1. Experimental findings and simulations of increasing transmural extent of ischemia. Upper panel contains ST-segment potential maps from an experiment with progressive reduction in flow through the LAD cannula. The lower panel contains the (approximately) equivalent simulation results with increasing transmural extent of the ischemic zone.
began with fitting to the nodes of the epicardial and endocardial surfaces analytic surfaces based on spherical harmonic basis functions, which allowed a reorganization of the geometry into concentric shells. A straightforward rule–based meshing scheme then allowed us to generate specific geometric models at any level of spatial density; however, the resulting model volume elements had curved surfaces, which required some additional complexity in the model description. The other significant advantage of this scheme was that it allowed very simple characterization of an ischemic region of prescribed circumferential extent that could be very simply extended from the endocardial surface, the innermost shell, to the epicardium, the outermost shell, or any transmural extent between. The most accurate and detailed approach to simulating electric potentials in the heart that is still computationally tractable is the bidomain [25], and for this, one needs not only a geometric model of volume elements but also values for the electrical conductivity both along and across the local fiber direction in the intracellular and extracellular space, 4 separate values. Values in the literature of myocardial conductivity vary quite widely [26-29], and to resolve these differences, we created a high-resolution model of myocardium that included realistic cell shape and connectivity, including a network of capillaries [1,3]. The model provided a means of computing conductivity of tissue under ischemic conditions of reduced capillary flow and varying levels of gap junction opening. 2.2. Simulations The bidomain approach allows simulation of cardiac potentials but is computationally expensive when used to calculate complete activation sequences over full heart geometries. To reduce the computational cost, we simulated cardiac potentials under the simplifying conditions of a fixed
potential difference between healthy and ischemic tissue. Smith and others [20] have suggested that this need not be the case, but there do not exist robust data describing more realistic distributions. Moreover, it is not likely that the fundamental findings from our simulations would differ significantly. We were able to take advantage of the increased efficiency of this approach to simulate a wide array of ischemic conditions and thus evaluate the effect of different transmural extents of ischemia, as well as the changes that arise under varying degrees of cell-to-cell coupling. 2.3. Experiments Experimental data for these studies came from 2 different preparations of canine hearts that we have described elsewhere in detail [30-32]. In the in situ case, the heart remained in the animal, exposed through a midsternal opening, and suspended in a pericardial cradle. In the isolated heart, the preparation consisted of an isolated dog heart suspended in a torso-shaped electrolytic tank. In both in situ and isolated heart cases, we controlled coronary perfusion either by means of a snare encircling the LAD artery or by cannulating the LAD artery and perfusing it by means of a calibrating precision pump. We alternated episodes of normal coronary flow with various levels of reduced flow, either stepwise or complete, to simulate a range of different forms of acute ischemia. Regular monitoring of blood gases ensured that perfusion was adequate and pH was normal between interventions. In each of preparations, we recorded epicardial potentials simultaneously from 490 epicardial sock electrodes contained in a flexible array attached to a nylon stocking fitted over the ventricles. We recorded all channels from the sock electrode using our custom built acquisition system, which permitted simultaneous recording of up to 1024
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
11
Fig. 2. Experimental findings and simulations of increasing duration of ischemia. Upper panel contains ST-segment potential maps from an experiment with prolonged ischemia under 10% of normal flow rates in the cannulated LAD artery. Values in minutes indicate the duration of the reduced flow from onset. The lower panel contains the simulation results with different values of gap junction resistance to mimic early (left map) and late (right map) durations of ischemia.
channels at a sampling rate of 1000 Hz with adjustable gain and each channel buffered by sample-and-hold circuitry, as described previously [32,33]. 3. Results We present here 2 representative results for which we have qualitative agreement through experiments. In the first, in Fig. 1, we show the effect of transmural extent of ischemia. In the experiment, this results from progressively reducing the flow through the cannula from 50 of normal flow to 25 and then 1, with 6 minutes of time between each of the measurements shown in the figure. The equivalent simulation result, shown in the lower panel, was to compute the epicardial potentials for subendocardial ischemia that ranged from 40 through 70 to 90 of the transmural extent. Note the rotation of the axis drawn in each panel that joins the peripheral minima through the central elevation of
Fig. 3. Proposed mechanism of epicardial potentials from subendocardial ischemia. The diagram shows schematically the ischemic and healthy zones, and the extracellular potential difference that arises at the border between the two. The size of plus and minus signs indicates the relative amplitude of the potential difference, which differs at lateral and transmural boundaries. Also shown is the assumed local fiber orientation.
potential. We will outline the significance of this finding in the Discussion section. Fig. 2 contains results from an experiment in which we performed a prolonged reduction in LAD arterial flow and approximately equivalent simulations. The upper panel of the figure presents an overview the ST-segment potentials after reduction in LAD flow to 10 of normal rates. One can clearly see the transition from control levels through a peak level of ST-segment elevation after 9 minutes of occlusion, followed by a gradual reduction of elevation almost to control levels after 24 minutes. The lower panel contains 2 ST-segment potential maps simulated at 70 transmural subendocardial ischemia with 2 different values of gap junctional resistance to also mimic early and late phases of ischemia. 4. Discussion The results of these simulations and experiments are still preliminary, yet they already suggest both qualitative agreement between the computational model and experiments, as well as mechanisms for ischemic injury current. By studying the potentials computed throughout the heart in the simulations, we have developed a mechanistic model of injury current that can explain both our findings and those of others in previous studies, the details of which are available elsewhere [2,6]. The key component of this proposed mechanism, shown in Fig. 3, is that although the intracellular potential difference across the border zone can be uniform, we assume 1 transmembrane potential for all the healthy cells and a second, uniform transmembrane potential for the entire ischemic zone; the resulting extracellular potentials depend on local tissue conductivity. Thus, as the local conductivity changes, as we know it does in an anisotropic model of the myocardium, then so too must the extracellular
12
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Fig. 4. Postmortem cardiac imaging of experimental hearts. The left-hand panel contains a photo of the heart used for the MRI and the right-hand panel a volume-rendered reconstruction of the anatomy and the markers based on volume rendering the set of magnetic resonance images. The photo shows as well the site of occlusion in the LAD artery.
potential. In the specific case in Fig. 3, we have assumed circumferentially oriented fibers so that the relationship between fiber orientation and ischemic border orientation is quite different between the lateral borders and the transmural border. As a result, so too are there potential differences at the respective boundaries. With these potential differences arising between lateral and transmural ischemic borders, there will be current flow within the heart, which will lead to potentials on the epicardial surface, as indicated in Fig. 3. In the configuration shown, the epicardial potential over the ischemic zone would appear positive and that over the lateral border zones would be relatively negative. On the body surface, one would then see ST-segment elevation and depression, which would appear at locations dictated by the locations of the heart within the torso and the ischemic region within the heart. Probably, the greatest challenge in performing ischemia experiments is establishing the degree of perfusion in the tissues to identity the regions most likely to experience ischemia. By reducing blood flow to a major coronary artery and observing reproducible and characteristic changes in epicardial potentials, we feel confident that we are, indeed, creating ischemic regions; their location and transmural extent, on the other hand, are still relatively poorly documentable. A related source of uncertainty is that the geometric models for the simulations came from a single geometry, the Auckland canine model [24], and thus can
agree only qualitatively with any data we obtained from actual experiments. There is every reason to expect that changes in cardiac geometry, even with the same ischemic regions included, will lead to different epicardial potentials. We have therefore begun to use magnetic resonance imaging (MRI) to acquire high-resolution scans of the hearts in our studies, with the goal of creating realistic geometric models, ideally in which we have also marked the perfusion bed of the cannulated vessel. Fig. 4 shows a photograph (left-hand panel) and a volume rendering (right-hand panel) of the MRI data from the same sample heart. Also shown in both panels are the reconstructions of the MRI contrast markers that we used to align geometries; we will attempt to use the same agent to mark the perfusion beds of the LAD in future experiments. At present, we are creating geometric models from this heart/torso tank geometry with which to carry out further studies. In future experiments, we will infuse magnetic resonance contrast agents into the coronary circulation and attempt to image them and thus the perfusion bed for a particular heart. Enhanced authenticity of the geometric model can only help in comparing computations with experiments and thus in refining our simulations and the mechanisms such simulations suggest. Acknowledgments Support for this research has come from awards from the Nora Eccles Treadwell Foundation, the Richard A and Nora
R.S. MacLeod et al. / Journal of Electrocardiology 38 (2005) 8 – 13
Eccles Harrison Fund for Cardiovascular Research (San Leandro, CA), and the National Institutes of Health NCRR National Center for Research Resources for the Center for Bioelectric Field Modeling, Simulation, and Visualization (HL P41RR12553). The authors express their gratitude to Matt Allison, Yonild Lian, Patti Larrabee, and Jayne Davis for expert technical assistance in the experimental studies, to Ted Dustman for development of analysis software, and to Phil Ershler and Bruce Steadman for the acquisition system. We also thank Drs Elliot McVeigh and Dan Ennis from the National Heart, Lung, and Blood Institute and Stanford University, respectively, for their generous assistance in generating magnetic resonance images of the heart.
References [1] Stinstra JG, Hopenfeld B, MacLeod RS. A model for passive cardiac conductivity. Int J Bioelectromagnetism 2003;5(1):185. [2] Hopenfeld B, Stinstra JG, MacLeod RS. Mechanism for ST depression associated with contiguous subendocardial ischemia. J Cardiovasc Electrophysiol 2004;15(10):1200. [3] Stinstra JG, Hopenfeld B, MacLeod RS. Using models of the passive cardiac conductivity and full heart anisotropic bidomain to study the epicardial potentials in ischemia. Proceedings of the IEEE Engineering in Medicine and Biology Society 26th Annual International Conference: IEEE Press; 2004. [4] Shome S, Stinstra JG, Hopenfeld B, et al. A study of the dynamics of cardiac ischemia using experimental and modeling approaches. Proceedings of the IEEE Engineering in Medicine and Biology Society 26th Annual International Conference: IEEE Press; 2004. [5] Shome S, Rossi S, Punske B, et al. Anisotropic epicardial ST potential patterns reveal location and depth of ventricular necroses. Proceedings of the BMES 2004 Annual Fall Meeting: BMES; 2004. [6] Hopenfeld B, Stinstra JG, MacLeod RS. The effect of conductivity on ST segment epicardial potentials arising from subendocardial ischemia. Annal Biomed Eng 2005;33(6):751. [7] Samson WE, Scher AM. Mechanism of ST-segment alteration during acute myocardial injury. Circ Res 1960;8:780. [8] Holland RP, Brooks H. The QRS complex during myocardial ischemia: an experimental analysis in the porcine heart. J Clin Invest 1976;57:541. [9] Holland RP, Brooks H. TQ-ST segment mapping: critical review and analysis of current concepts. Am J Cardiol 1977;4:110. [10] Holland RP, Arnsdorf MF. Solid angle theory and the electrocardiogram: physiologic and quantitative interpretations. Prog Cardiovasc Dis 1977;19(6):431. [11] Eyster JAE, Meek WJ, Goldberg H, et al. Potential changes in an injured region of the cardiac heart. Am J Physiol 1938;717:717. [12] Nahum LH, Hamilton WF, Heff HE. Injury current in the electrocardiogram. Am J Physiol 1943;139:202.
13
[13] Binetti G, Markiewicz W, Billingham M, et al. Problems in assessing infarction size by epicardial mapping: preliminary studies with quinidine. Eur J Cardiol 1976;4(4):429. [14] Fozzard HA, DasGupta DS. ST-segment potentials and mapping: theory and experiments. Circulation 1976;54:533. [15] Moroko PR, Libby P, Covell JW, et al. Precordial ST-segment elevation mapping: an atraumatic method for assessing alterations in the extent of myocardial ischemic injury. Am J Cardiol 1972;2:223. [16] Fozzard HA. ST-segment mapping is not a clinical tool. In: Rapport, RS, editor. Current controversies in cardiovascular disease. Saunders; 1980. p. 281. [17] Fozzard HA, Makielski JC. The electrophysiology of acute myocardial ischemia. Annu Rev Med 1985;3:275. [18] Kle´ber AG. Extracellular potassium accumulation in acute myocardial ischemia. J Mol Cell Cardiol 1984;16:389. [19] Kle´ber AG, Janse MJ, Wilms-Schopmann FJG, et al. Changes in conduction velocity during acute ischemia in ventricular myocardium of the isolated porcine heart. Circulation 1986;73:189. [20] Smith GT, Geary G, Ruf W, et al. Epicardial mapping and electrocardiographic models of myocardial ischemic injury. Circ Res 1979;60:930. [21] Smith GT, Geary GG, Blanchard W, et al. An electrocardiographic model of myocardial ischemia injury. J Electrocardiol 1983;16(3):223. [22] Li D, Li CY, Yong AC, et al. Source of electrocardiographic ST changes in subendocardial ischemia. Circ Res 1998;82:957. [23] Johnston PR, Kilpatrick D. The effect of conductivity values on ST segment shift in subendocardial ischaemia. IEEE Trans Biomed Eng 2003;50(2):150. [24] Nielsen PMF, Grice IJL, Smaill BH, et al. Mathematical model of geometry and fibrous structure of the heart. Am J Physiol 1991; 260:H1365. [25] Henriquez CS. Simulating the electrical behavior of cardiac tissue using the bidomain model. Crit Rev Biomed Eng 1993;21(1):1. [26] Clerc L. Directional differences of impulse spread in trabecular muscle from mammalian heart. J Physiol 1976;255:335. [27] Kle´ber AG, Janse MJ, Capelle FJLv, et al. Mechanism and time course of ST- and TQ-segment changes during acute regional myocardial ischemia in the pig heart determined by extracellular and intracellular recordings. Circ Res 1978;42:603. [28] Roberts DE, Hersch LT, Scher AM. Influence of cardiac fiber orientation on wavefront voltage, conduction velocity and tissue resistivity. Circ Res 1979;44:701. [29] Roberts DE, Scher AM. Effect of tissue anisotropy on extracellular potential fields in canine myocardium in situ. Circ Res 1982;50:342. [30] MacLeod RS, Punske B, Shome S, et al. The role of heart rate and coronary flow during myocardial ischemia. J Electrocardiol 2001:43. [31] MacLeod RS, Miller RM, Gardner MJ, et al. Application of an electrocardiographic inverse solution to localize myocardial ischemia during percutaneous transluminal coronary angioplasty. J Cardiovasc Electrophysiol 1995;6:2. [32] MacLeod RS, Ni Q, Punske B, et al. Effects of heart position on the body-surface ECG. J Electrocardiol 2000;33(Suppl):229. [33] MacLeod RS, Lux RL, Taccardi B. A possible mechanism for electrocardiographically silent changes in cardiac repolarization. J Electrocardiol 1997;30(Suppl):114.