Epicardial ST Depression in Acute Myocardial Infarction
Abstract—The presence of electrocardiographic ST depression in acute infarction remains controversial and poorly explained. A combined animal and modeling study was performed to evaluate the source of ST changes in acute infarction. In anaesthetized sheep, small infarcts showed uniform ST elevation over the infarction whereas larger infarcts showed marked ST depression over the normal myocardium in addition to the ST elevation. These findings were replicated by bidomain models of the heart. A hollow sphere was used to model a gradually increasing infarct, and this showed that there was a decrease in the ratio of ST elevation to ST depression as the infarct was increased. The current flowing out of the heart must be identical to the current flowing back into the heart. This means that any infarction will produce ST depression as well as ST elevation, the ratio between the two being related to the size of the infarction. Small infarction is associated with a small region of ST elevation and minor ST depression of the remaining myocardium, and as the infarct region increases, the amplitude of the epicardial ST elevation falls and the amplitude of the ST depression increases. Infarction size is proportional to both the height of the ST depression on the epicardium and the strength of the epicardial ST segment dipole.
The origins and significance of ST depression associated with acute myocardial infarction are poorly understood and controversial.1 2 3 4 5 6 7 8 As part of a study looking at partial-thickness ischemia in an experimental animal model,9 we observed that ST depression accompanied some episodes of full-thickness ischemia and not others. The literature reflecting experimental infarction has shown that full-thickness ischemia was associated with a region of epicardial ST elevation over the ischemia with minimal changes elsewhere.10 11 12 13 14 15 16 This discrepancy between clinical observation and experimental results has been more fully evaluated by detailed epicardial, endocardial, and body surface ECG mapping of acute infarction in different territories and of different sizes in an experimental sheep model. The electrical changes were correlated with regional blood flow measured by fluorescent microspheres. To explain the results of the experimental infarction, we have developed several levels of a bidomain model based on that described by Tung,17 including a hollow thick-walled sphere, and a finite element model of the heart that replicated the experimental observations.
Materials and Methods
A total of 33 sheep were randomized into three groups. Transmural ischemia was achieved by completely ligating the obtuse marginal branch (OM) in group 1, the proximal left circumflex coronary artery (LCX) in group 2, and the proximal left anterior descending coronary artery (LAD) in group 3 for a minimum of 20 minutes. The epicardial ST potential fields were recorded at 1, 2, 5, 10, 15, and 20 minutes for a period of 2 seconds, respectively. The regional myocardial blood flow (RMBF) was measured before and 20 minutes after the artery was occluded. In 10 animals of groups 2 and 3, the endocardial and the epicardial potential fields were also recorded simultaneously. The left ventricular pressure, the left atrial pressure, the coronary artery flow, and lead II of the ECG were also monitored before and during ischemia. The details of the surgical procedures have been published previously.9 The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1985). The RMBF was measured before ischemia and at 20 minutes of ischemia using fluorescent microspheres (Molecular Probes) as previously described.9 18
Potential Recording, Construction of Isopotential Maps, and Map Display
Epicardial potentials were recorded using an epicardial sock containing 64 electrodes (Cardiovascular Research and Training Institute, the University of Utah). Endocardial electrograms were recorded using a homemade 40-electrode basket mapping apparatus.9
The myocardium was represented by the bidomain model, in which intracellular and extracellular volumes occupy the same space and were separated everywhere by the membrane.
Hollow Thick-Walled Sphere
Consider two concentric spheres as a model of the left ventricle and its surrounding myocardium (Figure 1⇓). The inner sphere, of radius rb, contains the blood mass and the region between the inner sphere, and the outer sphere, of radius ra, represents the cardiac muscle. In terms of a spherical coordinate system (r, θ, φ), the blood mass is the region 0<r<rb and the myocardium is the region rb<r<ra.
We generated a region of transmural ischemia, which is axially symmetric about the azimuthal axis of the spherical coordinate system, so that the potentials have no φ dependence. Further assume that the region of ischemia covers the region 0<θ<θa in the myocardium. This set of ischemic regions has been solved analytically, as shown in the online Materials and Methods (see http://www.circresaha.org).
Isolated Heart Model
The bidomain equations were also solved in a realistically shaped isolated heart.9 Transmural ischemia was simulated using the above model on both small and large regions of the myocardium.
An expanded Materials and Methods section is available online at http://www.circresaha.org.
Immediately after the OM ligation, the heart rate, the left ventricular systolic pressure, and the LAD flow (measured by flow probe) often increased slightly and gradually recovered to the control level after 20 minutes. No ventricular arrhythmias and conduction block were observed, and all the animals survived until the end of the experiment in this group. Immediately after the LAD or the LCX ligation, the left ventricular systolic pressure and the coronary flow to the nonischemic region increased slightly, but the left ventricular systolic pressure and the coronary flow to the nonischemic region started to drop after 2 minutes in all but 2 animals. The left ventricular end-diastolic pressure increased in all but 4 animals within 15 minutes. In 9 of 25 animals with either the LAD or the LCX ligation, ventricular fibrillation developed within 5 minutes, and the animals died within 15 minutes (no data from these 9 animals were included). Ventricular fibrillation developed within 20 minutes in 7 animals, all of which died within 30 minutes. The occurrence of ventricular fibrillation was preceded by a period of sustained (0.5 to 3 minutes) ventricular tachycardia. Nine animals survived the 30- to 60-minute observation period; however, 5 developed ventricular ectopics and nonsustained ventricular tachycardia, and one developed atrioventricular block. The animal survivals in different artery occlusions are summarized in Table 1⇓.
Regional Myocardial Blood Flow and Hemodynamic Response
Selected hemodynamic results and changes in the myocardial blood flow after 20 minutes of ligation of the different vessels are presented in Table 2⇓ and Figure 2⇓, which show significantly different flow and hemodynamic changes with varied infarctions. Both the LCX and the LAD ligation caused a marked decrease in the flow to the infarcted regions (from 1.02±0.10 to 0.19±0.07 mL · min−1 · g−1 in the LCX ligation, from 0.99±0.20 to 0.16±0.09 mL · min−1 · g−1 in the LAD ligation, both P<0.001). In the noninfarcted regions, there was also a considerable decrease in RMBF (from 1.03±0.13 to 0.73±0.20 mL · min−1 · g−1 in LCX ligation, from 0.89±0.21 to 0.64±0.22 mL · min−1 · g−1 in LAD ligation, both P<0.05). The LCX and LAD ligations were accompanied by an increase in the left ventricular end-diastolic pressure (from 3±1 to 8±3 mm Hg in LCX ligation, from 1±4 to 6±7 mm Hg in LAD ligation, both P<0.05) and a decrease in the left ventricular systolic pressure (from 81±12 to 72±20 mm Hg in LCX ligation, from 92±11 to 69±13 mm Hg in LAD ligation, both P<0.05; Table 2⇓). However, in the OM ligation, the flow to the noninfarcted region increased nonsignificantly (from 1.09±0.19 to 1.12±0.15 mL · min−1 · g−1, P>0.05), although flow to the infarcted regions decreased by 62% (from 1.00±0.18 to 0.38±0.14 mL · min−1 · g−1, P<0.001). Diastolic pressure and the left ventricular systolic pressure were unchanged during the OM ligation (Table 2⇓).
The transmural flow distributions at 20 minutes of ischemia for the OM, the LCX, and the LAD ligations are presented in Table 3⇓. During infarction, there were similar changes in the RMBF to each third of the myocardium in both the infarcted and noninfarcted regions; the endocardial/epicardial flow ratio remained unchanged. Figure 2⇑ displays the spatial flow distributions of the left ventricles in different artery occlusions. It was plotted with the data of 3 animals from groups 1, 2, and 3, respectively. In OM occlusion, the flow reduction occurred only in the infarcted region (Figure 2A⇑). In either the LCX or the LAD occlusion, however, flow to the noninfarcted region also reduced considerably (Figure 2B⇑ and 2C⇑). In all 3 occlusions, flow reduction to each third of the ventricular wall was similar.
Linear correlation was used to determine whether a statistically significant relation existed between the RMBF change during ischemia in the noninfarcted regions and the simultaneous measurements of the left ventricular end-diastolic pressure, the left ventricular systolic pressure, and the mean left atrial pressure. The RMBF change during ischemia in the noninfarcted regions inversely correlated to the left ventricular end-diastolic pressure (r=−0.82, P=0.0002) and the mean left atrial pressure (r=−0.79, P=0.0001) and directly correlated to the left ventricular systolic pressure (r=0.82, P=0.0001; Figure 3⇓). During acute myocardial infarction, the RMBF changes in the noninfarcted regions were related to both the perfusion pressure and cardiac function.
Epicardial Potential Distribution in Different Sizes of Infarction
Representative maps of epicardial ST potential distribution from 3 typical experiments are displayed in Figure 4⇓ and show significantly different ST alterations with each different infarcted region. Total occlusion of the OM produced a graduated but even peak of ST elevation in the ischemic center, with the magnitude decreasing toward the border. The ST segment was depressed slightly at the surrounding regions (Figure 4⇓). The highest potential occurred between 5 and 10 minutes after the OM ligation.
Total ligation of either the LAD or the LCX produced a powerful dipole between ST elevation over the infarcted region and ST depression over the noninfarcted region. The region of ST elevation is markedly asymmetric, with the highest amplitude at the boundary, a pattern of ST elevation quite different from that of the total occlusion of the OM (Figure 4⇑).
Endocardial Potential Mapping in Large Infarction Compared With Epicardial Potential Mapping
From the endocardium, we recorded both ST elevation and ST depression during the ligation of either the LCX or the LAD. The results are shown in Figures 5⇓ and 6⇓ which display the simultaneous epicardial and endocardial ST potential recordings. The distribution patterns of ST changes in the endocardium are similar to those in the epicardium, except that potentials were lower in the endocardium. The changes in the amplitude of the ST potentials with time were also similar in the epicardium and the endocardium, ie, potential changes occurred within 30 seconds after the occlusion, reached their maximum within 5 to 10 minutes, and then decreased from 15 minutes onward.
Concentric Spheres Model
The bidomain model produced a set of curves around the circumference of the spheres that were symmetrical around the vertical axis for 8 different infarct sizes ranging from 0.2 radians, 6.36% of the total surface area, to 1.6 radians, or 50.1% of the surface area. The results are shown in Figure 7⇓. There was a clear increase in the ratio of negative to positive potentials as the size increased. We evaluated the physical dimensions against this shift looking at ischemic region on the sphere, volume of ischemia, surface area of ischemia, and the integral of current density under positive and negative regions. There was a one-to-one correspondence between the integral of current density under positive and negative curves for all infarct sizes but no relationship for the other parameters. This implies that the set of curves is produced within the constraints of the bidomain model, which requires, as the heart does, that the overall current lost from the heart is zero.
The model in Figure 8⇓ shows that the patterns are very similar to those measured in the sheep. We chose to analyze multiple infarcts using the spheres for simplicity because the real heart model has the right ventricle over the septum, which interferes with the epicardial field, thus making comparisons more difficult.
In our experimental work, it was clear that the large ischemic regions were associated with marked ST depression over the nonischemic region whereas the small ischemic regions had minimal ST depression elsewhere. These results differed from the findings in the literature mainly because the previously reported studies were from small regions of infarction, which maintained a stable condition of the animal.
To explain our findings, we looked first to see if the normal (nonischemic) region was truly not ischemic. There was some suggestion of ischemia in the normal regions because of the reduced flow from microsphere measurement, but there was no ST elevation on the normal endocardium, which would be expected if there was concomitant ischemia. The overall reduction of flow was also small and unlikely to produce ischemia in its own right. That the flow was reduced was surprising and probably due to a decreased pressure gradient across the coronary bed. This observation needs further investigation in terms of possible therapeutic gains from maneuvers to increase blood flow in the nonischemic regions.
The significance of this remained unclear until the concentric spheres model was analyzed. This enabled a series of incremental infarcts to be studied, as shown in Figure 7⇑. The results suggest that some basic balance between size of ischemia and ST elevation to ST depression ratio existed. An examination of physical factors has shown that the reason for this lies with basic physics.
The total current flowing out of the heart must flow back into the heart, and this paradigm was shown to be true in that the integral of current density over the ischemic region matches that over the normal region. This basic property of physics dictates that the overall current out of the heart must be zero; hence, all ST balances between elevation and depression are subject to this. The ST depression is part of the source, and the balance between elevation and depression is dependent on the zero line set by the requirement that the overall current from the heart is zero. In human practice, this requirement is modified by the use of the Wilson central terminal19 to set the reference potential.
Thus, any large infarct will have both ST elevation and ST depression generated at the ischemic boundary on the epicardium and the larger the infarct, the greater the ST depression. It must be pointed out, however, that these results apply only to the epicardial potentials not the body surface potentials. Although the body surface potentials are generated by the cardiac currents, they are sufficiently modified by the shape and conductivity of the organs within the thorax to make direct comparisons difficult.
There is support for this view from one of the few inverse transform studies carried out in humans with acute infarction.20 In this study, the epicardial potential distributions were derived from more than 200 patients with acute infarction. The results showed that a strong dipole on the epicardial surface predicted mortality for all patients including those with single vessel disease. At the time, no clear explanation for these results was advanced. Given the study presented here, it is clear that the dipole reflected the overall size of the infarction. Further support comes from a recent modeling study,21 which concluded that the source with full-thickness ischemia had both negative and positive ST components and which replicated the 12-lead appearance of angioplasty-induced ischemia.
The present study has provided an explanation for ST depression on the epicardium in patients with acute infarction. It also suggests that measurement of the negativity of the dipole on the epicardium should relate well to infarction size. Clearly, it is possible to have additional ischemia, but this will also be constrained by the need to have the overall current from the heart be zero. Thus, the presence of additional ischemia will probably increase the ST elevation as well as increasing the region of ST depression. However, it is not necessary to have ischemia other than the infarction to produce marked ST depression. It is important to remember that in electrocardiology the net current leaving the heart must always be zero.
This work was supported in part by grants from the Government Employees Medical Research Fund of Australia, the National Heart Foundation of Australia (the Sandy Baker Grant), and the Clive and Vera Ramaciotti Foundation of New South Wales. Drs Danshi Li and Chuan Yong Li were sponsored by the Australian Agency for International Development.
- Received May 25, 1999.
- Accepted August 30, 1999.
- © 1999 American Heart Association, Inc.
Gibson RS, Crampton RS, Watson DD, Taylor GJ, Carabello BA, Holt ND, Beller GA. Precordial ST-segment depression during acute inferior myocardial infarction: clinical, scintigraphic and angiographic correlations. Circulation. 1982;66:732–741.
Li DS, Li CY, Yong AC, Kilpatrick D. The source of electrocardiographic ST changes in subendocardial ischemia. Circ Res. 1998;82:957–970.
Samson WE, Scher AM. Mechanism of S-T segment alteration during acute myocardial injury. Circ Res. 1960;8:780–787.
Holland RP, Brooks H. Precordial and epicardial surface potentials during myocardial ischemia in the pig. A theoretical and experimental analysis of the TQ and ST segments. Circ Res. 1975;37:471–480.
Vincent MG, Abildskov JA, Burgess MJ. Mechanisms of ischemic ST-segment displacement: evaluation by direct current recordings. Circulation. 1977;56:559–566.
Kleber AG, Janse MJ, van Capelle FJC, Durrer D. Mechanism and time course of S-T and T-Q segment changes during acute regional myocardial ischemia in the pig heart determined by extracellular and intracellular recordings. Circ Res. 1978;42:603–613.
Richeson JF, Akiyama T, Shenk E. A solid angle analysis of the epicardial ischemic TQ-ST deflection in the pig: a theoretical and experimental study. Circ Res. 1978;43:879–888.
Smith GT, Geary G, Ruf W, Roelofs TH, McNamara JJ. Epicardial mapping and electrocardiographic models of myocardial ischemic injury. Circulation. 1979;60:930–938.
Tung L. “A bidomain model for describing ischemic myocardial DC potentials,” in PhD thesis, MIT, 1978.