Late potentials in normal subjects and in patients with ventricular tachycardia unrelated to myocardial infarction

Late Potentials in Normal Subjects and in Patients with Ventricular Tachycardia Unrelated to Myocardial Infarction HUMBERTO COTO, MD, CLAUDIC MALDONADO,

MA, PRASAD PALAKURTHY,

MD,

and NANCY C. FLOWERS, MD

Fifty normal male and female athletes, or athletically active subjects, were evaluated, and a search for low-amplitude late potentials in the terminal part of ventricular activation was performed. Recordings from 3 normal men met the definition of abnormal late potentials, and were indistinguishable by present analytic techniques from those encountered in patients who have ventricular tachycardia (VT) after myocardial infarction (Ml). Of 24 patients studied, 11 had VT, but only 2 had had an Ml, which occurred in the remote past. Another patient had 1 narrowed coronary artery on arteriography. Group differences could be demonstrated using amplitudes

and durations of late potentials, but late potentials generally did not prove the impressive marker of the patient with VT, which other workers, as well as ourselves, have encountered in patients after Ml. Late potentials were an impressive marker in a subset of the VT group in whom cardiomegaly developed. Thus, the absence of late potentials is an effective marker in the normal subject, but the presence of late potentials is not an effective marker in identifying the patient with non-Ml-related, nonsustained VT before development of cardiomegaly. (Am J Cardiol 1985;55:384-390)

In 1973, Boineau and Coxl demonstrated slow ventricular activation, in association with a recent myocardial infarction (MI), as a source of reentrant premature ventricular depolarizations. Berbari et al2 reported arrhythmogenic ventricular activity occurring during the ST segment in 1978. In the early 198Os, Simson, Breithardt et al4 and, more recently, other investigators, evaluated late potentials as a possible means of identifying certain patients prone to ventricular tachycardia (VT). Late potentials may be thought of as deflections occurring after the dominant QRSwave form that are relatively high in frequency and too low in amplitude to be distinguishable from noise without the application of additional noise reduction and signal enhancing techniques such as amplification, filtering and signal averaging. Based on either the duration of relatively low amplitude deflections in the terminal QRS or the ST segment, or the determination of the root-mean-square (RMS) of the voltage over a particular time period after the basic QRS wave form, certain group and possibly individual differences may

be shown between those patients who sustained MI and later had VT and those who did not. One difficulty in extending these initial nosologic observations further is the scarcity of information on the incidence of lowamplitude late potentials in normal subjects.5Our study characterizes the occurrence of late potentials in a normal and athletically active group who had not sought entry into a health care system, and evaluates the predictive accuracy of late potentials in distinguishing another population subset with VT from causes other than MI. Methods Fifty normal subjects, all of whom either were athletes (n = 38) or exercised regularly (n = 12), were studied. Before inclusion into the study, each subject had a cardiovascular history taken and underwent physical evaluation, including a detailed smoking and family history. None were included who had a family history of early heart diseaseor who smoked in excess of 10 cigarettes daily. All subjects had a normal routine electrocardiogram (ECG). The agesof the 26 men and 24 women ranged from 18 to 36 years (mean 25.0 f 5.8). An additional 24 patients, recently referred for problems relating to cardiac rhythm or conduction, were considered for the study. One patient had right bundle branch block with a standard QRS duration of 130 ms. Another patient had an intraventricular conduction defect with a QRS of 140 ms. All other patients had standard QRS durations less than 120 ms. Eleven with nonsustained VT will be considered in this report. Each was symptomatic, had repetitive episodes, and therapy was required for control. All of the 11 had episodes that

From the Division of Cardiology, University of Louisville, Louisville, Kentucky, and the Division of Cardiology, Medical College of Georgia, Augusta, Georgia. This study was supported by Grant HL 33692 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland. Manuscript received June 14, 1984; revised manuscript received October 15, 1984, accepted October 16, 1984. Address for reprints: Nancy C. Flowers, MD, Professor of Medicine, Division of Cardiology, Medical College of Georgia, Augusta, Georgia 30912. 384

February 1, 1985 THE AMERICAN JOURNAL OF CARDIOLOGY Volume 55

stopped within 30 seconds, and all had demonstrated at least a 5-beat salvo. One patient had an occasional episode (by definition a sustained VT), lasting longer than 30 seconds.No patient had acute MI and only 2 had remote MI. Another patient had an occlusion larger than 75%, visualized in 1 coronary artery and 1 had insignificant plaque formation. The remaining causeswere specifically not coronary artery disease and included mitral valve prolapse, idiopathic dilated cardiomyopathy, overdosage of Mellaril@ in 1 patient, essential hypertension and aortic valve disease. The recording and signal-averaging process was performed with a system from Arrhythmia Research Technology which uses essentially the instrumentation described by, and the software developed by, Simson et a1.3+sIn many instances, normal subjects and patients vere recorded on more than 1 occasion to verify stability. In brief, 3 relatively orthogonal sampling points were amplified at a gain of 1,000,with a band width of 0.05 to 300 Hz, with a noise level of 0.7 /.LVreferred to input. The signal from each lead was further amplified 5 times and passed through a 4-pole, 250-Hz low-pass filter. Analog-to-digital conversion was performed to 1Bbit accuracy at a sampling rate of 1,000/s on a Hewlett-Packard 9826 computer, and the information was stored on floppy disks. Each lead was recorded for 2l/z minutes. With the Z lead as temporal reference lead, an 8-point template began at a ref-

erencetime and extended for 128ms to include distal QRS and early ST segment. The initial 8-beat template was accepted if the mean standard deviation after direct current shift was under 20 pV. Subsequent beats were tested against the template and accepted when within twice template standard deviation. During averaging, a 512-ms segment of the heart beat was identified, beginning 100 ms before onset of the QRS, and the template updated every fourth beat. After averaging, the records were filtered to eliminate low frequencies using a high-pass filter (HPF) at both 25 and 50 Hz in the present protocol. A bidirectional digital filter modified by Simson et al6 that achieves the elimination of ringing was used. The filtered signals for the X, Y and Z leads were integrated into a vectormagnitude complex by the usual formula, M = dX2 + Ys + 2s. Onset and offset of QRS were determined by computer algorithm, by respective forward and backward searches noise level to determine the midpoint of the 5-ms segment in which the average signal exceeds the noise level mean f3 standard deviations. Measurements included the total duration of the integrated QRS complex, the total duration of the terminal QRS potentials that are less than 40 /.LVRMS, the total duration of potentials in terminal QRS less than 25 PV RMS, the RMS of the total voltage in the last 40 and the last 50 ms of the QRS.

B

R

200

P

V6

2 I

FIGURE 1. A, standard 12-lead electrocardiogram from S.L., a 21year-old male athlete, and B, recordings from 3 relatively orthogonaloriented bipolar surface leads. B, horizontal lead X, vertical lead Y, and anteroposterior lead Z are shown with thr ordinate cross-hatched at 100~yV intervals. Lower panel, dots represent an estimate of noise using the same ordinate with 1 PV represented per interval on the scale. Divisions on the abscissa are at 20-ms intervals. Top panel, recordings of the vectorma nitude are seen calculated by the standard formula Jc--r--2 X + Y + Z Deflections on the ordinate are at 20-PU intervals and along the abscissa at 20-ms intervals. Upper left panel, a vectormagnitude curve using the 25 Hz high-pass filter while in the right panel the high-pass filter was specified at 50 Hz. Measurements from all subjects included total duration of vectormagnitude recording marked by dense vertical bars below the abscissa. The root-mean-square (RMS) voltage was calculated for 50 ms (98.25) and 40 ms (65.12) before the recognized end of QRS (dense vertical bar). The last 50-ms time interval is not diagrammed. The 40- and 25-PV levels of voltage are indicated, the duration of amplitudes less than 25 fiV (16 ms) and less than 40 yV (27 ms) was measured. Fc = frequency of the high-pass filter; NML = normal; T- = true-negative; V(40) and V(50) = 40 and 50 ms of ventricular activation, respectively.

385

~a 25 VECTOR.S.L. NML T” (48) 65. 12 ” <5m !38.25

200

LATE POTENTIALS

386

IN NORMALS

AND IN PATIENTS

WITH VENTRICULAR

Results A normal ECG, X, Y and Z leads and a vectormagnitude record are shown in Figures IA and B from a 21-year-old athlete. The 12-lead standard ECG is normal. Note the relatively sharp decay after the intrinsicoid deflection and the relative lack of QRS notching in the X, Y, and Z leads; also note the very rapid decline toward the terminal end of the QRS in the vectormagnitude record. The last 40 ms of the vectormagnitude recording reveals RMS voltage values of 85 and 42 /LV at both HPF frequencies, values well within the normal range. Voltage during the last 40 ms of the QRS recorded with the 25 Hz HPF was considered normal if it

288 -

FC 25 JG. NML ” (40) 22.33 V1

Fo 50

F+

16.01

v (50)

20. 49

:I

TACHYCARDIA

was at least 25 PV RMS. Using the 50 Hz HPF, voltage of at least 18 PV RMS was considered normal. Figure 2 is an example of a false-positive result from a normal 29-year-old soccerplayer whose RMS voltage during the last 40 ms was 22.33 and 16.01 pV, respectively. A 22year-old male athlete had the lowest voltage in the group during the 40 ms at 25 Hz-12.69 PV RMS. As was the case with all normal subjects, the standard direct writer ECGs were normal in all 3 false-positive patients with QRS duration not in excessof 100 ms. The amplified and filtered vectormagnitude records exceeded 100 ms in 4 subjects in the normal group, 2 of whom had late potentials. Specifically, none revealed intraventricular conduction defects and none had an rSr’ pattern in lead VI. Using the 25 Hz HPF, the correlation in the group of 50 normal subjects between total QRS duration and the duration of terminal potentials less than 40 /.LVwas poor (r = 0.362), as was the correlation with terminal potentials less than 25 UV (r = 0.453). The correlation did not improve with 50 Hz HPF. We also examined the correlation between total QRS duration in the normal subjects and the RMS voltage during the last 40 and 50 ms at 50 Hz. A negative correlation resulted of 0.569 and 0.480, respectively. However, when the correlation was tested between QRS duration and terminal voltage using the 25 Hz HPF as indicated in Figure 3, the negative correlation was somewhat higher during both the last 40 ms (r = -0.678) and during the last 50 ms (r = -0.684). For purposes of classification we termed as truenegative any normal subject who had 25 r.lV or greater in the last 40 ms of the QRS using the 25 Hz HPF. Forty-seven of 50 subjects met this criterion. Three normal subjects failed to meet this criterion and demonstrated late potentials. No other distinguishing fea-

Normal Subjects 140 -

lrn-

0

*

HPF

25

Hz

0

I -1000

‘:

: E

I

I :’ 6) w

E -

d w

I

A

FIGURE 2. Recordings from J.G., a 29-year-old male varsity soccer player, whose voltage during the last 40 ms of the QRS was 22.33 yV root-mean-square with 25 Hz high-pass filter and 16.01 PV rootmean-square with 50 Hz high-pass filter. Contrast the sharp decline in the vectormagnitude record in Figure 1 with the more prolonged lower amplitude higher frequencies in this figure. Further, note the slight increase in mid-QRS notching and slurring visible in the X, Y and 2 leads in this illustration, which is less impressive in Figure 1. While the standard 1P-lead electrocardiogram did not reveal a QRS greater than 100 ms, the pattern was not that of an intraventricular conduction defect, nor did V, show an rSr’ pattern, although a relatively prominent S wave in X is noted. F-t = false-positive; other abbreviations as in Figure

t.

4

FIGURE 3. With the high-pass filter (HPF) specification at 25 Hz, negative correlation within the normal subject group between QRS duration and the root-mean-square (RMS) of the voltage during the last 40 (solid regression line r = -0.678) and the last 50 ms (dashed regression liner = -0.684).

February 1, 1985 THE AMERICAN JOURNAL OF CARDIOLOGY Volume 55

TABLE I

387

Late Potentials in Normal Subjects and in Patients with Ventricular Tachycardia Unrelated to Infarction* 25 Hz HPF Normal Subjects

ms QRS PV last 40 ms /.rV last 50 ms ms term. pot. <40 pv ms term. pot. <25 /.rV

96.86 76.15 126.93 20.80

f f f f

VT Patients 1.49 7.31 9.61 1.59

14.86 f 1.04

118.36 46.26 72.22 33.72

f 10.16 f 9.30 f 13.61 f 4.87

19.27 f 2.90

P NS <0.025
* Using the 25 Hz high-pass filter, this table shows results of group comparisons using the 5 measured parameters characterizing late potentials. All values are expressed as mean f standard error of the mean. Significant differences were established between the normal subjects and the subset with ventricular tachvcardia when voltaae durina the last 40 and 50 ms was considered and when the duration of terminal potentials shorter than 40 PV ias considered. HPF = high-pass filter; NS = not significant; term. pot = terminal potential; VT = ventricular tachycardia.

tures were identified, either clinically or analytically in these subjects. Thus, 94% true-negative and only 6% false-positive responses were seen. An example of a true-positive response from the VT group is from a patient with hypertensive etiologic factors who demonstrated low voltages of 13.36PV RMS and 9.11 p.cVRMS at 40 and 50 Hz, respectively (Fig. 4). A false-negative record from a patient with mitral valve prolapse is illustrated in Figures 5A and B. This patient had occasional sustained VT as well as recurrent nonsustained VT. The coronary arteriogram was normal. Note the sharp decay of the vectormagnitude record with a normal terminal voltage of 85.9 PV and 34.48 PV RMS with 25 and 50 Hz HPF, respectively. Thus, in the VT group, 5 of 11 patients (or only about 45%) met the true-positive criterion: voltage during the last 40 ms of QRS was under 25 yV RMS (although a sixth patient demonstrated only 25.71 /LV RMS). Both patients with standard QRS durations in excessof 120 ms also had low-amplitude late potentials. Table I lists the comparison of group mean f standard error of the mean of the 5 parameters measured in the normal group and in the VT group. Unlike other comparative experiences, the total duration of QRS in the normal group and in the VT group did not show statistically significant differences; however, the trend toward a difference is obvious. Although the means are quite different, the reason for lack of significance appears to be a wide standard deviation among the patients with VT. A significant group difference was demonstrated in the voltage during the last 40 and 50 ms and in the duration of terminal potentials using the 25 Hz HPF. Using the 50 Hz HPF, significant differences were shown for the voltage during the last 40 and 50 ms, but not in the other 3 parameters measured (Table II). Although it was meaningful to make group comparisons, for a diagnostic test to be clinically useful, it should accurately specify a person as belonging to 1 or another diagnostic category. In this light, the overall predictive accuracy of late potentials was 85% using the

~0 2s GE VT HTN T+

FIGURE 4. Recordings from G.F., a woman with a hypertension (HTN) as a cause of heart disease, who had repeated runs of ventricular tachycardia (VT) and who is representative of our true-positive patients with low-amplitude late potentials of 13.36 and 9.11 FV root-meansquare with 25 and 50 Hz high-pass filter, respectively. Note the trailing out of the low-amplitude tail during the terminal portion of ventricular activation. T-F = true-positive; other abbreviations as in Figure 1.

388

LATE POTENTIALS

IN NORMALS

AND IN PATIENTS

WITH VENTRICULAR

single 25 PV criterion as mentioned (i.e., voltage at least 25 WVfor normal subjects and less than 25 PV for patients with VT using the 25 Hz HPF). However, this criterion reflects the good performance in the larger normal group, rather than an equally high positive predictive accuracy. Using all 5 parameters tested, we sought single values that would better separate the normal subject from the patient with VT. Table III demonstrates that when the 50 Hz HPF was used, 75% of the patients with VT had a potential less than 25 PV RMS for a duration of at least 26 ms, whereas 84% of the normal subjects had a terminal potential less than 25 PV RMS for durations uniformly less than 26 ms. Computer searches at 1-ms intervals up and down the scale from 26 ms resulted in improvement in individual recognition in 1 category, but inevitably a more than equivalent loss in recognition ability in the other category. With use of the 25 Hz HPF, performance of single-point separation of patients with nonischemic, nonsustained VT from normal subsets was less effective. Disappointed that we were not seeing the degree of late potential discrimination previously encountered when patients with VT secondary to coronary artery disease and MI were studied, we turned our attention to looking at the heart size in this group of patients with VT from multiple, noninfarction causes.An interesting relation emerged (Table IV). The patients with VT, who also had cardiothoracic ratios in excess of 0.5 in the 72-inch upright posteroanterior roentgenogram of the chest, uniformly revealed the lower terminal voltage.

FIGURE 5. A, electrocardiogram, and B, X, Y and Z leads and vectormagnitude record from C.C., a patient representative of the falsenegatives with sustained ventricular tachycardia and a diagnosis of mitral valve prolapse (MVP). Coronary arteriography revealed no significant coronary artery disease. Note the normal routine electrocardiogram and the very sharp decline of QRS with the terminal voltage well above the 25 PV RMS level which was 85.9 yV with the 25 Hz high-pass filter, and 34.48 PV RMS with the 50 Hz high-pass filter. A relatively normal set of orthogonal leads X, Y and Z, as well as the filtered leads, was recorded. F- = false-positive: other abbreviations as in Figure 1.

TACHYCARDIA

Conversely, patients with VT with normal-sized hearts had the higher terminal voltages. There was a 34.2 PV RMS jump that occurred between the patient with the highest voltage with cardiomegaly and the patient with the lowest voltage with a normal cardiac silhouette. This latter patient had the largest heart (meeting the radiographic criterion for normal) and barely met the terminal potential magnitude criterion of normal. The etiology was cardiomyopathic. Because alcohol consumption was not a consideration, and because he had a severe viral infection before an acute onset of ventricular rhythm, the cause was presumed to be viral. Cultures and titers were not obtained. The left ventricular ejection fraction derived from catheterization data was 28% and the coronary arteries were normal. The standard ECG did not reveal left ventricular enlargement. Discussion Recordings with the 50 Hz HPF were found to be only slightly more reproducible on successivedays than they were with the 25 Hz HPF. The optimal single point of

FC 25

CC. VT MV? F-

V(40)

ES. 98

V

C50)

100.94

Fo

288

_

V(40)

50

34.48

V(50)

4,. 80

February 1, 1985 THE AMERICAN JOURNAL OF CARDIOLOGY Volume 55

TABLE II

Late Potentials in Normal Subjects and in Patients with Ventricular Tachycardia Unrelated to Infarction*

TABLE Ill

Late Potentials in Normal Subjects and in Patients with Ventricular Tachycardia Unrelated to Infarction*

50 Hz HPF Normal Subjects ms PV flV ms ms

QRS last 40 ms last 50 ms term. pot. <40 I*V term. pot. <25 PV

95.38 40.77 60.08 28.43 22.08

rt f f f f

1.53 5.86 6.71 1.88 1.66

Correct Classification 50Hz HPF

VT Patients 110.0 20.10 27.23 50.29 35.43

f f f f f

389

11.03 4.94 5.65 14.36 a.25

P <::25 <0:005 NS Ns

* Group comparisons using the 5 parameters measured from data recorded with the 50 Hz high-pass filter (HPF). Groups are statistically different when voltage during the last 49 or 50 ms of QRS were considered. Abbreviations as in Table I.

Last 40 ms Last 50 ms Last 40 ms Last 50 ms QRS duration Duration <40 /*V term. pot. Duration <40 PV term. pot. Duration <25 PV term. pot.

Patients with Non&stained VT(%)

Normal Subjects (%)

57 57 57 57 57 72 57 75

86 94 94 94 100 66 90 a4

<20 yv <30 /lv <15 /.Lv <25 /.LV >liOms 130 ms 137 ms 126 ms

%20 /.rv 130 /A/ 215 yv 125 /.LV 5110 ms <30 ms <37 ms <26 ms

Using the 5 parameters tested, single values were sought that would separate the normal subset from the patient with ventricular tachycardia. The 50 Hz high-pass filter was slightly more effective than the 25 Hz high-pass filter in achieving the best group separation. The most effective point of separation was the 26ms duration of terminal potentials shorter than 25 @/. In the normal subjects, 84% had terminal potentials shorter than 25 /IV that lasted under 26 ms, whereas in 75 % of patients with VT the duration was at least 26 ms. See text. Abbreviation as in Table I. l

discrimination between the normal subject and the patient with nonsustained VT was found with the 50 Hz HPF (Table III). Group differences were seenwith both HPF frequencies, but the most significant group difference (i.e., the voltage during the last 50 ms) was with the 50 Hz filter (Table II). In comparing effectiveness of discrimination of 2 different filter frequencies, 1 consideration should be recognized. For example, if frequency content between 25 and 50 Hz is abundantly represented at the end of QRS, the duration of the recording may be slightly shorter with the 50‘than with the 25 Hz HPF, In comparing the results of 1 study with another, filter frequencies that vary between studies more widely than those represented in this study could become a major consideration. A ready explanation was not forthcoming for the 3 normal subjects who revealed late potentials less than 25 PV RMS during the last 40 ms of the QRS. Whereas the hearts of these 3 subjects (all of whom are varsity soccer players) were at the upper limits of normal, none exceeded a cardiac thoracic ratio of 0.5. One subject was black, 1 was white, and 1 was Hispanic of South American Indian extraction. We thought it logical to attempt a correlation between the total QRS duration and the duration of late potentials. We reasoned that because total QRS duration tends to increase in patients after MI, often with some degree of conduction delay in the low-amplitude terminal QRS, it might also relate to a parallel spectrum of low-amplitude terminal potential durations in normal subjects. This was not found to be the caseas evidenced by the low r values. Although the ability of terminal potential amplitude or duration to distinguish the abnormal from the normal group was quite good, the ability to correctly specify persons within the VT group was much less impressive. All subjects had very symptomatic, recurrent episodes of nonsustained VT, although 1 also had an occasional episode of sustained VT, lasting longer than 30 seconds. The fact that only 2 subjects in the present study had a history of MI (and these events occurred in.the remote past) suggests that this, group of subjects with VT is basically very different from groups in earlier studies in which ischemic disease was implicit and the tachycardias were sustained.

The precise ultrastructural and functional relations of surface recorded late potentials, notches and slurs, and the fragmented epicardial electrogram remain unclear and is a rich field for investigation. Perhaps the most significant observation in this study is that, for whatever reason, the lower terminal voltage was clearly concentrated in patients with larger hearts. This observation may bear a relation to an earlier finding, that even in the absence of infarction, ventricular enlargement from any cause is also associatedwith an alteration in ventricular activation manifested by an increase in QRS notching.7,s Correspondence between fragmented electrograms recorded directly from the heart and late potentials

TABLE IV

Relations of Heart Size to Late Potentials in Patients with Ventricular Tachycardia

VT Hypertension CAD CAD CAD MVP Cardiomyopathy Mellaril CAD AVD MVP Cardiomyopathy

t Heart CXR

Last 40 ms WV RMS 13.36 17.77 21.14 21.92 24.30 25.71 59.93 64.34 81.39 85.98 93.02

1 + + i ‘0 : : 25 Hz HPF

The most likely cause of heart disease is listed for each patient. Two of the patients with coronary artery disease were so diagnosed because of a history of remote myocardial infarction. One other had greater than 50% diameter encroachment of a single coronary artery, and a fourth had minimal plaque formation only. Note the 34.22-p\/ difference between the patient with an enlarged heart with the highest voltage (25.71) and the patient with a normal-sized heart with the lowest (59.93). AVD = aortic valve disease: CAD = coronary artery disease; CXR = chest x-ray; HPF = high-pass filter; HTN = hypertension; MVP = mitral valve prolapse: RMS = root-mean-square; VT = ventricular tachycardia; + = yes; 0 = no.

390

LATE POTENTIALS

IN NORMALS

AND IN PATIENTS

WITH VENTRICULAR

recorded from the body surface has been pointed out by Simson et a1.gGardner et allo explored the nonuniform anisotropic border zone in the canine heart and pointed out that transmembrane action potentials may be normal at this time, though conduction remains very slow, presumably because the junctions between surviving myocardial fibers in the healed infarcts are very sparse. With premature stimulation, conduction block readily occurs, and this may cause reentry. These investigators emphasize that the structure of infarcts may be as important as the characteristics of the transmembrane potentials of the surviving muscle fibers in causing reentrant excitation. Along the same line, also in canine infarcts, evidence was presented that slow conduction alone does not cause fragmented activity. The appearance of fractionated electrograms in 2week-old infarction seemsto coincide with the increase in connective tissue in the epicardial border zone, which separates the myocardial fibers and eventually distorts the orientation. These investigators believe that the individual components of the fragmented electrograms most likely represent asynchronous electrical activity in separate bundles of surviving muscle under the recording electrode. They conclude that the anatomic substrate for reentry is present where fragmented electrograms can be recorded, because these electrograms indicate slow, inhomogeneous conduction. However, they emphasize that fragmented electrograms are probably found wherever myocardial fibers are separated by connective tissue, whether or not reentry occurs in the region.lO Can these very fundamental insights be applied to explain why, in this study of nonischemic patients with VT, surface recorded terminal potentials of low amplitude appeared only when cardiomegaly was present, but were not present in subjects with normal-sized

TACHYCARDIA

hearts? A possibility worth considering is that, because fragmented electrograms are found when myocardial fibers are separated by connective tissue, in the patient with a noninfarcted heart, a critical degree of connective tissue separation must occur as a result of gradual cardiac enlargement, before late potentials can be recorded by surface techniques, even with the use of signal averaging. Conversely, after MI in the patient with ischemic heart disease, such fibrotic separation may occur well before marked enlargement occurs. In conclusion, much basic work must be done to explain the relations of the various markers of alteration of the process of ventricular excitation. References 1. Boineau JP, Cox JL. Slow ventricular activation in acute myocardial infarction: a source of reentrant premature ventricular contraction. Circulation 1973;48:702-713. 2. Berbari EJ, Scherlag BJ, Hope RR, Lazzara R. Recording from the body surface of arrhythmogenic ventricular activity during the S-T segment. Am J Cardiol 1978;41:697-702. 3. Stmson MB. Use of signals in the terminal QRS complex to identify patients with ventricular tachvcardia after mvocardial infarction. Circulation 1981;64:235-242. ’ 4. Breifhardl 0, Becker R, R Seipel L, Abendroth RR, Ostermeyer J. Noninvasive detection of late potentials in man-a new marker for ventricular tachycardia. Eur Heart J 1981;2:1-11. 5. Denes P, Santarelli P, Hauser RG, Uretz EF. Quantitative analysis of the high-frequency components of the terminal portion of the body surface QRS in normal subjects and in patients with ventricular tachycardia. Circulation 1983;67:1129-1138. 6. Simson MB, Euler D, Michelson EL, Falcone RA, Spear JF, Moore EN. Detection of delayed ventricular activation on the body surface in dogs. Am J Physiol 1981;241:H363-H369. 7. Flowers NC, Horan LG, Thomas JR, Tolleson WJ. The anatomic basis for high-frequency components in the electrocardiogram. Circulation 1969; 39:531-539. 6. Flowers NC, Horan LG. Diagnostic import of QRS notching in high-frequency electrocardiograms of living subjects with heart disease. Circulation 1971:44:605-611. 9. Stmson MB, Untereker WJ, Spielman SR, Horowitz LN, Marcus NH, Falcone RA, Harken AH, Josephson ME. Relation between late potentials on the body surface and directly recorded fragmented electrograms in patients with ventricular tachycardia. Am J Cardiol 198351: 105-I 12. 10. Gardner PI, Ursell PC, Pham TD, Fenoglio JJ Jr, Wit AL. Experimental chronic ventricular tachycardia: anatomic and electrophysiologic substrate. In: Joseohson IWE. Wellens HJJ. eds. Tachvcardias: Mechanism. Diaanosis. and Treatment. Philadelphia: Lea & Febiger, 19842960. I