Spontaneous myocardial ischemia and the signal-averaged electrocardiogram

Spontaneous Myocardial lschemia and the Signal-Averaged Electrocardiogram Gioia Turitto, MD, Edward B. Caref, MA, Egidio Zanchi, MD, Fabio Menghini, MD, George Kelen, MD, and Nabil El-Sherif, MD

The effects of transient myocardial ischemia on the signal-averaged electrocardiogram were investigated in 13 patients with coronary artery disease and spontaneous angina undergoing 3-channel ambulatory electrocardiography. lschemia was seen as ST elevation in 2 patients or ST depression in 11; it was anterior in 5 patients, inferior in 4 and undefined in 4. Signal-averaged electrocardiograms with noise keels 51 pV were obtained from Holter tapes during 54 of 61 ischemic attacks recorded in the study group (66%), and compared with 54 tracings recorded within 60 minutes of the index attacks. Baseline tracings were normal in 8 patients (62%), showed a long QRS duration in 2 (lS%), and both a long QRS duration and a late potential in the remaining 3 (23%). Comparison of recordings at baseline and during ischemic attacks revealed no significant changes in signal-averaged electrocardiographic parameters. Absence of significant differences was also noted when analysis was performed according to the type of ischemic attacks (associated with ST elevation [n = 141 or ST depression [n = 40]), their location (anterior [n = 211 or inferior [n = 23]), their duration (>lO minutes [n = 261 or 510 minutes [n = 25]), and their magnitude (>2 mm [n = 161 or 12 mm [n = 361). It is concluded that spontaneous transient myocardial ischemia, independent of its type, location, duration and magnitude, does not generate a substrate for late potentials on the signal-averaged electrocardiogram. (Am J Cardiol 1661;67:676-660)

From the Cardiology Divisions, San Camillo Hospital, Rome, Italy, and the State University of New York, Health Science Center at Brooklyn, and the Veterans Administration Medical Center, Brooklyn, New York. This study was supported in part by Grant HL31341 from the National Institutes of Health, Bethesda, Maryland, and medical research funds from the Veterans Administration, Washington, DC. Manuscript received September 14, 1990; revised manuscript received and accepted November 27,199O. Address for reprints: Gioia Turitto, MD, Wichita Institute for Clinical Research, 551 North Hillside, Suite 340, Wichita, Kansas 67214.

676

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

67

elayed and fractionated ventricular activation potentials, usually referred to as late potentials, were initially observed on epicardial recordings during experimental myocardial ischemia and infarction and were subsequently documented on body surface signal-averaged electrocardiograms from the laboratory animal as well as in the clinical setting.le5 The ability to reliably record and analyze late potentials from Holter tapes has recently been reported.6 In the present study, the technique of signal-averaging from Holter tapes was used to investigate if transient myocardial ischemia resulted in any significant changes in signal-averaged electrocardiographic variables in a group of patients with coronary artery disease and spontaneous angina. Preliminary data have been published.7

D

METHODS Patients: Candidates for this study were all patients

hospitalized with spontaneous angina who underwent a 24-hour ambulatory electrocardiogram in the absence of any therapy except sublingual nitrates if needed. The study group included 13 patients who, in the course of their admission, met the following criteria: (1) presence of 21 transient myocardial ischemic attack during the ambulatory electrocardiogram; (2) absence of acute (
TABLE

During

I Clinical, Ambulatory

Electrocardiographic Electrocardiography

and Angiographic

Diagnosis Pt.

MI

Angina

Variables

ST Changes During lschemia

1 2 3 4 5 6 7 8 9 10 11

55,M 59,M 59,M 58,M 58,M 55.M 71,F 62,M 56,M 56,M 66,M

0 A.non-Q 0

0 I.nonQ 0 0 L.non-Q 0

M M M M M S S M M M M

12

69.M

I.L

M

LX

13

79,M

M

4,X.Y

A = anterior; I = inferior; L = lateral; LAD = left anterior mixed (spontaneous and exertional); MI = myocardial electrocardiogram leads; 0 = absent: - = not determined;

in 13 Patients

with Transient

Coronary Angiographic (“I0 Stenosis) LM 0 0 0 0 0 0 0 0 0 90 0 20

LAD

Myocardial

Attacks

Data

LC

85 95 90 70 75 0 30 0 0 0 90 100

lschemic

0 0 40 0 50 0 40 0 90 0 80 100

Right

Grafts

0 0 50 100 100

-

90 95 99 0

-

0 100 100

-

50 (LIMA to LAD) 50 (SVG to LAD)

descending coronary artery; LC = left circumflex artery; LIMA = left internal mammary artery; LM = left main trunk; M = infarction; S = spontaneous; Right = right coronary artery; SK = saphencu vein graft; x, y and z = ambulatory t = elevation; 1 = depression.

eral” lead); Y lead, with the positive and negative electrodes applied, respectively, at the V3 position and on the manubrium sterni, resembling Holter lead CM3 (“anterior” lead); Z lead, with the positive and negative electrodes applied, respectively, on the sacrum and the manubrium sterni, resembling Holter lead A (“inferior” lead). This configuration represented a modification of the orthogonal system originally described by Simson and was selected to maximize the detection and analysis of transient myocardial ischemic attacks. ST changes were identified by examining (1) the oscilloscopic replay of the tracing scanned at 120 times real-time; (2) the plots of the ST level and heart rate obtained from all 3 channels; (3) the transcription at a paper speed of 25 mm/s of all parts of the recording judged to be consistent with myocardial ischemia on the basis of the highspeed analysis and the ST level printouts. Ischemic attacks were defined as any episode of transient ST elevation or depression I1 mm from baseline or from the control level if resting ST-T abnormalities were present, having gradual onset and resolution, and lasting 11 minute.*T9 Signal-averaged electrocardiogram: Analog signals from the tapes were subjected to digital conversion during high-speed scanning at 120 times real-time, with a 1Zbit resolution and at a sampling rate of 694 Hz (real-time equivalent)/channel.6 Signals were amplified, averaged and filtered with a bidirectional filter at 25 Hz. The filtered leads were then combined into a vector magnitude dX2 + Y2 + Z2 and the QRS onset and offset were determined by computer algorithm. The QRS duration and root-mean-square voltage of the signals in the last 40 ms of QRS (RMS40) were calculated. The duration of low-amplitude signals (<40 PV) was also measured but not used for analysis. An abnormal signal-averaged electrocardiogram was defined as a recording showing a QRS duration >115 ms, an RMS40 <25 PV, or both. lo Late potentials were de-

fined as signals with abnormal RMS40.5 Signal-averaged electrocardiograms were obtained during all ischemit attacks. Acquisition of data began after ST changes were 21 mm, included the peak of ischemic attacks, i.e., the maximum degree of ST elevation or depression, and terminated at the onset of the resolution of ST changes. Baseline signal-averaged electrocardiograms were obtained within 60 minutes of each ischemit attack, at least 20 minutes before the onset or after the resolution of ST changes. Recordings were accepted for analysis if they showed a noise level I1 PV. Statistical analysis: Values of the signal-averaged electrocardiographic parameters on recordings performed at baseline and during ischemic attacks were compared using the analysis of variance and Student’s t test for paired data. A p value <0.05 was considered statistically significant. All variables were expressed as mean f standard deviation. RESULTS The clinical, electrocardiographic and angiographic characteristics of the 13 study patients are listed in Table I. During the ambulatory electrocardiogram, ischemit ST changes appeared on the anterior lead in 5 patients and were concordant with angiographic tindings, since these patients had a significant (>50%) stenosis of the left anterior descending coronary artery alone or in combination with an occlusion of the right coronary artery and previous inferior myocardial infarction. Ischemic ST changes appeared on the inferior lead in 4 patients and were also concordant with angiographic findings, since these patients had an isolated significant stenosis of the right coronary artery (n = 3) or the left circumflex artery (n = 1). The location of ischemia could not be determined in the remaining 4 patients because of the presence of diffuse ST changes or severe coronary artery disease, or the absence of angiographic data. THE AMERICAN

JOURNAL

OF CARDIOLOGY

APRIL

1, 1991

677

TABLE II Values of Signal-Averaged Electrocardiographic Parameters at Baseline and During Transient Myocardial lschemic Attacks Analyzed According to Their Type, Location, Duration and Magnitude

BAS

4 X

QRSD (ms)

STelevation (n = 14) ST depression (n = 40) Anterior (n = 21) Inferior (n = 23) Long(>lOminutes)(n=29) Brief(2 mm) (n = 18) Mild (52 mm) (n = 36)

RMS4O @/)

Baseline

lschemia

Baseline lschemia

115&16

116f18

45f29

38f33

113h19 113zkl8 115f17

114f17 113f17 116f18

40f28 43~31 38f26

42f27 44~1~31 38f27

114f16

114f18

39f32

4Of30

114f19

114f17

42f27

41f27

113f17 114f17

114&17 114f17

44*31 41&28

39f30 43*29

:--.‘/;

All values are mean f standard deviation. All differences were statistically not significant. QRSD = high-frequency QRS duration; RMS4O = root-mean-square voltage of last 40 ms of QRS.

ISCHEMIA RMS40:

A total of 61 transient myocardial ischemic attacks was recorded (range, 1 to 9 per patient). Among the 61 ischemic attacks, 43 (70%) were silent and 18 were symptomatic. Transient ischemia manifested as ST elevation in 2 patients and as ST depression in 11 patients. No patients developed ventricular tachyarrhythmias (13 ventricular premature complexes at a rate of >lOO beats/min) related to ischemic attacks. Signal-averaged electrocardiogram: Baseline signalaveraged electrocardiograms were abnormal in 5 of the 13 study patients (38%). A prolonged QRS duration was seen in 2 patients (15%), and both a prolonged QRS duration and a late potential in 3 patients (23%). Recordings with a noise level I1 PV could be obtained during 54 of 61 ischemic attacks (88%); the characteristics of these ischemic attacks are listed in Table II. These 54 recordings were matched for comparison with 54 tracings obtained within 60 minutes of the index ischemic attacks. QRS duration and RMS40 were similar at baseline and during ischemia (114 f 18 vs 114 f 17 ms and 41 f 29 vs 41 f 29 pV, respectively; difference not significant). Noise level and the number of averaged beats were also similar at baseline and during ischemia (0.8 f 0.2 vs 0.9 f 0.3 PV and 434 f 119 vs 415 f 112 beats, respectively; difference not significant). Data were analyzed separately according to the type of ischemic attacks, their location, their duration and their magnitude. None of these comparisons revealed any significant differences regarding QRS duration and RMS40 between baseline tracings and those obtained during ischemia (Table II). All normal recordings at baseline remained normal during ischemia, whereas all recordings with abnormal QRS duration or RMS40 at baseline were also abnormal during transient myocardial ischemia. Figures 1 and 2 show examples of signal-averaged recordings before and during transient myocardial ischemia. The latter was evident as ST elevation in Figure 1 and ST depression in Figure 2. The occurrence of ischemia did not result in any significant changes in the signal-averaged electrocardiographic paAmbulatory

678

Noise:

electrocardiogram:

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

67

50.0

P”

0.8 w

FIGURE 1. Signal-averaged electrocardiiams obtained from Holter tapes in patient 6 at baseline and during transient myocardial ischemia. The latter appears as ST elevation on the Z lead of the 34annel real-time tracing. Signal-averaged electhiographic parameters are normal at baseline and do not show any significant changes during ischsmia. LAS = duration of low-amplii signals 40 & QRSD = high-frequency QRS &ration; RIMS40 = root-mean-square voltage of lasl40 ms of QRS.

rameters: They were constantly normal in Figure 1 and abnormal in Figure 2. DISCUSSION Present study: This study documented the feasibility of obtaining signal-averaged electrocardiograms during episodes of spontaneous transient myocardial ischemia recorded on 3-channel Holter tapes. Our data showed no changes in signal-averaged electrocardiographic parameters between recordings at baseline and during transient myocardial ischemia. An analysis performed according to the characteristics of ischemia (type, location, duration and magnitude) did not change these results: ischemia did not lead to the onset of late potentials, even when it was prolonged and severe. Previous studies: The effects of transient myocardial ischemia on the signal-averaged electrocardiogram

were the objective of a small number of investigations.11-13 They focused on ischemia observed during percutaneous transluminal angioplasty,’ 1 dipyridamole testI or exercise test.13 In a study of Abboud et al,” a single surface electrocardiographic lead (Vs) was used to compute the RMS voltage of the QRS at a filter setting of 150 to 250 Hz. Ischemia due to angioplasty balloon inflation was reported to cause a significant decrease in the calculated RMS voltage of QRS, as well as the appearance of zones of reduced amplitude in the midportion of QRS. The significance of these findings is uncertain, since they were based on techniques for recording and analyzing the signal-averaged electrocardiogram which were at variance with those used in most recent reports.5J0J2-15 In the studies of Turitto12 and Caref13 and their co-workers, serial signal-averaged electrocardiograms were recorded before and during transient myocardial ischemia induced by dipyridamole infusion12 or by exercise13 in patients with documented coronary artery disease and angina. Ischemia did not significantly affect the signal-averaged electrocardiogram: Late potentials failed to appear on tracings obtained during positive dipyridamole or exercise tests. This was independent of the type of ischemia (with ST elevation12 or ST depression12J3), the presence of abnormal signal-averaged parameters at baseline,12J3 the presence and site of previous myocardial infarction,13 and the presence of ventricular arrhythmias during ischemia.13 On the other hand, some differences between both these reports and the present study must be considered. Induced ischemia was usually brief and relieved by drugsi or termination of exercise,i3 whereas spontaneous ischemia may considerably vary in duration.8,g There may be different pathogenetic mechanisms for provoked and spontaneous ischemia.16J7 Study limitations: The technique of signal-averaging from Holter tapes has inherent problems. Noise levels may be significantly higher than those seen with conventional recording systems; in fact, the mean number of beats required in our study to obtain noise levels I1 PV was >400, an amount that was shown to be associated with noise levels of 0.3 to 0.5 PV on conventional recordings.14J5 The inability to achieve low noise levels led to the loss of 12% of our sample of ischemic attacks. Differences in the frequency band width between the conventional and Holter instrumentation result in a tendency toward an increase in QRS duration and a decrease in RMS40 voltage on signal-averaged tracings from Holter tapes6 This issue is not relevant to our study, because each patient represented his own “control.” Finally, late potentials may vary in timing and duration and may not be amenable to time-domain signal-averaging, whereas they may be best studied on beat-to-beat recordings.18Jg Despite these limitations, the present study offered evidence that spontaneous transient myocardial ischemia does not affect the signal-averaged electrocardiogram. This confirmed observations relative to laboratory settings12J3 and suggests that the electrophysiologic changes associated with ischemia do not represent a

BASELINE

mm:

119 Ills

X

k

y-J/-

!I

I/I I

_I FIGURE 2. Signal-averaged ektrocardiograms obtained from Rolter tapes in patient 4 at baseline and duing transient myocardiai ischemii. The latter appears as ST depression on the X and Y lfds cd the 3-channel real-time tracing. Signal-averaged eiectrocardiigraphic parameters are abnormal at baseline and do not show any significant changes during ischemii. Abbreviations as in Figure 1.

substrate for late potentials on the signal-averaged electrocardiogram. REFERENCES 1. Ideker RE, Mirvis DM, Smith WM. Late, fractionated potentials. Am J Cardiol 1985;55:1614-1621. 2. El-Sherif N, Games JAC, Restive M, Mehra R. Late potentials and arrhythmogenesis. PACE 1985;8:440-462. 3. Simson MB, Euler DE, Michelson EL, Falcone RA, Spear JF, Moore EN. Detection of delayed ventricular activation on the body surface in dogs. Am J Physiol 1981;24l:H363-H369. 4. Simson 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 1983;51:105-112. 5. Simson MB. Use of signals in the terminal QRS complex to identify patients with ventricular tachycardia after myocardial infarction. Circulation 1981; 64~235-242. 6. Kelen G, Henkin R, Lannon M, Bloomfield D, El-Sherif N. Correlation between the signal-averaged electrocardiogram from Halter tapes and from realtime recordings. Am J Cardiol 1989;63:1321-1325. 7. Turitto G, Caref EB, El-Sherif N. Does transient myocardial ischemia create a substrate for late potentials? (abstr). Circulation 1990;82(suppl III):III-355.

THE AMERICAN

JOURNAL

OF CARDIOLOGY

APRIL 1. 1991

679

8. Turitto G, El-Sherif N. Alternans of the ST segment in variant angina. Incidence, time course and relation to ventricular arrhythmias during ambulatory electrocardiographic recording. Chest 1988;93:587-591. 9. Turitto G, Zanchi E, Maddahma A, Pellegrini A, Risa AL, Prati PL. Prevalence, time course and malignancy of ventricular arrhythmia during spontaneous ischemic ST-segment depression. Am J Cardiol 1989;64:900-904. 10. Caref EB, Turitto G, Ibrahim BB, Henkin R, El-Sherif N. Role of bandpass filters in optimizing the value of the signal-averaged electrocardiogram as a predictor of the results of programmed stimulation. Am J Cardiol1989;64: 16-26. 11. Abboud S, Cohen RJ, Selwyn A, Ganz P, Sadeh D, Friedman PL. Detection of transient myocardial ischemia by computer analysis of standard and signalaveraged high-frequency electrocardiograms in patients undergoing percutaneous transluminal coronary angioplasty. Circulakm 1987;76:585-596. 12. Turitto G, Zanchi E, Risa AL, Maddahma A, Saltarocchi ML, Vajola SF, Prati PL. Lack of correlation between transient myocardial ischemia and late potentials on the signal-averaged electrocardiogram. Am J Cardiol1990;65:290-

296. 13. Caref EB, Goldberg

680

THE AMERICAN

N, Mendelsbn

JOURNAL

L, Hanley

G, Okereke

OF CARDIOLOGY

R, Stein RA, El-

VOLUME

67

Sherif artery

N. Effects of exercise on the signal-averaged electrocardiogram in coronary disease. Am J Cardiol 1990;66:54-58. 14. Steinberg JS, Bigger JT Jr. Importance of the endpoint of noise reduction in analysis of the signal-averaged electrocardiogram. Am J Cardiol 1989;63:556-

560. 15. Borbola J, Denes P. Short- and long-term reproducibility of the signalaveraged electrocardiogram in coronary artery disease. Am J Cardiol 1988; 61:1123-1124. 16. Picano E. Dipyridamole-echocardiography test: historical background and physiologic basis. Eur Heart J 1989;10:365-376. 17. Maseri A, L’Abbate A, Chierchia S, Parcdi 0, Severi S, Biagini A, D&ante A, Marzilli M, Ballestra AM. Significance of spasm in the pathogenesis of ischemic heart disease. Am J Cardiol 1979;44:788-792. 18. El-Sherif N, Mehra R, Games JAC, Kelen G. Appraisal of a low noise electrocardiogram. J Am Coil Cardiol 1983;1:456-467. 19. Hombach V, Keppel U, Hopp HW, Winter U, Hirche H. Noninvasive beat by beat registration of ventricular late potentials using high resolution electrocardiography. Int J Cardiol 1984;6:167-183.