Body Surface Low-Level Potentials During Ventricular

Transcription

1 Body Surface Low-Level Potentials During Ventricular Repolarization with Analysis of the ST Segment Variability in Normal Subjects MADISON S. SPACH, M.D., ROGER C. BARR, PH.D., D. WOODRow BENSON, M.D., PH.D., ABE WALSTON, II, M.D., ROBERT B. WARREN, PH.D., AND SAM B. EDWARDS, M.D. SUMMARY Early ventricular repolarization was examined by measuring body surface maps in normal subjects, ages 1-60 years. Repolarization began with positive potentials. The time of appearance of these potentials from the onset of QRS increased with age. All normal subjects had positive ST segments on the anterior torso. Although the magnitude of the maximal ST-segment voltage varied from subject to subject, the maximal ST-segment voltage increased from early childhood into the twenties and thereafter decreased with age. The same age-related trend occurred during other phases of ventricular repolarization, i.e., the T wave and the U wave. Although the ST-segment maxima occurred in a circumscribed area near the heart in all adults, the normal ST-segment patterns varied in relation to the vertical distance over which low-level positive potentials extended on the left anterior torso. The ST-segment patterns differed greatly in the low-level potentials in areas distant to the heart. The ST-segment patterns indicate that the arrangement of ST-segment "current sources" is detectably different in normal adults. This variability suggests that the evaluation of ventricular ischemia might be enhanced by interpreting a questionably abnormal map (e.g., during exercise) compared with the patient's own baseline map. These results emphasize the importance of measuring the total cardiac electrical field on the body surface to account for the effects of the currents from all areas of the ventricles during the ST segment, rather than characterizing primarily the proximity effects as dohe with precordial maps. THE PRECORDIAL ST-SEGMENT elevation mapping method of Maroko and associates' (summation of precordial positive potentials at one instant) has stimulated clinical interest in various mapping methods as an extension of standard scalar ECGs. The value and limitations of this method were recently reviewed by Muller, Maroko and Braunwald.2 They note that this precordial method is of use primarily in patients who have ischemia or infarction located in two specific ventricular areas - the anterior and lateral left ventricle. Just before this technique was introduced, Reid et al.3 showed with precordial potential distributions (not summation maps) that precordial mapping gave variable results, depending on the site of the acute infarction. The rationale for mapping in the circumscribed area of the precordium is based on the theory that local myocardial electrical events are reflected in the "proximity potentials," as noted by Mirvis et al.4 in their recent exercise studies with left precordial isopotential maps. Most precordial ST-segment mapping studies have evaluated ventricular ischemia.' Since little attention is given to the normal background potential distributions, it is usually implied in the analysis that the ECG abnormality due to ischemia is judged against a background state where the ST segment is at zero From the Departments of Pediatrics, Physiology, Biomedical Engineering, and Medicine, Duke University School of Medicine, Durham, North Carolina. Supported in part by USPHS grants HL 11307, HL 05716, HL 07101, and a grant from the National Foundation of the March of Dimes. Address for reprints: Madison S. Spach, M.D., Box 3090, Duke University Medical Center, Durham, North Carolina Received March 10, 1978; revision accepted November 7, Circulation 59, No. 4, potential. In a theoretical discussion Fozzard and DasGupta5 noted that "the ST segment is usually isoelectric because the ventricular cells come to nearly the same depolarized potential, and current does not flow." On the other hand, ECG analyses indicate that there should be considerable current flow during the ST segment in some normal adult subjects whose precordial leads may have ST-segment elevation of 0.3 mv.6-10 Although the clinical use of precordial mapping is increasing, an essential question is unanswered: What is the control underlying state of the cardiac sources before the abnormal ischemic cardiac source is superimposed on them? Measurements of epicardial and body surface potential distributions in intact chimpanzees" and dogs'2 13 showed that during the ST segment of normal beats there was a single body surface maximum, while at the same time on the epicardium there were either single or multiple maxima, and the magnitudes of the "normal" epicardial positive ST segments were considerably greater than the 0.1 mv value used in acute dog experiments "to reflect the extent of myocardial ischemic injury."' A possible explanation for this discrepancy is that the baseline status of ventricular repolarization potentials is different in acute experimental animals' with air in the chest (more negative epicardial ST-T waves) than in chronic healthy intact animal preparations"l-'3 (more positive epicardial ST-T waves). The above emphasizes the paucity of associated studies concerning the basic problem of the relationship between epicardial and body surface potential distributions in the presence of ventricular ischemia. Holland and associates'1416 used the solid angle theorem to describe complex functions of the geometry of the ischemic muscle to interpret abnor-

2 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. 823 mal ST-T-segment scalar ECG deflections resulting from ischemia. However, the solid angle theorem is difficult to use for the analysis of many ST-T waves because the basic assumption17 used with it, i.e., that the transmembrane potential is the same within the area of the ventricle that defines the boundary used to compute the solid angle, is not fulfilled.'8 Only recently have any qualitative", 19 or quantitative20' 21 analyses explained the relationship between known ST-T-wave epicardial and body surface potential distributions. Inverse calculation studies by Barr and Spach2' of epicardial potential distributions in intact dogs indicate that a workable solution probably can be extended to humans for relating body surface potential distributions to those on the epicardium. The use of epicardial potential distributions as an index of cardiac electrical activity assumes significance, since they can be related in a detailed way to the underlying intramural potentials,'2 22 and they can be computed quantitatively from intracellular potential distributions during the ST segment and T wave (SI theory).'8 Since ST-segment epicardial potential distributions are determined by the spatial distribution of intracellular potentials throughout the ventricle, these background distributions affect the patterns of current flow during the ST segment when varying degrees of ischemia occur at specific locations in the ventricle. Therefore, it is important to know if there are variations during the ST segment in the body surface potential distribution in normal subjects. If present, they would provide insight to the normal variability of the spatial distribution of intracellular potentials during the ST segment. Although variability in the magnitude of positive ST segments in healthy subjects has long been realized and emphasized in scalar electrocardiography,8-10 only a few ST-segment surface maps in humans23-27 are available to form a basis for interpreting ST-segment potentials in terms of their origin. In this report we consider different phases of ventricular repolarization in normal subjects as a reflection of the major components of ventricular action potentials,28 with emphasis on the spatial analysis of ST-segment potentials. The interpretations of the STsegment body surface patterns are discussed in terms of different cardiac sources'8 (different spatial distribution of intracellular potentials) in normal hearts during the plateau phase of ventricular action potentials. These interpretations agree with the inverse STsegment epicardial potential distributions that were calculated2' from the surface maps; the inverse epicardial patterns in these subjects were consistent with those measured"' 12 and computed21 in intact experimental animals. Methods The same 66 normal subjects between 1-60 years old whose scalar deflections were analyzed in the accompanying report29 were evaluated in this study. Their waveforms served as a check that the low-level repolarization potentials were accurately represented in the maps. Each subject's body surface maps were the basis for analysis of the following: 1) the time from the onset of QRS to the appearance of ventricular repolarization potentials, 2) the magnitudes of the voltages during the various phases of repolarization, and 3) the variability in the patterns of the STsegment potential distributions. For this study we chose to study the ST segment in the form of T.(0-50 msec) maps,29 a map representing the average of the potentials during the early ST segment, i.e., during the first 50 msec of the ST segment after ventricular excitation potentials disappeared. Clinical Information In the normal group, there were 59 males and seven women; two of the women were 42 and 46 years old, and five were years old. The major criterion of normality was a state of good health and well-being and no evidence of cardiovascular disease; i.e., none had a history of known or suspected coronary artery disease, hypertension or peripheral vascular disease. All subjects older than 22 years had had a physical examination within 1 year with a normal chest x-ray, a normal routine ECG and a blood pressure that did not exceed 140/90 mm Hg. None of the subjects were taking medications. The surface maps were recorded with each subject in the supine position and at a basal heart rate that varied from beats/min for subjects older than 5 years and from beats/min in the younger children. Recording Methods The details of the recording methods, inspection of the waveforms, and procedures for constructing the maps have been presented in detail previously.29 The methods emphasized the accurate depiction of lowlevel potentials (less than 250 AV). The body surface maps were recorded with an array of 150 electrodes, and the waveforms were recorded in subsets of 20 each with an additional time reference trace. All potentials were measured in reference to Wilson's central terminal. The baseline was chosen during the UP interval (TP interval in young children) in all of the waveforms. The maps were displayed automatically, the isopotential lines drawn, and each map photographed in a movie format. Each patient's maps were viewed millisecond-by-millisecond during late QRS to detect the instant when ventricular excitation potentials disappeared. Once this instant was identified, the ST (0-50 msec) map for each subject was constructed by reprocessing the digitally stored map data to obtain the average potential at each of the 150 positions during the next 50 msec. Results First, we will review body surface patterns associated with the appearance of ventricular repolarization potentials, and then analyze the appearance time of these repolarization potentials in the normal subjects. Each phase of ventricular

3 824 CIRCULATION VOL 59, No 4, APRIL 1979 (I) G-800,uV 0~~~~~~~~~160,4V 0)-64#V 0+ 16O0AiLV ;Al,-B2 m -3 JI -4 52am& 0 ''f-" ~~~~~~~ ;-20 *.3 t \ ' -X0z L:.6.1.2~~~~~~L L *10.13~~ S -*f ~~~~ ~,3 7.I..lIlo.4* It ~ ~ ~ ~ ~ ~ , msec (3 140,u.V (2) 0~~~~~~~()70,uV OII10 70 msec kv 0-120okVO aoiv ( Rep.) e % l0-2 - /,-o z x \'Z2 -,- -:4-2~~. -i j 0-2 *f msec (3) 0B-50uV 0+40kiV -1 1, Z ~ ~ 1 2 2~~~~~ l2 4 F.-2 fl 2 l. 6 I II 0 I --I Z I.0~ 945 msec Type 0- Au %- -.z -0.7 I. 0-30kuV 78 msec 0~130M.v L -2 Nr ~~~~~.3.2.~~~~~ i msec 320kV o-M.14%4* ~ & k kuv msec Type B

4 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. 825 repolarization is characterized in the form of the maximal voltage values during the ST-T-U wave for each normal subject. Then the ST segment is analyzed for the extrema and for the general low-level potential patterns in the ST (0-50 msec) maps of the total normal group. The analysis procedure is described for each type of evaluation and, when appropriate, the results of others are noted for comparison. Appearance of Ventricular Repolarization Potentials in Normal Subjects In Taccardi's surface maps in normal adults23 the patterns became complex during the last 30 msec of QRS because 1) multiple excitation maxima frequently were present, and 2) a repolarization maximum developed either in the midsternal or left axillary area during the last 5-10 msec. To learn if there were variations in the way ventricular repolarization potentials appear and increase thereafter, we used the criteria of Taccardi23 for defining excitation maxima and minima as those which develop during QRS and then disappear before or with the onset of the ST segment and assigned potentials to repolarization as those which evolve late in QRS, persist, and continue in the ST segment. These criteria can be justified on an experimental basis from simultaneous epicardial and body surface potential measurements in intact animals."' 13, 19, 21 We observed at least four types of changing potential distributions during late QRS as the patterns changed from pure excitation potentials during late QRS to pure repolarization potentials at the J point. In all subjects repolarization appeared in the form of positive potentials, as would be expected from epicardial potential distributions.", 18 The types of changing patterns are illustrated in figures 1 and 2 for each of four subjects to emphasize the variable, rather than constant, types of changing patterns encountered in normal subjects during late QRS. The major differences in the four types of sequential change presented in figures 1 and 2 involved 1) the initial background state of late ventricular excitation (row 1) with either one or two widely separated excitation maxima and the subsequent order of their disappearance, and 2) the positive potentials of repolarization (row 2) initially developed in locations that varied from the lower torso to the top of the chest at the upper sternum and then shifted to the left precordium as indicated by the arrows in row 2. The patterns at the time of the J point (row 3) and during the ST segment (row 4) did not demonstrate a difference as striking from subject-to-subject as the variability during late QRS. The first type was characterized by minimal overlap of repolarization and excitation potentials (fig. 1, type A). There was a single excitation maximum located low on the back in association with a more prominent excitation minimum on the anterior precordium (A, 1 and 2). This pattern, presumably due to excitation at the lateral-inferior left ventricle, remained stable until the last 2-6 msec of QRS. At this time low-level positive potentials due to repolarization developed on the lower left precordium (A,3). They increased on the anterior torso with the development of an ST-segment maximum in the left parasternal area and a low-level minimum in the back (A,4). This pattern occurred in 10 subjects, seven of whom were older than 50 years; the remaining three were years old. In the second pattern (fig. 1, type B) there were two excitation maxima during late QRS, one at the upper sternum and the other on the middle of the back, with a minimum on the anterior precordium (B,1). This is the typical late QRS pattern described by Taccardi.23 Presumably, the two maxima were due to wavefronts in the posterior lateral left ventricle and in the upper right ventricular area. The upper sternal excitation maximum persisted while the one on the back disappeared. Simultaneously positive potentials resulting from repolarization extended across the lower left axilla with the rapid development of a prominent left precordial repolarization maximum (B,2). The repolarization maximum was associated with waveforms in that area that had a prominent notch at the end of QRS, as illustrated in the waveform beneath the maps of figure 1, type B. The repolarization maximum developed in an area which had been occupied 20 msec earlier by a prominent excitation FIGURE 1. Appearance of ventricular repolarization potentials in two normal subjects. The sequential changes in the body surface potential distributions during late QRS and the ST segment are shown. Each subject's series ofmaps represents one offour types ofchanges that occurred during late QRS. Thefirst map (1) of each subject represents a time instant late in QRS before positive repolarization potentials appeared. The second map (2) shows the pattern produced when repolarization potentials (Rep) appeared. The third map (3) is the instant immediately after the disappearance ofexcitation potentials (J point of the map). The fourth instant (4) is the ST-segment potential distribution 50 msec after the J point. The isopotential lines were drawn automatically according to a logarithmic scale to represent progressively the following values in microvolts: 40, 60, 100, 150, 250, 400, 600 and A log separation was considered beneficial, since it accentuated the details of the low-level potential changes, without loss of detail near the extrema. The thick dashed line is the zero reference potential line. The potential value at each recording point is displayed in tens of microvolts. Although the individual values are difficult to see, the numbers serve as an indicator of the 150 electrode positions. The column of values on the right side of the back is duplicated at the lateral borders of the map. The time instant represented by each map is indicated by the vertical line superimposed on the waveform associated with each map. F = Front; B = Back. The values of the maxima (plus sign) and minima (minus sign) are indicated above each map. The peaks of the QRS and T wave deflections on the scalar reference trace are truncated.

5 826 (I)-.22 -I: ( ) 420,aV ,uV =' s2-.p -6-t. *1} -- -@ :I @ -H 19 -:1 -> -to -it ': -Lo -II-X 65 msec () 140,uV -120,iiV -I4O M I100 p.v(rep) -3 *3 o= ~~F 72 *- *, -* ~~~~~~~~-tt -'*@,sj *. 1.2 *Z *] CIRCULATION VOL 59, No 4, APRIL 1979 I zy~~~~~~~~~~~~~~-i of 2 o) I0o4V A (0)-510A1V 74 msec 120,uV -60,(p ) 70,uV ( Rep.) - '3263 * - (0313 0li.V 2.! 4 *) '* *s *Z>=,> *@ F g. *1 _ ~~~~~-1-2 *8 "s I 2-.~ -_Ss2 ;-E _s-ilf*0 )-: *2.2 2-I : 4X ss -2 -: * * ==I *2 * ** X2 I;1 K S ;.... s * * * 2 *2 *1 *2 -*la S = X0 msec ,uV ( 3) 0S)- 70.LV 2 * ] * * > \ 3 *^- *6 21* : * * *1 * ( &.2 z >, - #- * s e t.-. 2 t -2L _ * -, ^ S A *f Z - N4 =.2 0:9 78 msec (4) (i)~-70,av L * /.-% *e ) s -\ * \ 2 - O-30,AV 93 msec 90QV B * 3 '2 *2 * I-I -2-2 ; *: *; *3 ' -. * *.3.2..: 3.2 *1 1 3 I I I 97 msec (j) 1 70 AlV -Il 3 I a e -J t I I I ,2V -Zṯz 3 * ; * j3l --* -IL* -2-1* 3 i *^ A 09\.e *1 2 \ 2 * - -s * *, 1 *98t Le * *-Z *-* *1 *3 *1 *1 *3 * t * - 2 h,* *.3 6 *Z 1 *1 * *2.2 *: *2 - *1 - *3 *25 * 2 *1.*2 728 msec Type C 147 msec Type O

6 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. 827 maximum. Because of this, and because these positive potentials continued through the end of ventricular excitation (B,3) and the early ST segment (B,4), the origin of this maximum, as well as the prominent notch at the end of QRS, was ascribed to repolarization. The end of excitation (B,3) was characterized by a gradual disappearance of the upper sternal maximum, after which the anterior minimum quickly shifted to the back during the ST-segment (B,4). This pattern occurred in 20 subjects, 15 of whom were years old; the remaining five occurred in older subjects, as old as 55 years. The third type (fig. 2, type C) occurred in 20 subjects in all age groups. There was a single excitation maximum during late QRS, in this case at the upper sternum with a minimum nearby (C,1). The first evidence of repolarization occurred in the negative potentials on the left precordium with the extension of less negative potentials (lower absolute value than the surrounding negative area) from the upper back (C, 1). Prominent positive repolarization potentials rapidly developed in this precordial area and resulted in a distribution with two maxima and two minima (C,2) similar to that described by Taccardi.23 Excitation ended with the disappearance of the excitation maximum and minimum that were located over the sternum (C,3), while the repolarization maximum on the anterior precordium continued to increase during the first 50 msec of the ST segment, and the minimum persisted on the lower anterior torso (C,4). The type C and type B patterns were commonly encountered in subjects years old. Both frequently were associated with waveforms in the midprecordium that had prominent positive notches at the end of QRS; these positive notches were due to repolarization rather than excitation. However, in some subjects waveforms occurred on the upper left precordium and upper sternum with similar positive notches that were due to excitation, since the maximum associated with these notches disappeared rather than persisted during the ST segment. The origin of these small notches in the waveforms could be determined only by viewing the potential distributions with respect to terminal excitation vs early repolarization. The final pattern (fig. 2, type D) occurred in 16 subjects, again in all age groups. During late QRS there were two excitation maxima, one at the upper sternum and the other in the middle of the back with a prominent minimum on the anterior chest (D,1). The maximum in the back disappeared. The upper sternal excitation maximum persisted and repolarization positive potentials extended from that area inferiorly down to the left parasternal area to produce a discrete repolarization maximum as shown in (D,2). It was difficult to determine the precise time of onset of repolarization potentials since they extended directly from the positive area on the upper sternum that surrounded the excitation maximum. As the excitation maximum disappeared, the lower precordial repolarization maximum increased (D,3) with a shift of the minimum to the back during the early ST segment (D,4). Time Duration of Overlapping Excitation and Repolarizing Potentials The different patterns showed that the J point in the waveforms was not a precise index of either the end of excitation or the beginning of repolarization. Therefore, for the body surface maps, we applied the term J point to that instant when ventricular excitation potentials disappeared in the body surface maps. The variability in the number and location of the excitation maxima, which occurred either on the back or at the upper sternum, or both, with either one disappearing first, indicated there were normal variations in the late ventricular activation sequence. These differences in the excitation maxima during late QRS were consistent with the differences in the ventricular excitation sequences measured by Durrer et al.30 in perfused human hearts. In most subjects the best criterion of the end of excitation was the disappearance of either the excitation maximum at the upper sternum or of the one on the back. However, in several subjects, the best index was the disappearance of the anterior torso minimum when it was more prominent than either excitation maximum. A comparison of the time of the J point of the map, as measured from the onset of QRS, and of the time of the appearance of positive repolarization potentials for all normal subjects, is shown in figure 3. With increasing age there was an increase in the time of appearance of repolarization potentials. This occurred in association with the expected increase in the time of the J point. Conversely, the trend was for the duration of overlap of excitation and repolarization potentials to decrease with increasing age. Although the graph in figure 3 indicates the trend of change with age, similar times of the J point and similar times of the FIGURE 2. Appearance of ventricular repolarization potentials. The body surface potential distributions are presented in the same format as in figure 1. They are shown for two normal subjects. Each series represents one offour types of changes in the body surface potential distributions that occurred with the appearance of ventricular repolarization potentials during late QRS. In addition to the variable patterns of the excitation potentials during late QRS, the major repolarization differences illustrated in the first two rows of this figure andfigure I included thefollowing: In type A, repolarization potentials did not appear until the excitation potentials had almost disappeared. In the remaining three types, repolarization potentials developed while prominent excitation potentials were still present. Note that the major differences in these three types were that the positive potentials of repolarization shifted to the left precordium from the lower torso (type B), high left axilla (type C), and upper sternum (type D) as shown by the arrows in row 2.

7 828 CIRCULATION VOL 59, No 4, APRIL 1979 n=l I n=10 n=13 n10 n-22 msec AGE IN YEARS appearance of repolarization potentials occurred in all age groups. The longest duration of overlap of excitation and repolarization potentials occurred with the types of change during late QRS shown in sequences B and C of figures 1 and 2 in subjects younger than 30 years. In some subjects the overlap lasted for msec. Maximal Body Surface Potential Values During ST-T-U Wave four parameters were The values of the following measured as a survey of ventricular repolarization in each normal subject: 1) the J-point voltage maximum; FIGURE 3. Duration ofoverlapping excitation and repolarization potentials. The number of subjects in each age group is indicated at the top ofeach column. The time of the end of ventricular excitation (solid circles) and the time of appearance of repolarization positive potentials (open circles) for each subject were measured in reference to the onset ofqrs by viewing the sequential maps millisecond by millisecond. For each age group, the average value is indicated by the horizontal solid lines. The mean time duration of overlap for each age group is the time difference between the two horizontal lines of each column, the thick vertical bar at the bottom of each column representing the mean time duration of overlapping excitation and repolarization of that age group. The dashed lines are included to emphasize the decrease in the time duration of overlapping excitation and repolarization potentials with increasing age. be = body surface potentials. 2) the voltages of the extrema of the ST (0-50 msec) maps; 3) the maximal voltage that occurred during the T wave; and 4) the maximal voltage of the U wave. As shown in figure 4, the results indicated a similar trend with increasing age for all four parameters. The maximal value of each increased from childhood into the twenties and declined thereafter. The lowest values of each parameter occurred in the youngest and oldest subjects, but there was considerable overlap, with similar values of each parameter spanning all age groups. The decrease in the magnitude of the maximal voltage of the ST segment and T wave with increasing age after 30 years is in agreement with the results of

8 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. 829 A. (I) J Point Moximum 4(e B.,aV t I i I.- t I i ~ ~ ~ i AGE IN YEARS FIGURE 4. Maximal potential values during ST- T- U wave in normal subjects. Each value represents the designated potential measuredfrom the body surface maps ofeach subject. The potential values of the body surface maps were measured with respect to the baseline chosen during the UP interval (TP interval in small children) of the original waveforms. The number ofsubjects and the age range in years represented by each column is indicated at the bottom of each panel. In A (1) the J-point maximal body surface potential (be) is the highest potential of the body surface map at the instant when ventricular excitation potentials disappeared; i.e., the first instant when only repolarization potentials were present. In A (2) the ST(0-50 msec) max-min be values are the highest and lowest values of the ST (0-50 msec) map ofeach subject. The T-wave maximum 4'e in B (1) and the U-wave maximum (be in B (2) are the highest potential values that occurred on the body surface during inscription of these waveforms. In each column the horizontal line indicates the average value of the potential measurements ofthat age group. There were no U waves (B,2) in the 1-7-yearold age group.29 Note that for each phase of ventricular repolarization a similar age trend is indicated, although for each variable similar maximum values were encountered in some subjects at all ages.

9 830 CIRCULATION VOL 59, No 4, APRIL 1979 the precordial lead analysis of the ST segment and T wave by Hiss et al.8 However, neither their scalar ECG data nor our surface map results agree with the results of Parisi et al.,10 who found increasing values for the ST segment and T wave in healthy adult men older than 30 years. There were no U waves in the 11 children under 8 years. We have been unable to find standard ECG analyses for age-related trends of the U wave amplitude for comparison with these map results. However, these results (fig. 4B) are in agreement with those of Lepeschkin3' and Surawicz et al.32 who noted in their scalar lead analyses that the tallest U waves were associated with the tallest T waves. Lepeschkin31 noted that the U wave was usually 5-20% of the T wave. In these map results, the maximal voltage of the U wave was 4-28% of that of the T wave. Although the highest values of the U waves were associated with the largest T waves, the U/T wave maximal voltage ratios provided no direct index of the absolute magnitude of either, since the highest ratios occurred with the smallest T waves. For example, maximal voltages of the U waves were 15-28% of the T waves when the T-wave maximal voltage was less than 0.5 mv, and the U waves were 4-12% of the T waves with T waves that had a maximal voltage greater than 1 mv. It is well-known that the heart rate affects the amplitude of the T wave and the U wave.3' Within the range of the basal heart rates of the total group (53-80 beats/min), we could find no evidence that heart rate affected the differences in the maximal T-wave and U- wave voltages, or those of the ST segment, at different ages. Also, for the 50 subjects over 18 years of age we could find no evidence that variations in body size (body weight range kg) affected the age trend of the voltages of the four repolarization parameters (fig. 4). This was similar to the conclusion of Simonson and Keys33 in their detailed analysis of standard ECGs of normal adults at different ages. There were positive potentials during the ST segment in all 66 normal subjects (fig. 4A). Note that in several of the normal year-old subjects the early ST-segment maximum was greater than 200,V, a value used as an index of ST-segment change due to ischemia in summation ST-segment elevation precordial maps in older patients after acute myocardial infarction." 3 However, the maxima in the ST (0-50 msec) maps and at the J point were less than 200 AV in the 22 subjects over 40 years of age. The minima of the ST (0-50 msec) maps were of considerably lower absolute value than the maxima in all subjects greater than 7 years of age. Although the age trend of the minima was less obvious than that of the maxima, the most negative ST-segment potentials occurred in the children and the least prominent STsegment minima occurred in the oldest adults (fig. 4, A2). Location of Positive and Negative ST-Segment Deflections To determine the body surface distribution of positive and negative ST-segment deflections, we analyzed the following in the form of NT (0-50 msec) maps in each subject: 1) the location of the maximum and minimum, and 2) the location of each absolute value that exceeded +30 uv. Since previous analysis of the waveforms29 of these subjects indicated that absolute values in excess of +30,uV should be accurate in these maps, we apply the terms positive and negative ST-segment deflections to correspond to these limits in the ST (0-50 msec) maps. Figure 5 shows composite maps of the locations of the ST-segment extrema and the locations of the positive and negative ST segments for the total group. The composite map in A(l) represents the locations of the maxima and minima of the 11 children younger than 8 years, and the composite map in B(l) represents the locations of the extrema of the 55 subjects 8-60 years of age. Each of these composite maps was constructed by plotting the location of the maximum and the minimum of the ST (0-50 msec) map of each subject to produce a composite picture of the locations of the ST-segment maxima and minima in the younger (A) and older (B) groups. Each of the accompanying composite maps in A(2) and B(2) show the locations of positive and negative ST-segment deflections. Each of these composite maps was. constructed in the following way: As measured in the ST (0-50 msec) map, the polarity of the ST segment (+30 AV) was marked at each of the 150 locations in each subject. The composite was made by overlay markings of the locations of the positive and negative STsegment deflections subject-by-subject. Each composite map in A(2) and B(2) shows the locations where the ST (0-50 msec) map positive and negative potentials exceeded the absolute value of ±30 AV for the total group of individual subjects. In all subjects 8-60 years of age the ST-segment maxima were located within a circumscribed area on the anterior torso that covered the lower sternum and adjacent upper abdomen and precordium (fig. 5, Bl). The minima were located in multiple areas distant to the heart, including the right upper lateral chest, the back, and right anterior lower torso. These ST (0-50 msec) map results and the ST-segment potential distributions shown in figures 1 and 2 differ from those of Taccardi,23' 25 who found the ST-segment minimum in normal adults to be located most frequently on the right upper chest beneath the clavicle. The magnitudes of the minima were of low value (-15 to -90,V), and it was frequently difficult to assign them to a specific location, as noted by Taccardi.23 In only four subjects, all older than 50 years, did minima occur on the left side of the back in an area adjacent to the left axilla (fig. 5, Bl). The younger children (A,1) differed from the older subjects in that the ST-segment maxima were located on the upper left precordium and the minima occurred nearby in the right parasternal area. The locations of the positive and negative STsegment deflections showed a similar difference between the younger children and the older group. In the group of younger children the areas in which there was no overlap of positive and negative ST-segment

10 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. 831 A (I) ---Vl. I-7 Years (n-lj) l (2) B. (I) ; Max ST (0-50 msec) 60 to 370?OV Min ST(0-50msec)-20 to-120p,v 8-60 Yeors (n-55) (2) > 30 AV ST (O - 50 msec) B < -30 V ST (O - 50 msec) Max ST (0-50 msec) 60 to 370,V (n 55) = M in ST(0-50 msec)-15 to -90,uV (n -51) Min ST(0-50 msec)-15 to -40,uV (n-4 ) > 30jiV ST(0-50msec).-30,pV ST(0-50msec) FIGURE 5. Normal subjects: Composite maps of the locations of the ST-segment extrema and the locations of the positive and negative STsegments in normal subjects. The locations of the ST(0-50 msec) map maxima and minima of the 11 children younger than 8 years are shown in A (1) and the locations of the ST (0-50 msec) map maxima and minima of the 55 normal subjects between 8-60 years old are shown in the composite map in B (1). The positions indicated by the dots on the composite maps are the same as those of the isopotential maps. Each composite map represents the designated group of individual subjects. Note that the locations of the ST-segment maxima and minima in the 11 children (A,1) were located with a transverse (left-to-right) orientation, while the orientation of the maxima and minima in the older subjects was more in an anterior-toposterior direction. In B (I) the lightly slashed area is shown to designate the area where the minima (-15 to -40,u V) occurred in four subjects, all of whom were in the year group. The accompanying composite maps in A (2) and B (2) show the locations of the ST (0-50 msec) map potentials that exceeded ±30,u Vfor all of the subjects within each group. The stippled areas represent the locations where ST(0-50 msec) map positive potentials were higher than 30,u V, and the slashed areas represent the location ofthe negative potentials lower than -30, V. The plus and minus signs indicate the maxima and minima in the composite maps on the left, and they indicate the areas where only positive and negative ST-segment deflections occurred in the composite maps on the right. deflections were oriented in a left-to-right direction (A,2), compared with a more anteroposterior orientation of these areas in the older subjects (B,2). Note in the younger children the wide vertical area on the left lateral chest and adjacent back where only positive ST-segment deflections occurred, and on the lateral right side of the torso where only negative ST segments occurred (A,2). In the older group (B,2), the

11 832 CIRCULATION VOL 59, No 4, APRIL 1979 region within which only positive ST segments occurred was located anteriorly in an area extending from the sternum to the left axilla (to a location near lead V6), and the area where only negative ST-segment deflections occurred was on the upper back and right shoulder. Note also in figure 5, B2 that for the total group of older subjects, both positive and negative ST segments occurred on the lower torso and in the left axilla at the location of lead V,. Spatial Analysis of ST-Segment Potential Distributions To find if there were different ST-segment potential distributions in the 55 normal subjects 8-60 years old, we analyzed the ST (0-50 msec) map of each subject to determine the overall pattern produced by the lowlevel potentials that exceeded an absolute value of ± 30,uV. In these maps the location of the minima and the associated low-level negative potentials varied markedly on the back and right side of the torso, and they seemed to form many types of ST-segment patterns that were difficult to classify. However, there were easily discernible differences in the positive potentials for this total group. The pattern of the positive potentials could be described by one of four types of patterns, as illustrated in the ST (0-50 msec) maps of four normal subjects in figure 6. While the maximum in all of these subjects was located within a circumscribed area on the lower left precordium and adjacent upper abdomen, the patterns varied primarily in relation to the vertical distance over which the low level positive potentials extended on the left anterior torso. In the first pattern shown in figure 6A the lower torso was relatively negative, with positive potentials extending superiorly as a broad area to the limits of the upper anterior chest. In the second pattern (fig. 6B) there was an opposite vertical pattern with positive potentials extending to the inferior limits of the map, and more negative potentials occurred at the top. The third pattern (C) was characterized by a vertical column of positive potentials that covered the left anterior torso from the lower to the upper limits of the map. The final pattern (D) consisted of a circumscribed region of positive potentials on the anterior torso with more negative potentials at the top and the bottom. These variations in the ST (0-50 msec) maps were intriguing in relation to the similar way there were four types of change in the potential distributions during late QRS (figs. 1 and 2). The late QRS patterns differed from subject to subject in the way positive repolarization potentials appeared in areas that varied from the lower torso to the upper anterior chest, and then they shifted to the mid precordium. Some general relationships could be identified between the patterns of the ST (0-50 msec) maps of figure 6 and the types of change produced by the appearance of repolarization potentials during late QRS (figs. 1 and 2). All of the ST (0-50 msec) map patterns were found in each age group from 8-60 years of age. The patterns shown in B and C of figure 6 were most often encountered in the 8-30-year-old subjects (15 of 23 subjects). These ST-segment patterns were associated most frequently with the late QRS changes shown for types B and C of figures and 2. Thereby, these two types of appearance of repolarization potentials (figs. and 2) and these ST (0-50 msec) map patterns (figs. 6B and 6C) were most commonly found in young adults and adolescents. The ST (0-50 msec) map patterns shown in B and C of figure 6 also were associated with the largest magnitudes of the ST-segment maxima, since the young adults and adolescents represented the age groups where these maximal values occurred (fig. 4A). The fourth ST (0-50 msec) map pattern (fig. 6D) occurred most frequently in the year age group (10 of 12 subjects), and this ST-segment pattern, therefore, was frequently associated with the type of change produced by the appearance of repolarization potentials during late QRS shown for the potential distributions of type A in figure 1. Finally, the ST (0-50 msec) map pattern shown in figure 6A was associated with all but one of the types of change during late QRS; it did not occur with type A of figure 1. Discussion There were two major features of the body surface ST-segment potential distributions in normal subjects. First, all subjects had positive ST segments on the anterior torso, and the magnitude of the maximal STsegment voltage varied from subject to subject. The magnitude of the maximal ST-segment voltage increased until years of age and decreased thereafter. The same age-related trend occurred for the other phases of ventricular repolarization, e.g., the T wave and the U wave. Second, although the STsegment maxima were located in a circumscribed area near the heart in all adults, the spatial distribution of the low level positive potentials varied in distant areas and resulted in at least four identifiable patterns in the normal ST-segment body surface maps. These two features of the ST segment in normal subjects have one general interpretation. During early ventricular repolarization (plateau phase of the action potential), there was a difference in the spatial arrangement of the ventricular intracellular potentials on a subject-by-subject basis, and there was a trend for the ventricular intracellular gradients (or the currents they produce) to change with age. The different normal ST-segment patterns (fig. 6) indicate that the arrangement of the ST-segment "current sources" is detectably different in normal subjects, i.e., there is no one specific spatial arrangement of intracellular potentials during the plateau phase of repolarization that can be classified as normal. Therefore, the ventricular locations and magnitudes of the intracellular gradients vary from one normal subject to another to produce the different ST-segment patterns illustrated in figure 6 and the considerable variations in the magnitude of the maximal ST-segment potentials shown in figure 4A. A comment is appropriate about the ST-segment patterns of figure 6, whether the different patterns of positive potentials represent different categories that are real or whether they represent minor and non-

12 A. 25 Years NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al B. 22 Years ) 180,AV G-30pV ( 250pV 9-0ioo Av v-i- =me C. 39 Years (3 220AV, ) -50,uV D.52 Years O-20V ) 8OiV,-F, : B -I -I -2s _1.D+1-t.0^ L *2 *2 * * _- -1 -I E11 30OAV *1 t 2. 1? 212e *.2*1 *s2 *.2 *2 *1~ +1 *2 *2 *1.0.0 *L +2 * *+1 *1 *1 * *1 K0.0 *1 40 #1 41.I 42 *2 *a *1 *2 *2 SE <-30 AiV FIGURE 6. ST-segment potential distribution patterns in normal subjects. The ST (0-50 msec) maps are those of individual normal subjects, each of whom represents one ofthefour ST-segment patterns encountered in the 8-60-year-old age group. The stippled area represents the region ofpositive potentials that exceeded 30 p V in the S-T (0-50 msec) maps and the slashed area represents the region ofnegative potentials that were less than -30,u V. The locations of the maxima and minima are indicated by the plus and minus signs, respectively. The brackets superimposed on the waveform below each map indicate the first 50- msec interval of the ST segment for which the average potential was determined at each of the 150 positions to make each M7T (0-50 msec) map. Note that in all four subjects the maximum occurred within a circumscribed area near the heart. The major differences in the patterns occurred in relation to the vertical distance over which the low-level positive potentials extended on the left anterior torso. Ir reproducible differences. By definition of the method for demarcation of those potentials greater than 30,uV, the assignment of the ST (0-50 msec) maps to one of the four categories was reproducible from observer to observer. The opposite vertical distribution of positive potentials in patterns A and B strongly indicates that they reflect differences in the cardiac sources. However, although patterns C and D were easily defined as described, they present a more difficult problem in general interpretation, whether their origin was from different cardiac sources or from different effects of the volume conductor (e.g., proximity effects of the heart). The patterns in figure 6 are presented to illustrate the detectable variability of normal ST-segment potential distributions rather than to provide a fixed classification of patterns.

13 834 CI RCULATION VOL 59, No 4, APRIL 1979 Normally, in the adult hearts of experimental animals the ST-T-wave extracellular potentials in the walls of the ventricles are influenced primarily by a unidirectional gradient across the wall with positive potentials at the epicardium and negative potentials at the endocardium."' 12 This unidirectional extracellular gradient results from lower intracellular potentials at the epicardium and higher intracellular potentials at the endocardium of both ventricles. The ST-segment maps in the normal subjects 8-60 years of age were consistent with these experimental data, i.e., the maximum occurred anteriorly within a circumscribed area near the heart and the minimum and associated negative potentials occurred in the back and in other areas distant to the heart. However, the variations in the low-level, positive ST-segment potentials on the left anterior torso (fig. 6) indicate that in normal human hearts there is another effect superimposed on the predominant transmural ventricular gradient. The ST-segment maps indicate that on a subject-to-subject basis there should be different areas in the ventricle where the intracellular potentials are lower than in other areas during the plateau stage of the action potential, irrespective of total action potential duration. If mild ventricular ischemia should occur and produce currents of small-to-moderate magnitude, rather than the massive ischemic currents of acute infarction,'"3 the specific pattern of the STsegment potential distribution will depend not only on the location of the ischemic muscle, but also on the background "normal" spatial arrangement of the intracellular action potentials. The presence of variability in the low-level, positive ST-segment potentials on the left anterior torso in normal adults suggests that the evaluation of ventricular ischemia might be enhanced by interpreting a questionably abnormal surface map, such as one obtained during exercise, with resepct to the patient's own control surface map. There is no way to prove or disprove the above interpretations based on the body surface maps alone. However, they present a way of interpreting the STsegment body surface maps on a valid theoretical basis,'8' 21 and they are consistent with the qualitative and quantitative explanation of similar ST-segment patterns found in the epicardial and body surface potential distributions measured in intact animals."' 18 They are consistent also with the simulation studies of the ST segment by Miller and Geselowitz.35 The results and interpretations emphasize the importance of incorporation into the study of the origin of body surface ST segments based on intracardiac events, such as those associated with ventricular ischemia, measurements of the total cardiac electrical field on the body surface. Body surface maps provide a way of accounting for the effects of the currents that arise from all areas of the heart, rather than characterizing primarily the proximity effects as embodied in precordial maps. The major differences in the ST-segment patterns occurred in the low-level potentials in areas distant to the heart, rather than in the proximal precordial region. It is this feature that distinguishes the body surface map from a map limited to the precordial region, i.e., the potential distribution over the entire volume conductor is required to account for the currents generated from all areas of the heart. Inverse Calculation of ST-Segment Epicardial Maps from the Body Surface Maps If epicardial potential distributions computed from the body surface potential distributions could be viewed, rather than body surface maps, the interpretation could be tested in more detail. As noted earlier, based on experimentally measured epicardial and body surface potential distributions in intact animals, the recent inverse calculation studies of Barr and Spach2' have indicated that it should be possible to apply this approach to humans. To test the interpretation of the patterns found in the ST (0-50 msec) maps (fig. 6) of the normal subjects (i.e., there should be different areas in the ventricle where the intracellular potentials are lower than in others during the plateau phase of the action potential), the inverse calculation method was used to derive the associated epicardial potential distributions of the normal ST-segment patterns of figure 6. These preliminary computed epicardial maps in humans were based on human body surface and heart geometric measurements. The methods have been described in detail for dogs.21 They represent the potentials computed at 75 locations on the heart from the 150 body surface points of the surface maps. Figure 7 shows the epicardial potential distributions computed from two of the normal ST-segment body surface patterns (figs. 6A and B). In both of these normal subjects the computed epicardial pattern was predominantly one of positive potentials over the ventricle and negative potentials over the atrium with the minimum at the atrioventricular ring at the lateral base of the left ventricle. The magnitudes of the epicardial voltages varied from -0.7 mv to 1.0 mv. In figure 7A the body surface ST-segment pattern with the superiorly oriented positive potentials produced a single maximum high on the anterior epicardial surface, a location similar to that found in intact chimpanzees." The normal ST-segment body surface potential distribution shown in figure 7B had a single maximum and minimum, with low-level positive potentials on the inferior torso. The "inverse" epicardial map (B) resulted in two epicardial maxima, again similar to animal experimental results,'2 with an anterior and a diaphragmatic epicardial ST-segment maximum. The low-level positive potentials on the lower torso were related to the epicardial maximum on the diaphragmatic surface of the heart. Both the patterns and the magnitudes of the computed epicardial ST-segment potentials were consistent with previous intact animal results."' The animal experiments have shown that on the epicardium and the body surface, positive potentials during ST-T waves can be interpreted to reflect areas of earlier repolarization with respect to areas of later repolarization where less positive or negative potentials occur.

14 NORMAL ST-T-U WAVE BODY SURFACE MAPS/Spach et al. A. NORMAL: 25 YEARS Q-30,V I180 AV 0-3,,iV IS? I.N, 0)0.6 mv (0-0,4 mv 835 p 1 (0 o N~~~~~~~~11 ~I B. NORMAL:22 YEARS 250p.V eq-ioo,av (30.Gmv Q my El > 30,uV E.-30,AV.-60iAO IV FIGURE 7. Inverse ST-segment epicardial potential distributions calculated from ST (0-50 msec) body surface maps. The body surface maps on the left are reproduced from those of the normal subjects shown for patterns A and B offigure 6. Each associated epicardial map was drawn as an overlay tracingfrom the photographs of the computed epicardial maps, as done for previous experimentally measured epicardialpotential distributions in chimpanzees", and dogs."2 21 The epicardial outlines of the heart show the anterior surface ofthe heart on the left and the diaphragmatic surface on the right. The values of the maxima and minima are indicated above each map and the values of the isopotential lines within the designated positive and negative regions are shown below. The units are in,uv for the body surface and in mv for the epicardium. RV = right ventricle; LV = left ventricle. These computed "inverse" ST-segment epicardial maps are consistent with the general interpretation noted previously for normal adults: During the ST segment there is a predominant transmural gradient in the walls of the ventricles due to lower intracellular potentials at the epicardium (positive extracellular potentials) and higher intracellular potentials at the endocardium (negative extracellular potentials reflected on the atrium).", 12 However, the differences in the location and the number of the epicardial maxima reflect differences in ventricular areas of "early repolarization" (lower intracellular potentials) during the ST segment, as suggested from the differences in the low-level positive potentials in the original body surface maps. The results of this study indicate that the use of total body surface maps in the form of potential distributions offers several advantages in the evaluation of selective heart events during the ST segment: 1) STsegment differences in normal subjects were apparent

15 836 CIRCULATION VOL 59, No 4, APRIL 1979 in the low-level potentials that occurred in areas distant to the heart; 2) the body surface patterns and their differences could be interpreted on an intracellular basis; and 3) the interpretation could be compared with computed inverse epicardial potential distributions. References 1. Maroko PR, Libby P, Covell JW, Sobel BE, Ross J Jr, Braunwald E: Precordial S-T segment elevation mapping: an atraumatic method for assessing alterations in the extent of myocardial ischemic injury. The effects of pharmacologic and hemodynamic interventions. Am J Cardiol 29: 223, Muller JE, Maroko PR, Braunwald E: Precordial electrocardiographic mapping. A technique to assess the efficacy of interventions designed to limit infarct size. Circulation 57: 1, Ried DS, Pelides LJ, Shillingford JP: Surface mapping of RS-T segment in acute myocardial infarction. Br Heart J 33: 370, Mirvis DM, Keller FW, Cox JW Jr, Zettergren DG, Dowdie RF, Ideker RE: Left precordial isopotential mapping during supine exercise. Circulation 56: 245, Fozzard HA, DasGupta DS: ST-segment potentials and mapping. Theory and experiments. Circulation 54: 533, Sokolow N, Friedlander RD: The normal unipolar precordial and limb lead electrocardiogram. Am Heart J 38: 665, Goldman MJ: RS-T segment elevation in mid- and left precor-.dial leads as a normal variant. Am Heart J 46: 817, Miss RG, Lamb LE, Allen MF: Electrocardiographic findings in 67,375 asymptomatic subjects. X. Normal values. Am J Cardiol 6: 200, iwasserburger RH, Alt WJ, Lloyd CJ: The normal RS-T seg-,ment elevation variant. Am J Cardiol 8: 184, Parisi AF, Beckmann CH, Lancaster MC: The spectrum of ST segment elevation in the electrocardiograms of healthy adult men. J Electrocardiol 4: 137, Spach MS, Barr RC, Lanning CF, Tucek PC: Origin of body surface QRS and T wave potentials from epicardial potential distributions in the intact chimpanzee. Circulation 55: 268, Spach MS, Barr RC: Ventricular intramural and epicardial potential distributions during ventricular activation and repolarization in the intact dog. Circ Res 37: 243, Ramsey M III, Barr RC, Spach MS: Comparison of measured torso potentials with those simulated from epicardial potentials for ventricular depolarization and repolarization in the intact dog. Circ Res 41: 660, Holland RP, Brooks H: Precordial and epicardial surface potentials during myocardial ischemia in the pig. Circ Res 37: 471, Holland RP, Brooks H: Spatial and nonspatial influences on the TQ-ST segment deflection of ischemia. J Clin Invest 60: 197, Holland RP, Arnsdorf MF: Solid angle theory and the electrocardiogram: physiologic and quantitative interpretations. Progr Cardiovasc Dis 19: 431, Plonsey R: Bioelectric Phenomena. New York, McGraw-Hill, 1969, pp Spach MS, Barr RC: Origin of epicardial ST-T wave potentials in the intact dog. Circ Res 39: 475, Spach MS, Barr RC, Lanning CF: Experimental basis for QRS and T wave potentials in the WPW syndrome. The relation of epicardial to body surface potential distributions in the intact chimpanzee. Circ Res 42: 103, Barr RC, Ramsey M III, Spach MS: Relating epicardial to body surface potential distributions by means of transfer coefficients based on geometry measurements. IEEE Trans Biomed Eng 24: 1, Barr RC, Spach MS: Inverse calculation of QRS-T epicardial potentials from body surface potential distributions for normal and ectopic beats in the intact dog. Circ Res 42: 661, Spach MS, Barr RC: Analysis of ventricular activation and repolarization from intramural and epicardial potential distributions for ectopic beats in the intact dog. Circ Res 37: 830, Taccardi B: Body surface distribution of equipotential lines during atrial depolarization and ventricular repolarization. Circ Res 19: 865, Taccardi B, DeAmbroggi L, Viganotti C: Characteristic features of surface potential maps during QRS and S-T intervals. Adv Cardiol 10: 248, Taccardi B, DeAmbroggi L, Viganotti C: Body-surface mapping of heart potentials. In The Theoretical Basis of Electrocardiology, edited by Nelson CV, Geselowtiz DB. Oxford, Clarendon Press, 1976, pp Abildskov JA, Burgess MJ, Lux RL, Wyatt R, Vincent GM: The expression of normal ventricular repolarization in the body surface distributiop of T potentials. Circulation 54: 901, Block P, Lenears A, Tiberghein J, Coussert E, van Thiel E, Lebedelle M, Raadschelders I, Bourgain R, Kornreich F: Surface maps and myocardial scanning at rest and during exercise: comparison with coronary angiography. Acta Cardiol 31: 467, Hoffman BF, Cranefield PF: Electrophysiology of the Heart. New York, McGraw-Hill, 1960, pp Spach MS, Barr RC, Warren RB, Benson DW, Walston A II, Edwards SB: Isopotential body surface mapping in subjects of all ages: Emphasis on low level potentials with analysis of the method. Circulation 59: 805, Durrer D, van Dam R Th, Freud GE, Janse MJ, Meijler FL, Arzbaecher RC: Total excitation of the isolated human heart. Circulation 41: 899, Lepeschkin E: The U wave of the electrocardiogram. Arch Intern Med 96: 600, Surawicz B, Kemp RL, Bellet S: Polarity and amplitude of the U wave of the electrocardiogram in relation to that of the T wave. Circulation 15: 90, Simonson E, Keys A: The effect of age and body weight on the electrocardiogram of healthy men. Circulation 6: 749, Pelides LJ, Reid DS, Thomas M, Shillingford JP: Inhibition by 3-blockade of the ST segment elevation after acute myocardial infarction in man. Cardiovasc Res 6: 295, Miller WT III, Geselowitz DB: Simulation studies of the electrocardiogram. 11. Ischemia and infarction. Circ Res 43: 315, 1978