- Email: [email protected]
How many ECG leads are required for in vivo studies in safety pharmacology?
Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Contents lists available at ScienceDirect
Journal of Pharmacological and Toxicological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p h a r m t ox
Appraisal of state-of-the-art
How many ECG leads are required for in vivo studies in safety pharmacology? Robert L. Hamlin ⁎ The Ohio State University, Scientific Director, QTest Labs, 1900 Coffey Rd., Columbus, Ohio 43210, United States
a r t i c l e
i n f o
Article history: Received 20 August 2007 Accepted 13 March 2008 Keywords: Electrocardiography Equivalent generator Leads Methods
a b s t r a c t This article explains the principles of electrocardiography, and explains how it is used by Safety Pharmacology, with a focus on the requirement for multiple leads in Safety Pharmacology assessment. Electrocardiography as used in different disciplines (e.g., medicine, anesthesiology, physiology, and pharmacology/toxicology/safety pharmacology) has different requirements for the number of electrodes applied. Electrodes may be placed at an infinite number of points on the body, and voltages (electrocardiograms) may be registered between/ among them. However in safety pharmacology there is little evidence that more than 1—or at most 3—lead(s) is (are) required to provide all of the information that might be present using an infinite number. This is based upon (1) the biophysics of the heart as a generator of electrical potential/voltage, (2) the fact that most properties of electrophysiology affected adversely by drugs are expressed as changes in durations, and (3) experience. A single, unipolar lead (V3) recorded from the left sternal border at the 5th intercostal space possesses minimal artifact and large, stable deflections. This lead allows for accurate measurement of heart rate and rhythm, durations of component deflections (e.g., PQ, QRS, QT), and J-point deviation. A greater number of leads seldom or never yield additional information that detects liabilities. Commonly voltages recorded between the right thoracic and left pelvic limbs (lead II) provides information similar to lead V3, and lead II is easier to apply, and produces voltages with less artifact and similar to those in lead V3. A lead measuring the voltage between the left and right thoracic limbs (lead I) along with lead II allows for estimating orientation of vectors in the frontal plane, but knowledge of these vectors seldom or never indicates liability of a test article. © 2008 Published by Elsevier Inc.
1. Introduction
1.3. What are electrocardiographic leads?
1.1. Why is electrocardiography used in safety pharmacology?
Lead is used three times in electrocardiography. It stands for a position (e.g., V3, rV2) on the torso from which voltages are recorded. It stands for a wire extending from the patient to the electrocardiograph. It stands for a combination of electrodes between which voltages are measured. This discussion will be about points on the torso surface from which voltages should be registered.
Electrocardiography is a method essential for studies in safety pharmacology, because it detects, simply, inexpensively, and noninvasively, many (most) drug-induced changes in electrophysiological properties of the heart (e.g., chronotropy, dromotropy, bathomotropy) that might translate to morbidity or mortality if expressed in man.
1.4. What is an electrocardiograph and what are types of leads? 1.2. What is the electrocardiographic system? There are 4 components to the electrocardiographic system (Fig. 1): (1) the heart as a generator of voltage, (2) the body volume which allows for transmission of those voltages to the torso surface, (3) the voltmeter (electrocardiograph) which detects, amplifies and records the voltages, (4) the electrocadiographer who interprets the electrocardiograms (the recordings of the voltages) to determine the state of electrophysiological health of the heart, and who often makes recommendations for treatment or identifies toxicity.
⁎ Corresponding author. Fax: +1 614 292 3646. E-mail address: [email protected]. 1056-8719/$ – see front matter © 2008 Published by Elsevier Inc. doi:10.1016/j.vascn.2008.03.004
An electrocardiograph is a voltmeter which registers differences of electrical potential (the force that drives electrons from place to place). As such a voltmeter requires 2 inputs, termed positive (+) and negative (−). These inputs come from positions (leads) on the torso and/or limbs of the subject. The electrocardiograph measures voltage (potential to drive electrons) between these points, and of course there is an infinite number of points on the torso/limbs between which potential differences can be measured and might have meaning to the electrocardiographer. For example: (1) lead I measures the voltage difference between the left and right thoracic limbs, where the lead from the left thoracic limb goes to the positive pole of the electrocardiograph, and the lead from the right thoracic limb goes to the negative pole; (2) lead V3 measures the voltage difference between an electrode on the left portion of the thorax and a
162
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 1. Overview of the electrocardiographic system, including: the heart as a generator of potential, the body as a volume conductor, electrodes placed on the limbs or torso surface, leads (wires) extending from the electrodes to the voltmeter (electrocardiograph), the electrocardiograph, the electrocardiogram, the interpreter.
combination of electrodes from the right and left thoracic and the left pelvic limbs. The electrode from the thorax goes to the positive pole of the electrocardiograph and the 3 electrodes (combined) from the limbs go to the negative pole; (3) lead aVF measures the voltage difference between an electrode on the left pelvic limb and a combination of electrodes from the left and right thoracic limbs, where the lead from the left pelvic limb goes to the positive pole of the electrocardiograph and the 2 leads (combined and modified by high resistance) from the thoracic limbs go to the negative pole of the electrocardiograph. Voltage differences between only 2 points form so-called bipolar leads; voltage differences between 1 point and the combination of the 3 limbs form so-called unipolar leads; voltage differences between 1 point and the combination of points from the 2 other limbs form socalled augmented unipolar leads. 1.5. What determines the numbers of leads and which one(s) is (are) required? Of course the number of leads possible (permutations of inputs to the poles of the electrocardiograph) is infinite, therefore it is important to know how many, and which ones, are required to provide the information necessary for the mission of electrocardiography. That depends upon the mission. In clinical electrocardiography of humans (Morganroth & Gussak, 2004; Zipes & Jalife, 2004), to detect chamber enlargement, to localize myocardial ischemia and infarction, and to quantify dispersion of QTc among leads, 12 or more leads may be necessary. In conventional electrocardiography in veterinary clinical medicine (Detweiler, 1988; Tilley, 1993), 6 leads (of which only 2 or 3 of the 6 present information that is unique) are usually taken. But if the interest is only in determining heart rate, rhythm, durations of various electrophysiological processes (e.g., rate of discharge of the SA node, speed of conduction through the atria or ventricles or from atria to ventricles, durations of both ventricular depolarization and repolarization) or to search for abnormal discharge from within the atria or ventricles, then only 1 lead may be necessary (Malik & Batchvarov, 2000). But it better be a “good” lead, i.e., one from which deflections can be measured accurately (Horan, Flowers, & Brody, 1964)! The decision on the numbers of leads is based upon experience. However the decision may be based upon theoretical, biophysical
considerations as well; for example, how does the heart behave as an equivalent generator of potential? Of course the final decision on how many, and which, leads are necessary in studies on safety pharmacology may be answered only by the medical community and drug regulatory agencies who could state how often 1, 2, 3+….+n leads produced information that predicted safety with 100% sensitivity and 100% specificity. Since 100% sensitivity and specificity are impossible to achieve, rather we seek the fewest number of leads that produces results as good as the greatest number of leads ever used. 2. The heart as a generator of potential The voltages are produced on the body surface by waves of depolarization (Durrer & Van der Tweel, 1965; Scher, 1964) and repolarization traversing the atria and ventricles (Fig. 2). These waves are in fact sheets of dipoles (positive and negative charges separated by a small distance) in which the face of the sheet in stimulated (depolarized) myocardium is comprised of negative charges, while the face of the sheet in the resting (repolarized) portion of the myocardium is comprised of positive charges (Wilson & Bayley, 1950; Bayley et al., 1954). Fig. 3 shows a dog, its heart (albeit an abnormally large one) within the torso, and a single sheet of dipoles, frozen at an instant, traveling through the heart in a general direction from head to tail. The sheet of dipoles has negative signs in the region that is depolarized, and positive signs in the region that is yet to be depolarized. We may predict/estimate the magnitude and sign of the voltage produced by this wave at any point (P), say on left hind leg of the torso, using (1) the solid angle (Fig. 4) concept (Holland & Ornsdorf, 1977), (2) the vector (Fig. 5) concept (Frank, 1954; Grant & Estes, 1951), or (3) by gross approximation (Fig. 6) (Hamlin & Smith, 1960). Using the solid angle concept (Fig. 4), we construct radii from the open ends of the sheet to the point, measure the angle (Ω1) determined by the radii, and state that the sign will be positive because the point is in the general milieu of the positive face of the sheet, and that the voltage will be proportional to the magnitude of the angle. For mathematical correctness the magnitude of the angle should be multiplied by 2 constants, one dependent upon the resistive 1 In this example the angle is in one plane, whereas the boundary is actually 3 dimensional, therefore the angle should be a solid angle.
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
163
Fig. 2. A schematic diagram of the heart, showing the wave of depolarization emitted from the SA node, traveling first through the atria, and then through the ventricles (Fig. 2A). An ECG, shown to the right, has colors which match up to which waves of depolarization produce which voltages (Fig. 2B).
characteristics of the body, the other upon certain geometrical considerations (Bayley et al., 1954; Brody, 1956). The sheet of dipoles may be represented, also, by a single vector (the wide arrow in Fig. 5) whose length is proportional to the area of the sheet of dipoles and whose direction represents the direction the sheet is moving (Grant & Estes, 1951). The voltage, at a point on the torso (P), generated by the sheet of dipoles represented by the vector, is proportional to the projection of the vector on a lead axis (dotted line) constructed from the point on the torso through the origin of the vector or the center of the heart. The length of the projection is proportional to the cosine of the angle (θ) between the vector and the lead axis, times the magnitude (length) of the vector. For mathematical correctness the projection should be divided by the square of the distance between the vector and the point, and again multiplied by a constant which represents the conductive properties of the body. But it is assumed that the distance between the heart and the electrode is 1, therefore squaring the distance can be neglected. The sheet or vector may be represented by (resolved into) a single dipole—the so-called equivalent dipole. As the waves of depolarization
and repolarization (not considered here) traverse the myocardium, the sheets of dipoles change geometry (size and configuration) and direction of motion (Fig. 2), but at any and all times the sheet can still be represented by an equivalent dipole whose magnitude and orientation changes in concert with the changing sheet of dipoles, and that can be represented by a single vector whose length represents the magnitude of the equivalent dipole and whose spatial orientation represents the spatial orientation with which the sheets of dipoles traverses the torso (Armoundas, Feldman, Mukkamala, & Cohen, 2000; Armoundas et al., 2003; Bu & Berbari, 2006; Corbin & Scher, 1977; Frank, 1952; Fukuoka, Armoundas, Oostendorp, & Cohen, 2003). That being the case (i.e., that is if a single vector can represent the waves of depolarization and repolarization (Scher, Young, & Meredith, 1960; Stilli et al., 1986; Taccardi, 1990)) then it follows that both magnitude and orientation of that vector—and therefore of the sheets of dipoles—can be characterized unequivocally2 by 2 The vector may be characterized unequivocally, but there may not be a unique boundary.
164
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 3. A left-lateral view of a dog showing the heart and a wave of depolarization traversing the heart. The wave is a sheet of dipoles (separations of − and + signs), in which the face covered with − signs is in myocardium that has been depolarized, and the face with + signs is in the myocardium yet to be depolarized.
Fig. 5. The same dog and sheet of dipoles as in Fig. 3, but a vector is thrust through the sheet of dipoles. The length of the vector represents the size of the sheet of dipoles, and the direction it points is the direction the sheet is moving through the heart and torso. The amplitude of the voltage recorded from this lead may be predicted as the length of the projection of the vector on the lead axis (the imaginary line drawn from the point on the torso through the origin of the vector or the center of the heart). The magnitude of the projection of the vector is equal to the magnitude of the vector times the cosine of the angle (θ) between the vector and the lead axis.
3 leads (Fig. 7). In the case of this illustration, the 3 leads are orthogonal (i.e., they intersect at right angles and at the center of the heart) (Fig. 8). Leads I, aVF, and V10 could constitute the 3 leads required to characterize all so-called dipolar properties of the heart represented by an equivalent dipole (McFee & Johnston,1953; Bayley et al.,1954). In fact to estimate the sheets of dipoles most accurately, the orthogonal leads should be based upon electrodes equidistant from the heart, the heart should be in the center of the body, and the body volume should conduct uniformly. None of these requirements is fulfilled by any orthogonal lead system; however this limitation seldom compromises the clinical or toxicological usefulness of the characterization of the sheet of dipoles (Bayley et al., 1954; Flowers, 1964; Wilson & Bayley, 1950). If a vector is placed in a torso, any electrode on the torso surface may have a voltage with a sign of 0, + or −, and if + or − the magnitude may be from very small (b0.1 mV) to reasonably large (up to 5 mV). If the vector points towards an electrode, the sign (deflection) will be +; if it points away from the electrode, the sign will be −; if it points at right angles to the electrode, the sign will be 0 (i.e., there will be no deflection). A third method (Fig. 6) of predicting the magnitude and sign of a potential at a point on the body surface may be to imagine
yourself sitting at the point on the torso, looking at the heart, and determining whether the sheet is traveling toward you (+) or away from you (−), and how big the sheet is (the magnitude of the voltage). This approximation is semiquantitative but useful. A vector may be mimicked by a flashlight. Instead of producing a voltage on the torso surface, the flashlight produces either lightness or darkness. A positive voltage or a perception of light will appear at a point on the torso towards which a vector/flashlight points. If a single vector/flashlight represents the sheets of dipoles accurately (i.e., if the heart behaves as an equivalent dipole) (Fig. 9), then there can be only 1 set of + signs (maxima)/brightness and − signs (minima)/darkness on the torso surface (Scher et al., 1960; Taccardi, 1990). At the top of this figure are 2 boundaries of different sizes and traveling in different directions. Notice the lightness is greater for the big boundary than for the small boundary and that different regions on the torso are either light or dark depending upon where the vector/flashlight points. When both boundaries exist simultaneously (bottom of figure), both
Fig. 4. The same dog as in Fig. 3, but with a point on the left pelvic limb for which a voltage is to be predicted, and how the amplitude and sign of the voltage generated at the instant of the wave may be predicted by the magnitude of the angle (Ω) produced by radii constructed from the open ends of the boundary (sheet of dipoles).
Fig. 6. The same dog as in Fig. 3, but an imaginary person is sitting on the electrode and looking at the heart and sheet of dipoles. The voltage will be roughly proportional to the size of the boundary the person sees. If looking at the + face, the voltage will be +, if looking at the − face, the voltage will be −, and if the boundary is traveling at right angles to the person, no voltage will be present.
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
165
Fig. 7. An “ideal” ECG showing the 3 orthogonal leads (I, aVF, V10) and V3 at high paper speed, and lead aVF at slow paper speed. Measurements are more accurate from the trace at higher paper speed, but rhythm is observed more easily from lower paper speed.
flashlights are on, both illuminate regions, but the brighter one dominates the smaller one and the brightness it produces may only be observed when you are close to it—a so-called proximity effect or biased by proximity. Studies have been conducted on many species to validate that their hearts behave as an equivalent dipole. In some of these studies unipolar electrocardiograms were recorded from hundreds of points on the torso surface (Taccardi, Musso, & De Ambroggi, 1972). At many instances during the QRS complex, maxima and minima were sought. At each instance only a single maximum and minimum occurred, but of course the magnitudes and positions of these maxima and minima varied from time to time, as the magnitudes, geometry and locations of sheets of dipoles varied within the heart. Thus a single vector (i.e., one flashlight) with changing magnitude/brightness and orientation/ direction could explain all of the electrocardiographic information present in the hundreds of leads. Had 2 vectors been necessary at any instant, then the heart would have behaved as an equivalent quadrupole; there would have been 2 sets of maxima and minima, and more than 3 leads would have been required to characterize the 2 vectors. Again with reference to Fig. 9, bottom, this is illustrated by 2
flashlights, one bright and one dim, representing two waves of depolarization traveling in different directions. Although in general the larger flashlight dominates and makes the hind end of the dog light and the front end dark, the smaller flashlight has a local “contribution” and produces a dim light in the darkness and slight darkness in the lightness. 3. Which 3 leads are necessary? Although any 3 leads would be satisfactory to characterize the magnitude and orientation of a vector, it is easier (more intuitive) for an electrocardiographer to identify a vector if the 3 leads are mutually orthogonal, (i.e., intersect at right angles). One lead (lead I) determines if the vector points to the left or right; 1 lead (aVF) determines if the vector points to the head or tail (inferiorly in man); 1 lead (V10) determines if the vector points dorsad (posteriorly in man) or ventrad (anteriorly in man) (Fig. 8). Any lead other than those 3 can contribute nothing further to characterizing the vector (or sheets of dipoles), therefore any other lead may be considered redundant and superfluous, and serves as an unnecessary distraction to the electrocardiographer (Hamlin, Burton, Leverett, & Burns, 1974). 4. What are the exceptions (there always are some)?
Fig. 8. The orthogonal axes with a vector at the origin, and projections of the vector on each axis. The voltage recorded on each axis is proportional to the projection of the vector on that axis. Of course this is idealized, since at the very least electrodes seldom form truly orthogonal axes, the heart is not truly in the center of the axial system (i.e., it may be closer to one axis than to another), and there are serious inhomogeneities in the volume conductor which skew the projections.
Potential for non-dipolar properties and leads biased by proximity constitute possible exceptions (Flowers, Horan, & Brody, 1965; Flowers, Horan, Thomas, & Tolleson, 1969; Zhou, Yin, Nong, & Yu, 2002). Although 3 leads may be all that is necessary (theoretically) (Burger, van Brummelen, & Herpen, 1961) in practice some leads placed on the thorax close to the heart have deflections from which measurements can be made more easily and accurately. For example, the unipolar thoracic lead, V3, may appear qualitatively similar to lead aVF, however the end of the T wave is much less ambiguous (Fig. 7) in that is it terminates in the isoelectric (base) line rather sharply rather than fusing into the baseline gradually (Detweiler, 1988). Since the QT interval is so important in studies in safety pharmacology, it is obvious that lead V3 may be a welcome addition to the electrocardiographer who may not need it for characterizing vectors but who may find measurements of QT more precise when it is used. In addition there are some studies which claim significant nondipolar properties to the equivalent cardiac generator. Still these
166
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 9. A left-lateral view of a dog with its heart within and with electrodes placed at 3 points on the torso. Waves of depolarization are represented as flashlights pointing in the direction the waves are traveling. The voltages registered at each point will be mimicked by the relative brightness (if the flashlight points towards the point) or darkness (if the flashlight points away from the point). In the upper left panel, a single boundary/flashlight travels/points to the head and to the back, and the region toward which it travels/points becomes light whereas the region away from which it is pointing becomes black, the relative intensity related to the square of the distance between the source and the electrode. Because the boundary/flashlight is so small, no or very small voltages/lightness or darkness will be present at electrodes aVF or VN, but only at VP because it is so close. In the upper right panel showing a much larger/more intense boundary/light source, now voltage/lightness or darkness will be observed at aVF and VN, albeit greater at VP because of its proximity. In the bottom panel, both boundaries/flashlights exist simultaneously. The larger/brighter dominates the electrodes placed at a distance (aVF, VN), but the smaller dominates the single electrode (VP) so close to it. Thus although electrode VP is in the region that should contain darkness based upon darkness at VN and lightness at aVF, because of its proximity to the smaller flashlight and it being dominated due to that proximity, light will be observed. This is a so-called proximity effect, and can only exist when the heart is not represented as a single dipole (flashlight) that changes amplitude and orientation.
studies do not demonstrate superiority, over assuming that the heart is a dipolar source, in detecting either disease or drug effects. However there may be an even more important requirement for additional leads…if the cardiac generator is not only dipolar, but possesses multipolar (quadrupolar, octapolar) properties as well. If the cardiac activation process, at an instant, has 2 waves of depolarization, 1 large the other small, and 1 oriented spatially differently from the other, these 2 waves may be represented as 1 wave (or vector) when “viewed” from a relatively great distance, i.e., from electrodes more than 3 heart's diameter from the heart. However the larger wave will dominate the observation, so information about the smaller wave will be lost. It would be more accurate to express these 2 waves using 2 vectors. If each vector produces a single maximum and a single minimum, then 2 vectors will produce 2 maxima and 2 minima. Thus within a zone in which a positive deflection would be produced by the major vector, a negative deflection may be attributable to the smaller vector which influences electrodes placed close to smaller wave. Let us return to the analogy between a vector and a flashlight. A very weak (dim) flashlight is placed within the thorax and points dorsad and craniad (Fig. 9, top left). Lightness is observed any place on the torso towards which the light is pointing; darkness any place away from which the light is pointing. The further from the light source, the less bright, or dim, the light appears. Thus an electrode placed at VP will observe light, whereas electrodes at aVF or aVN will not be influenced. A very strong (bright) flashlight is placed within the thorax and points caudad (Fig. 9, top right). As with the dim flashlight, lightness is observed any place on the torso towards which the light is pointing, darkness is observed any place away from which the light is pointing, and the further from the light source, the less bright, or dim, the light appears. Thus electrodes placed at VP and aVN will observe darkness, whereas the electrode placed at aVF will observe light. Now
if both flashlights are activated together (i.e., both are turned on) (Fig. 9, bottom), electrodes at aVF and aVN will be influenced only by the brighter flashlight, whereas the electrode at VP is influenced, still, by the dimmer one; it is “biased by proximity.” Each flashlight mimics a dipolar source, and if you only measured light from a distance, the dimmer light source is relatively inconsequential and you will claim a dipolar source, but if you measure light from the distant point and from close to the dim source, you will identify a quadrapole. Non-dipolar sources usually contribute so-called high-frequency components in the electrocardiograms (Geselowitz, 1976; Tysler et al., 2004). These high-frequency components are identified as beads (Fig. 10A), slurs (Fig. 10B) or notches (Fig. 10C). Because by definition the deflections contain high-frequencies, which form is manifested on an electrocardiogram depends upon the frequency characteristics of the recorder. If the frequency content of a high-frequency component is 80 Hz and an electrocardiograph records only to 40 Hz, then a highfrequency component will either not be registered or it will be registered, as a form fruste, as a notch, bead, or slur. Non-dipolar sources generating high-frequency components to the QRS complex may occur with myocardial infarction or right ventricular hypertrophy. With right ventricular hypertrophy there appears to be 2 separate generators competing for dominance of the vector when both ventricles are activated nearly simultaneously. Without right ventricular hypertrophy, the left ventricle is normally so dominant that activation of the right ventricle is trivial, however, with right ventricular hypertrophy the dominance might shift from one ventricle to the other, so the vector shifts, rapidly, from endocardium to epicardium of the left ventricle, to endocardium to epicardium of the right ventricle, back to the left ventricle, and so on. Notice that both ventricles are nearly equal in mass. Using the flashlight analogy, with right ventricular hypertrophy, it is as if there are 2 flashlights, one pointing through each ventricle. Each is
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
167
particularly in certain leads. These imply more than dipolar sources, but, again, there is no data supporting the contention that multiple leads are necessary to identifying their presence or that they become toxicologically relevant in studies of safety pharmacology. 5. A recommended number of ECG leads for safety pharmacology. Does one need all 3 leads? All 3 leads are necessary in order to characterize the vector precisely. However there appears to be no instances in studies in safety pharmacology and/or toxicology where characterization of the vector has lead to identification of a toxic potential. Characterization of the vector may be necessary for clinical reasons, to detect and even to semiquantify chamber enlargement, but not to identify toxic potential of a drug. 6. What leads have proven useful? Given that characterization of the vector appears to be unnecessary to predict toxic potential, what is necessary? These measurements are periodicity of the heart (P–P interval), rhythm (sequences of depolarization and repolarization), atrioventricular conduction (PQ interval), duration of ventricular conduction (QRS duration), duration of ventricular repolarization (QT interval and QT interval corrected for heart rate), and rarely J-point deviation. These measurements can be made using a single lead (usually unipolar thoracic lead V3) (Bayley and Schmidt, 1954), and should determine potential toxic effects on the important electrophysiological properties of cardiac function (i.e., chronotropy, dromotropy, bathmotropy) which, if altered, may translate into morbidity and/or mortality. We must await information from a coalition of regulatory agencies, the pharmaceutical industry, and academia to document instances when more than a single lead has been more predictive of toxic potential than a single lead (Malik & Batchvarov, 2000; Morganroth & Gussak, 2004; Redfern et al., 2003). Then we must ask, “Which lead, and what feature(s) of the multiple leads resulted in greater predictive value?” Fig. 10. Examples of a deflection containing a high-frequency—usually representing non-dipolar information—component in which there is a notch, bead or slur in the down-stroke of the R-wave. The form of the disturbance depends upon the frequency characteristics of the electrocardiograph. If the electrocardiograph responds only to low frequencies, the notch becomes a bead or a slur.
Conflict of interest statement
turned on for an instant when it dominates the other, but then when the other is turned on, it becomes dominant. For two reasons neither myocardial infarction nor massive right ventricular hypertrophy occur in studies in safety pharmacology. First there are so few (or no) drugs which produce them. Second, the studies are of relatively short duration so there is insufficient time to develop massive hypertrophy. But of greatest importance, there never seems to have been an instance in which an electrocardiogram developed non-dipolar, high-frequency components.[This may be because they have never been sought!] Of course high-frequency components are manifested quite commonly as notches in the QRS complex (Flowers et al., 1969). However these occur most commonly because a vector is oriented nearly normal (i.e., at right angles) to the axis of the lead. Thus with only slight shifts in orientations of a single wave of depolarization, the wave may be directed either toward, away from, or at right angles to the electrode going to the positive pole of the electrocardiograph, and notches will occur. Such notches may become more obvious when conduction velocities are slowed as with INa blockers (i.e., class I antiarrhythmics). There have been only limited studies on the equivalent generator responsible for P waves or the ST-T (Spach & Barr, 1976). Frequently the P wave contains high-frequency components (notches, beads and slurs), and the ST-T contour may be of the ‘dome-dart” configuration,
References
The author serves as a paid consultant to many pharmaceutical houses and contract research organizations, but none of these influenced the contents of this manuscript.
Armoundas, A. A., Feldman, A. B., Mukkamala, R., & Cohen, R. J. (2000). Accuracy of a single equivalent moving dipole model in a realistic anatomic geometry torso model. Computers in Cardiology, 27, 439−442. Armoundas, A. A., Feldman, A. B., Mukkamala, R., He, B., Belk, P. A., Lee, Y. Z., et al. (2003, Dec). Statistical accuracy of a moving equivalent dipole method to identify sites of origin of cardiac electrical activation. IEEE Transactions on Biomedical Engineering, 50(12), 1360−1370. Bayley, R. H., & Schmidt, A. E. (1954). The problem of adjusting the Wilson central terminal to zero in the living human subject. Circulation Research, 3, 94−1101. Bayley, R. H., Reynolds, E. W., Kennard, C. L., & Head, J. F. (1954). The zero of potential of the electrical filed produced by the heart beat. The problem with reference to homogeneous volume conductors. Circulation Research, 2, 5−13. Brody, D. A. (1956). A theoretical analysis of intracavitary blood mass influence on the heart-lead relationship. Circulation Research, 4, 731−738. Bu, G., & Berbari, E. J. (2006). A single moving dipole model of ventricular depolarization. Biomedical Sciences Instrumentation, 42, 237−242. Burger, H. C., van Brummelen, A., & Herpen, G. (1961). Heart vector and leads. American Heart Journal, 61, 317−320. Corbin, L. V., & Scher, A. M. (1977). The canine heart as an electrocardiographic generator. Dependence on cardiac cell orientation. Circulation Research, 41, 58−67. Detweiler, D. K. (1988). The electrocardiogram: A critical review. In P. Q. MacFarlane & T. D. V. Lawrie (Eds.), Comprehensive electrocardiography: Theory and practice in healthy and disease (pp. 1267−1329). New York: Pergamon Press. Durrer, D., & Van der Tweel (1965). Excitation of the left ventricular wall of the dog and goat. Annals of the New York Academy of Sciences, 65, 779−803. Flowers, N., Horan, L., & Brody, D. (1965). Evaluation of multipolar effects in the highfidelity standard electrocardiogram by means of factor analysis. Circulation, 32, 273−280.
168
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Flowers, N. C., Horan, L. G., Thomas, J. R., & Tolleson, W. J. (1969). The anatomic basis for high-frequency components in the electrocardiogram. Circulation, 39, 531−547. Frank, E. (1952). Electric potential produced by two point current sources in a homogenous conducting sphere. Journal of Applied Physics, 23, 1225−1331. Frank, E. (1954). General theory of heart-vector projection. Circulation Research, 2, 258−263. Fukuoka, Y., Armoundas, A. A., Oostendorp, T. F., & Cohen, R. J. (2003). A single equivalent moving dipole model: An efficient approach for localizing sites of origin of ventricular electrical activation. Annals of Biomedical Engineering, 31, 564−576. Geselowitz, D. B. (1976). Determination of multipole components. In C. Nelson & D. Geselowitz (Eds.), The theoretical basis of electrocardiology (pp. 202−212). Oxford: Clarendon Press ch. 9. Grant, R. P., & Estes, E. H. (1951). Spatial vectorelectrocardiography. New York: The Blakiston Co., Inc. Hamlin, R. L., & Smith, C. R. (1960). Anatomical and physiologic basis for interpretation of the electrocardiogram. American Journal of Veterinary Research, 21, 701−708. Hamlin, R. L., Burton, R., Leverett, S., & Burns, J. (1974). The electrocardiogram from miniature swine recorded with the McFee-axial reference program. Journal of Electrocardiology, 7, 155−162. Holland, M. P., & Ornsdorf, M. F. (1977). Solid angle theory and the electrocardiogram: Physiologic and quantitative interpretations. Progress in Cardiovascular Diseases, 19, 431−437. Horan, L., Flowers, N., & Brody, D. (1964). The limits of information in the vectocardiogram: Comparative resynthesis of body surface potentials with different lead systems. American Heart Journal, 68, 362−369. Malik, M., & Batchvarov, V. N. (2000). Measurement, interpretation and clinical potential of QT dispersion. Journal of the American College of Cardiology, 36, 1749−1766. McFee, R., & Johnston, F. D. (1953). Electrocardiographic leads. I. Introduction. Circulation, 8, 554−562. Morganroth, J., & Gussak, I. (Eds.). (2004). Cardiac safety of noncardiac drugs: Practical guidelines for clinical research and drug development. Totowa, NJ: Humana Press.
Redfern, W. S., Carlsson, L., Davis, A. S., Kynch, W. G., MacKenzie, I., et al. (2003). Relationships between preclinical cardiac electrophysiology, clinical QT interval prolongation and torsade de pointes for a broad range of drugs: Evidence for a provisional safety margin in drug development. Cardiovascular Research, 58, 32−456. Scher, A. M. (1964). The sequence of ventricular excitation. American Journal of Cardiology, 14, 287−296. Scher, A. M., Young, A. C., & Meredith, W. M. (1960). Factor analysis of the electrocardiogram. Test of electrocardiographic theory: Normal hearts. Circulation Research, 8, 519−526. Spach, M. S., & Barr, R. C. (1976). Origin of epicardial ST-T wave potentials in the intact dog. Circulation Research, 39, 475−487. Stilli, D., Musso, E., Macchi, E., Manaca, C., Dei Cas, L., Vasini, P., et al. (1986). Body surface potential mapping in ischemia with normal resting ECG. Canadian Journal of Cardiology, 8, 519−526. Taccardi, B. (1990). Body surface mapping and cardiac electric sources. A historical review. Journal of Electrocardiology, 23(Suppl:150–3). Taccardi, B., Musso, E., & De Ambroggi, L. (1972). Current and potential distribution around an isolated dog heart. Proceedings of the Satellite Symposium of the 25th International Congress on Physiological Science (The electrical field of the heart) and the 12th Colloquium Vectorcardiographicum. Brussels: Presses Academiques Europeenes. Tilley, L. P. (1993). Essentials of canine and feline electrocardiography. Interpretation and treatment, 3rd edition. Baltimore: Lippincott/William & Wilkins. Tysler, M., Turzova, M., & Slavomira, F. (2004). Assessment of local repolarization changes using model based BSPM interpretation. Advances in Electrocardiography. World Scientific Publishing Co. Pte., Ltd. Wilson, F. N., & Bayley, R. H. (1950). The electric field of an eccentric dipole in a homogeneous spherical conducting medium. Circulation, 1, 84−93. Zhou, X., Yin, B. S., Nong, D. B., & Yu, D. K. (2002). Study on distribution of cardiac electrical field around the isolated heart of guinea pig. Di Yi Junyi Daxue Xuebao, 22, 130−131. Zipes, D. P., & Jalife, J. (2004). Cardiac electrophysiology 4th edition, From cell to bedside. Saunders: Philadelphia.
Contents lists available at ScienceDirect
Journal of Pharmacological and Toxicological Methods j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p h a r m t ox
Appraisal of state-of-the-art
How many ECG leads are required for in vivo studies in safety pharmacology? Robert L. Hamlin ⁎ The Ohio State University, Scientific Director, QTest Labs, 1900 Coffey Rd., Columbus, Ohio 43210, United States
a r t i c l e
i n f o
Article history: Received 20 August 2007 Accepted 13 March 2008 Keywords: Electrocardiography Equivalent generator Leads Methods
a b s t r a c t This article explains the principles of electrocardiography, and explains how it is used by Safety Pharmacology, with a focus on the requirement for multiple leads in Safety Pharmacology assessment. Electrocardiography as used in different disciplines (e.g., medicine, anesthesiology, physiology, and pharmacology/toxicology/safety pharmacology) has different requirements for the number of electrodes applied. Electrodes may be placed at an infinite number of points on the body, and voltages (electrocardiograms) may be registered between/ among them. However in safety pharmacology there is little evidence that more than 1—or at most 3—lead(s) is (are) required to provide all of the information that might be present using an infinite number. This is based upon (1) the biophysics of the heart as a generator of electrical potential/voltage, (2) the fact that most properties of electrophysiology affected adversely by drugs are expressed as changes in durations, and (3) experience. A single, unipolar lead (V3) recorded from the left sternal border at the 5th intercostal space possesses minimal artifact and large, stable deflections. This lead allows for accurate measurement of heart rate and rhythm, durations of component deflections (e.g., PQ, QRS, QT), and J-point deviation. A greater number of leads seldom or never yield additional information that detects liabilities. Commonly voltages recorded between the right thoracic and left pelvic limbs (lead II) provides information similar to lead V3, and lead II is easier to apply, and produces voltages with less artifact and similar to those in lead V3. A lead measuring the voltage between the left and right thoracic limbs (lead I) along with lead II allows for estimating orientation of vectors in the frontal plane, but knowledge of these vectors seldom or never indicates liability of a test article. © 2008 Published by Elsevier Inc.
1. Introduction
1.3. What are electrocardiographic leads?
1.1. Why is electrocardiography used in safety pharmacology?
Lead is used three times in electrocardiography. It stands for a position (e.g., V3, rV2) on the torso from which voltages are recorded. It stands for a wire extending from the patient to the electrocardiograph. It stands for a combination of electrodes between which voltages are measured. This discussion will be about points on the torso surface from which voltages should be registered.
Electrocardiography is a method essential for studies in safety pharmacology, because it detects, simply, inexpensively, and noninvasively, many (most) drug-induced changes in electrophysiological properties of the heart (e.g., chronotropy, dromotropy, bathomotropy) that might translate to morbidity or mortality if expressed in man.
1.4. What is an electrocardiograph and what are types of leads? 1.2. What is the electrocardiographic system? There are 4 components to the electrocardiographic system (Fig. 1): (1) the heart as a generator of voltage, (2) the body volume which allows for transmission of those voltages to the torso surface, (3) the voltmeter (electrocardiograph) which detects, amplifies and records the voltages, (4) the electrocadiographer who interprets the electrocardiograms (the recordings of the voltages) to determine the state of electrophysiological health of the heart, and who often makes recommendations for treatment or identifies toxicity.
⁎ Corresponding author. Fax: +1 614 292 3646. E-mail address: [email protected]. 1056-8719/$ – see front matter © 2008 Published by Elsevier Inc. doi:10.1016/j.vascn.2008.03.004
An electrocardiograph is a voltmeter which registers differences of electrical potential (the force that drives electrons from place to place). As such a voltmeter requires 2 inputs, termed positive (+) and negative (−). These inputs come from positions (leads) on the torso and/or limbs of the subject. The electrocardiograph measures voltage (potential to drive electrons) between these points, and of course there is an infinite number of points on the torso/limbs between which potential differences can be measured and might have meaning to the electrocardiographer. For example: (1) lead I measures the voltage difference between the left and right thoracic limbs, where the lead from the left thoracic limb goes to the positive pole of the electrocardiograph, and the lead from the right thoracic limb goes to the negative pole; (2) lead V3 measures the voltage difference between an electrode on the left portion of the thorax and a
162
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 1. Overview of the electrocardiographic system, including: the heart as a generator of potential, the body as a volume conductor, electrodes placed on the limbs or torso surface, leads (wires) extending from the electrodes to the voltmeter (electrocardiograph), the electrocardiograph, the electrocardiogram, the interpreter.
combination of electrodes from the right and left thoracic and the left pelvic limbs. The electrode from the thorax goes to the positive pole of the electrocardiograph and the 3 electrodes (combined) from the limbs go to the negative pole; (3) lead aVF measures the voltage difference between an electrode on the left pelvic limb and a combination of electrodes from the left and right thoracic limbs, where the lead from the left pelvic limb goes to the positive pole of the electrocardiograph and the 2 leads (combined and modified by high resistance) from the thoracic limbs go to the negative pole of the electrocardiograph. Voltage differences between only 2 points form so-called bipolar leads; voltage differences between 1 point and the combination of the 3 limbs form so-called unipolar leads; voltage differences between 1 point and the combination of points from the 2 other limbs form socalled augmented unipolar leads. 1.5. What determines the numbers of leads and which one(s) is (are) required? Of course the number of leads possible (permutations of inputs to the poles of the electrocardiograph) is infinite, therefore it is important to know how many, and which ones, are required to provide the information necessary for the mission of electrocardiography. That depends upon the mission. In clinical electrocardiography of humans (Morganroth & Gussak, 2004; Zipes & Jalife, 2004), to detect chamber enlargement, to localize myocardial ischemia and infarction, and to quantify dispersion of QTc among leads, 12 or more leads may be necessary. In conventional electrocardiography in veterinary clinical medicine (Detweiler, 1988; Tilley, 1993), 6 leads (of which only 2 or 3 of the 6 present information that is unique) are usually taken. But if the interest is only in determining heart rate, rhythm, durations of various electrophysiological processes (e.g., rate of discharge of the SA node, speed of conduction through the atria or ventricles or from atria to ventricles, durations of both ventricular depolarization and repolarization) or to search for abnormal discharge from within the atria or ventricles, then only 1 lead may be necessary (Malik & Batchvarov, 2000). But it better be a “good” lead, i.e., one from which deflections can be measured accurately (Horan, Flowers, & Brody, 1964)! The decision on the numbers of leads is based upon experience. However the decision may be based upon theoretical, biophysical
considerations as well; for example, how does the heart behave as an equivalent generator of potential? Of course the final decision on how many, and which, leads are necessary in studies on safety pharmacology may be answered only by the medical community and drug regulatory agencies who could state how often 1, 2, 3+….+n leads produced information that predicted safety with 100% sensitivity and 100% specificity. Since 100% sensitivity and specificity are impossible to achieve, rather we seek the fewest number of leads that produces results as good as the greatest number of leads ever used. 2. The heart as a generator of potential The voltages are produced on the body surface by waves of depolarization (Durrer & Van der Tweel, 1965; Scher, 1964) and repolarization traversing the atria and ventricles (Fig. 2). These waves are in fact sheets of dipoles (positive and negative charges separated by a small distance) in which the face of the sheet in stimulated (depolarized) myocardium is comprised of negative charges, while the face of the sheet in the resting (repolarized) portion of the myocardium is comprised of positive charges (Wilson & Bayley, 1950; Bayley et al., 1954). Fig. 3 shows a dog, its heart (albeit an abnormally large one) within the torso, and a single sheet of dipoles, frozen at an instant, traveling through the heart in a general direction from head to tail. The sheet of dipoles has negative signs in the region that is depolarized, and positive signs in the region that is yet to be depolarized. We may predict/estimate the magnitude and sign of the voltage produced by this wave at any point (P), say on left hind leg of the torso, using (1) the solid angle (Fig. 4) concept (Holland & Ornsdorf, 1977), (2) the vector (Fig. 5) concept (Frank, 1954; Grant & Estes, 1951), or (3) by gross approximation (Fig. 6) (Hamlin & Smith, 1960). Using the solid angle concept (Fig. 4), we construct radii from the open ends of the sheet to the point, measure the angle (Ω1) determined by the radii, and state that the sign will be positive because the point is in the general milieu of the positive face of the sheet, and that the voltage will be proportional to the magnitude of the angle. For mathematical correctness the magnitude of the angle should be multiplied by 2 constants, one dependent upon the resistive 1 In this example the angle is in one plane, whereas the boundary is actually 3 dimensional, therefore the angle should be a solid angle.
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
163
Fig. 2. A schematic diagram of the heart, showing the wave of depolarization emitted from the SA node, traveling first through the atria, and then through the ventricles (Fig. 2A). An ECG, shown to the right, has colors which match up to which waves of depolarization produce which voltages (Fig. 2B).
characteristics of the body, the other upon certain geometrical considerations (Bayley et al., 1954; Brody, 1956). The sheet of dipoles may be represented, also, by a single vector (the wide arrow in Fig. 5) whose length is proportional to the area of the sheet of dipoles and whose direction represents the direction the sheet is moving (Grant & Estes, 1951). The voltage, at a point on the torso (P), generated by the sheet of dipoles represented by the vector, is proportional to the projection of the vector on a lead axis (dotted line) constructed from the point on the torso through the origin of the vector or the center of the heart. The length of the projection is proportional to the cosine of the angle (θ) between the vector and the lead axis, times the magnitude (length) of the vector. For mathematical correctness the projection should be divided by the square of the distance between the vector and the point, and again multiplied by a constant which represents the conductive properties of the body. But it is assumed that the distance between the heart and the electrode is 1, therefore squaring the distance can be neglected. The sheet or vector may be represented by (resolved into) a single dipole—the so-called equivalent dipole. As the waves of depolarization
and repolarization (not considered here) traverse the myocardium, the sheets of dipoles change geometry (size and configuration) and direction of motion (Fig. 2), but at any and all times the sheet can still be represented by an equivalent dipole whose magnitude and orientation changes in concert with the changing sheet of dipoles, and that can be represented by a single vector whose length represents the magnitude of the equivalent dipole and whose spatial orientation represents the spatial orientation with which the sheets of dipoles traverses the torso (Armoundas, Feldman, Mukkamala, & Cohen, 2000; Armoundas et al., 2003; Bu & Berbari, 2006; Corbin & Scher, 1977; Frank, 1952; Fukuoka, Armoundas, Oostendorp, & Cohen, 2003). That being the case (i.e., that is if a single vector can represent the waves of depolarization and repolarization (Scher, Young, & Meredith, 1960; Stilli et al., 1986; Taccardi, 1990)) then it follows that both magnitude and orientation of that vector—and therefore of the sheets of dipoles—can be characterized unequivocally2 by 2 The vector may be characterized unequivocally, but there may not be a unique boundary.
164
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 3. A left-lateral view of a dog showing the heart and a wave of depolarization traversing the heart. The wave is a sheet of dipoles (separations of − and + signs), in which the face covered with − signs is in myocardium that has been depolarized, and the face with + signs is in the myocardium yet to be depolarized.
Fig. 5. The same dog and sheet of dipoles as in Fig. 3, but a vector is thrust through the sheet of dipoles. The length of the vector represents the size of the sheet of dipoles, and the direction it points is the direction the sheet is moving through the heart and torso. The amplitude of the voltage recorded from this lead may be predicted as the length of the projection of the vector on the lead axis (the imaginary line drawn from the point on the torso through the origin of the vector or the center of the heart). The magnitude of the projection of the vector is equal to the magnitude of the vector times the cosine of the angle (θ) between the vector and the lead axis.
3 leads (Fig. 7). In the case of this illustration, the 3 leads are orthogonal (i.e., they intersect at right angles and at the center of the heart) (Fig. 8). Leads I, aVF, and V10 could constitute the 3 leads required to characterize all so-called dipolar properties of the heart represented by an equivalent dipole (McFee & Johnston,1953; Bayley et al.,1954). In fact to estimate the sheets of dipoles most accurately, the orthogonal leads should be based upon electrodes equidistant from the heart, the heart should be in the center of the body, and the body volume should conduct uniformly. None of these requirements is fulfilled by any orthogonal lead system; however this limitation seldom compromises the clinical or toxicological usefulness of the characterization of the sheet of dipoles (Bayley et al., 1954; Flowers, 1964; Wilson & Bayley, 1950). If a vector is placed in a torso, any electrode on the torso surface may have a voltage with a sign of 0, + or −, and if + or − the magnitude may be from very small (b0.1 mV) to reasonably large (up to 5 mV). If the vector points towards an electrode, the sign (deflection) will be +; if it points away from the electrode, the sign will be −; if it points at right angles to the electrode, the sign will be 0 (i.e., there will be no deflection). A third method (Fig. 6) of predicting the magnitude and sign of a potential at a point on the body surface may be to imagine
yourself sitting at the point on the torso, looking at the heart, and determining whether the sheet is traveling toward you (+) or away from you (−), and how big the sheet is (the magnitude of the voltage). This approximation is semiquantitative but useful. A vector may be mimicked by a flashlight. Instead of producing a voltage on the torso surface, the flashlight produces either lightness or darkness. A positive voltage or a perception of light will appear at a point on the torso towards which a vector/flashlight points. If a single vector/flashlight represents the sheets of dipoles accurately (i.e., if the heart behaves as an equivalent dipole) (Fig. 9), then there can be only 1 set of + signs (maxima)/brightness and − signs (minima)/darkness on the torso surface (Scher et al., 1960; Taccardi, 1990). At the top of this figure are 2 boundaries of different sizes and traveling in different directions. Notice the lightness is greater for the big boundary than for the small boundary and that different regions on the torso are either light or dark depending upon where the vector/flashlight points. When both boundaries exist simultaneously (bottom of figure), both
Fig. 4. The same dog as in Fig. 3, but with a point on the left pelvic limb for which a voltage is to be predicted, and how the amplitude and sign of the voltage generated at the instant of the wave may be predicted by the magnitude of the angle (Ω) produced by radii constructed from the open ends of the boundary (sheet of dipoles).
Fig. 6. The same dog as in Fig. 3, but an imaginary person is sitting on the electrode and looking at the heart and sheet of dipoles. The voltage will be roughly proportional to the size of the boundary the person sees. If looking at the + face, the voltage will be +, if looking at the − face, the voltage will be −, and if the boundary is traveling at right angles to the person, no voltage will be present.
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
165
Fig. 7. An “ideal” ECG showing the 3 orthogonal leads (I, aVF, V10) and V3 at high paper speed, and lead aVF at slow paper speed. Measurements are more accurate from the trace at higher paper speed, but rhythm is observed more easily from lower paper speed.
flashlights are on, both illuminate regions, but the brighter one dominates the smaller one and the brightness it produces may only be observed when you are close to it—a so-called proximity effect or biased by proximity. Studies have been conducted on many species to validate that their hearts behave as an equivalent dipole. In some of these studies unipolar electrocardiograms were recorded from hundreds of points on the torso surface (Taccardi, Musso, & De Ambroggi, 1972). At many instances during the QRS complex, maxima and minima were sought. At each instance only a single maximum and minimum occurred, but of course the magnitudes and positions of these maxima and minima varied from time to time, as the magnitudes, geometry and locations of sheets of dipoles varied within the heart. Thus a single vector (i.e., one flashlight) with changing magnitude/brightness and orientation/ direction could explain all of the electrocardiographic information present in the hundreds of leads. Had 2 vectors been necessary at any instant, then the heart would have behaved as an equivalent quadrupole; there would have been 2 sets of maxima and minima, and more than 3 leads would have been required to characterize the 2 vectors. Again with reference to Fig. 9, bottom, this is illustrated by 2
flashlights, one bright and one dim, representing two waves of depolarization traveling in different directions. Although in general the larger flashlight dominates and makes the hind end of the dog light and the front end dark, the smaller flashlight has a local “contribution” and produces a dim light in the darkness and slight darkness in the lightness. 3. Which 3 leads are necessary? Although any 3 leads would be satisfactory to characterize the magnitude and orientation of a vector, it is easier (more intuitive) for an electrocardiographer to identify a vector if the 3 leads are mutually orthogonal, (i.e., intersect at right angles). One lead (lead I) determines if the vector points to the left or right; 1 lead (aVF) determines if the vector points to the head or tail (inferiorly in man); 1 lead (V10) determines if the vector points dorsad (posteriorly in man) or ventrad (anteriorly in man) (Fig. 8). Any lead other than those 3 can contribute nothing further to characterizing the vector (or sheets of dipoles), therefore any other lead may be considered redundant and superfluous, and serves as an unnecessary distraction to the electrocardiographer (Hamlin, Burton, Leverett, & Burns, 1974). 4. What are the exceptions (there always are some)?
Fig. 8. The orthogonal axes with a vector at the origin, and projections of the vector on each axis. The voltage recorded on each axis is proportional to the projection of the vector on that axis. Of course this is idealized, since at the very least electrodes seldom form truly orthogonal axes, the heart is not truly in the center of the axial system (i.e., it may be closer to one axis than to another), and there are serious inhomogeneities in the volume conductor which skew the projections.
Potential for non-dipolar properties and leads biased by proximity constitute possible exceptions (Flowers, Horan, & Brody, 1965; Flowers, Horan, Thomas, & Tolleson, 1969; Zhou, Yin, Nong, & Yu, 2002). Although 3 leads may be all that is necessary (theoretically) (Burger, van Brummelen, & Herpen, 1961) in practice some leads placed on the thorax close to the heart have deflections from which measurements can be made more easily and accurately. For example, the unipolar thoracic lead, V3, may appear qualitatively similar to lead aVF, however the end of the T wave is much less ambiguous (Fig. 7) in that is it terminates in the isoelectric (base) line rather sharply rather than fusing into the baseline gradually (Detweiler, 1988). Since the QT interval is so important in studies in safety pharmacology, it is obvious that lead V3 may be a welcome addition to the electrocardiographer who may not need it for characterizing vectors but who may find measurements of QT more precise when it is used. In addition there are some studies which claim significant nondipolar properties to the equivalent cardiac generator. Still these
166
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Fig. 9. A left-lateral view of a dog with its heart within and with electrodes placed at 3 points on the torso. Waves of depolarization are represented as flashlights pointing in the direction the waves are traveling. The voltages registered at each point will be mimicked by the relative brightness (if the flashlight points towards the point) or darkness (if the flashlight points away from the point). In the upper left panel, a single boundary/flashlight travels/points to the head and to the back, and the region toward which it travels/points becomes light whereas the region away from which it is pointing becomes black, the relative intensity related to the square of the distance between the source and the electrode. Because the boundary/flashlight is so small, no or very small voltages/lightness or darkness will be present at electrodes aVF or VN, but only at VP because it is so close. In the upper right panel showing a much larger/more intense boundary/light source, now voltage/lightness or darkness will be observed at aVF and VN, albeit greater at VP because of its proximity. In the bottom panel, both boundaries/flashlights exist simultaneously. The larger/brighter dominates the electrodes placed at a distance (aVF, VN), but the smaller dominates the single electrode (VP) so close to it. Thus although electrode VP is in the region that should contain darkness based upon darkness at VN and lightness at aVF, because of its proximity to the smaller flashlight and it being dominated due to that proximity, light will be observed. This is a so-called proximity effect, and can only exist when the heart is not represented as a single dipole (flashlight) that changes amplitude and orientation.
studies do not demonstrate superiority, over assuming that the heart is a dipolar source, in detecting either disease or drug effects. However there may be an even more important requirement for additional leads…if the cardiac generator is not only dipolar, but possesses multipolar (quadrupolar, octapolar) properties as well. If the cardiac activation process, at an instant, has 2 waves of depolarization, 1 large the other small, and 1 oriented spatially differently from the other, these 2 waves may be represented as 1 wave (or vector) when “viewed” from a relatively great distance, i.e., from electrodes more than 3 heart's diameter from the heart. However the larger wave will dominate the observation, so information about the smaller wave will be lost. It would be more accurate to express these 2 waves using 2 vectors. If each vector produces a single maximum and a single minimum, then 2 vectors will produce 2 maxima and 2 minima. Thus within a zone in which a positive deflection would be produced by the major vector, a negative deflection may be attributable to the smaller vector which influences electrodes placed close to smaller wave. Let us return to the analogy between a vector and a flashlight. A very weak (dim) flashlight is placed within the thorax and points dorsad and craniad (Fig. 9, top left). Lightness is observed any place on the torso towards which the light is pointing; darkness any place away from which the light is pointing. The further from the light source, the less bright, or dim, the light appears. Thus an electrode placed at VP will observe light, whereas electrodes at aVF or aVN will not be influenced. A very strong (bright) flashlight is placed within the thorax and points caudad (Fig. 9, top right). As with the dim flashlight, lightness is observed any place on the torso towards which the light is pointing, darkness is observed any place away from which the light is pointing, and the further from the light source, the less bright, or dim, the light appears. Thus electrodes placed at VP and aVN will observe darkness, whereas the electrode placed at aVF will observe light. Now
if both flashlights are activated together (i.e., both are turned on) (Fig. 9, bottom), electrodes at aVF and aVN will be influenced only by the brighter flashlight, whereas the electrode at VP is influenced, still, by the dimmer one; it is “biased by proximity.” Each flashlight mimics a dipolar source, and if you only measured light from a distance, the dimmer light source is relatively inconsequential and you will claim a dipolar source, but if you measure light from the distant point and from close to the dim source, you will identify a quadrapole. Non-dipolar sources usually contribute so-called high-frequency components in the electrocardiograms (Geselowitz, 1976; Tysler et al., 2004). These high-frequency components are identified as beads (Fig. 10A), slurs (Fig. 10B) or notches (Fig. 10C). Because by definition the deflections contain high-frequencies, which form is manifested on an electrocardiogram depends upon the frequency characteristics of the recorder. If the frequency content of a high-frequency component is 80 Hz and an electrocardiograph records only to 40 Hz, then a highfrequency component will either not be registered or it will be registered, as a form fruste, as a notch, bead, or slur. Non-dipolar sources generating high-frequency components to the QRS complex may occur with myocardial infarction or right ventricular hypertrophy. With right ventricular hypertrophy there appears to be 2 separate generators competing for dominance of the vector when both ventricles are activated nearly simultaneously. Without right ventricular hypertrophy, the left ventricle is normally so dominant that activation of the right ventricle is trivial, however, with right ventricular hypertrophy the dominance might shift from one ventricle to the other, so the vector shifts, rapidly, from endocardium to epicardium of the left ventricle, to endocardium to epicardium of the right ventricle, back to the left ventricle, and so on. Notice that both ventricles are nearly equal in mass. Using the flashlight analogy, with right ventricular hypertrophy, it is as if there are 2 flashlights, one pointing through each ventricle. Each is
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
167
particularly in certain leads. These imply more than dipolar sources, but, again, there is no data supporting the contention that multiple leads are necessary to identifying their presence or that they become toxicologically relevant in studies of safety pharmacology. 5. A recommended number of ECG leads for safety pharmacology. Does one need all 3 leads? All 3 leads are necessary in order to characterize the vector precisely. However there appears to be no instances in studies in safety pharmacology and/or toxicology where characterization of the vector has lead to identification of a toxic potential. Characterization of the vector may be necessary for clinical reasons, to detect and even to semiquantify chamber enlargement, but not to identify toxic potential of a drug. 6. What leads have proven useful? Given that characterization of the vector appears to be unnecessary to predict toxic potential, what is necessary? These measurements are periodicity of the heart (P–P interval), rhythm (sequences of depolarization and repolarization), atrioventricular conduction (PQ interval), duration of ventricular conduction (QRS duration), duration of ventricular repolarization (QT interval and QT interval corrected for heart rate), and rarely J-point deviation. These measurements can be made using a single lead (usually unipolar thoracic lead V3) (Bayley and Schmidt, 1954), and should determine potential toxic effects on the important electrophysiological properties of cardiac function (i.e., chronotropy, dromotropy, bathmotropy) which, if altered, may translate into morbidity and/or mortality. We must await information from a coalition of regulatory agencies, the pharmaceutical industry, and academia to document instances when more than a single lead has been more predictive of toxic potential than a single lead (Malik & Batchvarov, 2000; Morganroth & Gussak, 2004; Redfern et al., 2003). Then we must ask, “Which lead, and what feature(s) of the multiple leads resulted in greater predictive value?” Fig. 10. Examples of a deflection containing a high-frequency—usually representing non-dipolar information—component in which there is a notch, bead or slur in the down-stroke of the R-wave. The form of the disturbance depends upon the frequency characteristics of the electrocardiograph. If the electrocardiograph responds only to low frequencies, the notch becomes a bead or a slur.
Conflict of interest statement
turned on for an instant when it dominates the other, but then when the other is turned on, it becomes dominant. For two reasons neither myocardial infarction nor massive right ventricular hypertrophy occur in studies in safety pharmacology. First there are so few (or no) drugs which produce them. Second, the studies are of relatively short duration so there is insufficient time to develop massive hypertrophy. But of greatest importance, there never seems to have been an instance in which an electrocardiogram developed non-dipolar, high-frequency components.[This may be because they have never been sought!] Of course high-frequency components are manifested quite commonly as notches in the QRS complex (Flowers et al., 1969). However these occur most commonly because a vector is oriented nearly normal (i.e., at right angles) to the axis of the lead. Thus with only slight shifts in orientations of a single wave of depolarization, the wave may be directed either toward, away from, or at right angles to the electrode going to the positive pole of the electrocardiograph, and notches will occur. Such notches may become more obvious when conduction velocities are slowed as with INa blockers (i.e., class I antiarrhythmics). There have been only limited studies on the equivalent generator responsible for P waves or the ST-T (Spach & Barr, 1976). Frequently the P wave contains high-frequency components (notches, beads and slurs), and the ST-T contour may be of the ‘dome-dart” configuration,
References
The author serves as a paid consultant to many pharmaceutical houses and contract research organizations, but none of these influenced the contents of this manuscript.
Armoundas, A. A., Feldman, A. B., Mukkamala, R., & Cohen, R. J. (2000). Accuracy of a single equivalent moving dipole model in a realistic anatomic geometry torso model. Computers in Cardiology, 27, 439−442. Armoundas, A. A., Feldman, A. B., Mukkamala, R., He, B., Belk, P. A., Lee, Y. Z., et al. (2003, Dec). Statistical accuracy of a moving equivalent dipole method to identify sites of origin of cardiac electrical activation. IEEE Transactions on Biomedical Engineering, 50(12), 1360−1370. Bayley, R. H., & Schmidt, A. E. (1954). The problem of adjusting the Wilson central terminal to zero in the living human subject. Circulation Research, 3, 94−1101. Bayley, R. H., Reynolds, E. W., Kennard, C. L., & Head, J. F. (1954). The zero of potential of the electrical filed produced by the heart beat. The problem with reference to homogeneous volume conductors. Circulation Research, 2, 5−13. Brody, D. A. (1956). A theoretical analysis of intracavitary blood mass influence on the heart-lead relationship. Circulation Research, 4, 731−738. Bu, G., & Berbari, E. J. (2006). A single moving dipole model of ventricular depolarization. Biomedical Sciences Instrumentation, 42, 237−242. Burger, H. C., van Brummelen, A., & Herpen, G. (1961). Heart vector and leads. American Heart Journal, 61, 317−320. Corbin, L. V., & Scher, A. M. (1977). The canine heart as an electrocardiographic generator. Dependence on cardiac cell orientation. Circulation Research, 41, 58−67. Detweiler, D. K. (1988). The electrocardiogram: A critical review. In P. Q. MacFarlane & T. D. V. Lawrie (Eds.), Comprehensive electrocardiography: Theory and practice in healthy and disease (pp. 1267−1329). New York: Pergamon Press. Durrer, D., & Van der Tweel (1965). Excitation of the left ventricular wall of the dog and goat. Annals of the New York Academy of Sciences, 65, 779−803. Flowers, N., Horan, L., & Brody, D. (1965). Evaluation of multipolar effects in the highfidelity standard electrocardiogram by means of factor analysis. Circulation, 32, 273−280.
168
R.L. Hamlin / Journal of Pharmacological and Toxicological Methods 57 (2008) 161–168
Flowers, N. C., Horan, L. G., Thomas, J. R., & Tolleson, W. J. (1969). The anatomic basis for high-frequency components in the electrocardiogram. Circulation, 39, 531−547. Frank, E. (1952). Electric potential produced by two point current sources in a homogenous conducting sphere. Journal of Applied Physics, 23, 1225−1331. Frank, E. (1954). General theory of heart-vector projection. Circulation Research, 2, 258−263. Fukuoka, Y., Armoundas, A. A., Oostendorp, T. F., & Cohen, R. J. (2003). A single equivalent moving dipole model: An efficient approach for localizing sites of origin of ventricular electrical activation. Annals of Biomedical Engineering, 31, 564−576. Geselowitz, D. B. (1976). Determination of multipole components. In C. Nelson & D. Geselowitz (Eds.), The theoretical basis of electrocardiology (pp. 202−212). Oxford: Clarendon Press ch. 9. Grant, R. P., & Estes, E. H. (1951). Spatial vectorelectrocardiography. New York: The Blakiston Co., Inc. Hamlin, R. L., & Smith, C. R. (1960). Anatomical and physiologic basis for interpretation of the electrocardiogram. American Journal of Veterinary Research, 21, 701−708. Hamlin, R. L., Burton, R., Leverett, S., & Burns, J. (1974). The electrocardiogram from miniature swine recorded with the McFee-axial reference program. Journal of Electrocardiology, 7, 155−162. Holland, M. P., & Ornsdorf, M. F. (1977). Solid angle theory and the electrocardiogram: Physiologic and quantitative interpretations. Progress in Cardiovascular Diseases, 19, 431−437. Horan, L., Flowers, N., & Brody, D. (1964). The limits of information in the vectocardiogram: Comparative resynthesis of body surface potentials with different lead systems. American Heart Journal, 68, 362−369. Malik, M., & Batchvarov, V. N. (2000). Measurement, interpretation and clinical potential of QT dispersion. Journal of the American College of Cardiology, 36, 1749−1766. McFee, R., & Johnston, F. D. (1953). Electrocardiographic leads. I. Introduction. Circulation, 8, 554−562. Morganroth, J., & Gussak, I. (Eds.). (2004). Cardiac safety of noncardiac drugs: Practical guidelines for clinical research and drug development. Totowa, NJ: Humana Press.
Redfern, W. S., Carlsson, L., Davis, A. S., Kynch, W. G., MacKenzie, I., et al. (2003). Relationships between preclinical cardiac electrophysiology, clinical QT interval prolongation and torsade de pointes for a broad range of drugs: Evidence for a provisional safety margin in drug development. Cardiovascular Research, 58, 32−456. Scher, A. M. (1964). The sequence of ventricular excitation. American Journal of Cardiology, 14, 287−296. Scher, A. M., Young, A. C., & Meredith, W. M. (1960). Factor analysis of the electrocardiogram. Test of electrocardiographic theory: Normal hearts. Circulation Research, 8, 519−526. Spach, M. S., & Barr, R. C. (1976). Origin of epicardial ST-T wave potentials in the intact dog. Circulation Research, 39, 475−487. Stilli, D., Musso, E., Macchi, E., Manaca, C., Dei Cas, L., Vasini, P., et al. (1986). Body surface potential mapping in ischemia with normal resting ECG. Canadian Journal of Cardiology, 8, 519−526. Taccardi, B. (1990). Body surface mapping and cardiac electric sources. A historical review. Journal of Electrocardiology, 23(Suppl:150–3). Taccardi, B., Musso, E., & De Ambroggi, L. (1972). Current and potential distribution around an isolated dog heart. Proceedings of the Satellite Symposium of the 25th International Congress on Physiological Science (The electrical field of the heart) and the 12th Colloquium Vectorcardiographicum. Brussels: Presses Academiques Europeenes. Tilley, L. P. (1993). Essentials of canine and feline electrocardiography. Interpretation and treatment, 3rd edition. Baltimore: Lippincott/William & Wilkins. Tysler, M., Turzova, M., & Slavomira, F. (2004). Assessment of local repolarization changes using model based BSPM interpretation. Advances in Electrocardiography. World Scientific Publishing Co. Pte., Ltd. Wilson, F. N., & Bayley, R. H. (1950). The electric field of an eccentric dipole in a homogeneous spherical conducting medium. Circulation, 1, 84−93. Zhou, X., Yin, B. S., Nong, D. B., & Yu, D. K. (2002). Study on distribution of cardiac electrical field around the isolated heart of guinea pig. Di Yi Junyi Daxue Xuebao, 22, 130−131. Zipes, D. P., & Jalife, J. (2004). Cardiac electrophysiology 4th edition, From cell to bedside. Saunders: Philadelphia.