Effect of thrombolytic therapy on the predictive value of signal-averaged electrocardiography after acute myocardial infarction

Effect of thrombolytic therapy on the predictive value of signal-averaged electrocardiography after acute myocardial infarction

Effect of Thrombolytic Therapy on the Predictive Value of Signal-Averaged Electrocardiography After Acute Myocardial Infarction Marek Malik, PhD, MD, ...

618KB Sizes 24 Downloads 53 Views

Effect of Thrombolytic Therapy on the Predictive Value of Signal-Averaged Electrocardiography After Acute Myocardial Infarction Marek Malik, PhD, MD, Piotr Kulakowski, MD, Olusola Odemuyiwa, MD, Jan Poloniecki, PhD, Anne Staunton, RN, Teri Millane, MD, Thomas Farrell, MD, and A. John Camm, MD

Standard time domain variables from signai-averaged electrocardiography were examined in a population of 331 survivors of acute myocardial infarction. Of these subjects, 130 received early (<24 hours) thrombolytic therapy. During a follow-up of L 10 months, there were 17 arrhythmic events (8.5%) (sudden death or sustained symptomatic ventricular tachycardia) in the group without thrombolysis and 8 (6.2%) in those with thromboiysts. Statistkally, highly significant differences between the stgnal-averaged electrocardiographic variables of patients with and without arrhythmic events were found in the group without thrombolysis, whereas only root-mean-square voltage of the terminal 40 ms of the signal-averaged QRS complex was stattstically associated with outcome (the differences in the other 2 indexes being not slgnifkant) in patients with thrombolysis. When udng 2 previously published categoric criteria for the diagnosis of abnormal signal-averaged electrocardiography, the performance of these criteria in predicting arrhythmic events was substantially better in the group without thrombolysis than in those with thrombolysis (positive predictive accuracy >3 times lower). Retrospectively adjusted receiver-operator characteristics showed that for a sensitivity of 30% the maximum achievable positive predictive accuracy of signal-averaged electrocardiography for arrhythmic events was 100% in the group without thrombolysis, but only 27% in those with thrombolysis. It is concluded that standard signal-averaged electrocardiography after acute myocardial infarction is less informative in patients who receive thrombolytic treatment. (Am J Cardiol 1662;70:21-25)

T

he value of signal-averagedelectrocardiography for the prediction of arrhythmic complications after acute myocardial infarction (AMI) has been clearly established.‘-’ The presenceof late potentials is an important risk stratification factor, but the effect of therapeutic interventions on the diagnosisof late potentials and on their predictive value should be considered. This particularly applies to thrombolytic therapy, and severalstudies already showedthe effectsof thrombolysis on the incidence of late potentials in patients after AMI.E-‘2 With the now frequent use of thrombolytic therapy in AMI, it has been suggestedthat the categoric criteria for risk stratification after AM1 may need moditication.i3 The effect of thrombolysis on the predictive value of late potentials recorded on signal-averaged electrocardiographyhas not been established.This study comparesthe value of signal-averagedelectrocardiographic late potentials for the prediction of serious arrhythmic events in patients who did and did not receive early thrombolytic therapy during AMI. METHODS

Patlen& We examined 331 patients (aged 29 to 74 years) who were admitted to the hospital with AM1 during the period of 1986to 1990for whom signal-averaged electrocardiography was available and who survived to discharge.AM1 was diagnosedusing previously describedstandard criteria.14 Patients were excluded if they had noncardiac diseaselikely to increasemortality, important nonischemiccardiac disease,history of cardiac surgery, or permanentpacemakerinsertion, or if they were aged >75 years or were unable to be followed up. Patients with bundle branch block or ventricular preexcitation were excluded, becausethe wide QRS patterns disturb the time domain signal-averagedelectrocardiographic analysis. History of AM1 was present in 39 patients who were included. Patients were followed up for L6 months (mean 20) during which 32 died; 12 died suddenly (according to the Cardiac Arrhythmia Pilot Study definition),i5 and 13 had spontaneoussustainedventricular tachycardia. Thrambolytie Beginning in the secondhalf of 1988, eligible patients were given thrombolytic theraFrom the Department of Cardiological Sciences,St. George’sHospital py. In the absenceof contraindications, 1.5 million units Medical School, London, England. This study was supported by the of streptokinase were administrated to patients seen British Heart Foundation,London, England. Manuscript receivedDe- within 24 hours of the onset of major symptoms.Concember2,1991; revisedmanuscript receivedand acceptedFebruary 26, traindications were active bleeding or history of coagu1992. Address for reprints: Marek Malik, PhD, MD, Department of lopathy, suspectedaortic dissection, recent head injury Cardiological Sciences,St. George’s Hospital Medical School, Cran- (<3 months), documentedproliferative diabetic retinop athy, fractured ribs secondary to traumatic resuscitamer Terrace, London SW 17ORE, England.

therapy:

THROMBOLYSIS AND LATE POTENTIALS AFTER AMI 21

events. This comparison was performed for the total population and separately for groups of patients who did and did not receive thrombolytic therapy. The nonparametric Mann-Whitney U test was usedfor this purpose.

TABLE I Characteristics of Patient Population Pt. Group

No TT

TT Early

TT Late

Total

No. Mean age (yr) Women Anterior AMI Q-wave AMI

201 58 41 104 128

100 56 19 53 75

30 55 8 8 21

331 57 68 165 224

CATEGORIC CRITERIA FOR SIGNAL-AVERAGED ELECTROCARDIOGRAPHIC DIAGNOSIS: The signal-averagedelectro-

Characteristics of patients are shown for those receiving no thrombolytic therapy (77, and for those receiving it ~6 (Early) and >6 (Late) hours after onset of symptoms. AMI = acute myocardial infarction.

TABLE II Numbers of Patients in Individual Groups Pt. Group

VT

SD

Others

Total

No thrombolysis Thrombolysis All pts.

8 5 13

9 3 12

184 122 306

201 130 331

SD = sudden death; VI = sustained

symptomatic

ventricular

tachycardia.

tion, systolic blood pressure <85 mm Hg or history of sensitivity to streptokinase. Streptokinase was also not used if it was previously administered within 5 to 365 days. Thrombolytic therapy was used in 130 patients of the study population. Signal-averaged eleebowdiogrclphy: A signal-averaged electrocardiographic record was obtained before discharge (day 5 to 11 after AMI) in each patient. Recordings were obtained on the Arrhythmia Research Technology system (model 1200 EPX), using Frank orthogonal leads and a sampling rate of 1 kHz. In each patient, 200 to 500 ventricular complexes were averaged,the achievednoiselevel was 10.5 pv. The analysis was performed with filter settings of 25 to 250 Hz, and the following standard quantitative variables were computed in each case:root-mean-squarevoltage of the terminal 40 ms of the signal-averaged QRS complex (RMS-40), total duration of the signal-averagedQRS complex (tQRS), and duration of low-amplitude signals <40 pv (LAS). Several patients (n = 100) were treated with p blockers, but in all cases,signal-averagedelectrocardiography was performed when the patient was receiving no other antiarrhythmic therapy. Left ve&ieuIar ejection fraction: Before discharge from the hospital, patients performed a symptom-limited treadmill exercise test, using a Bruce protocol. Depending on the result of this test, left ventricular ejection fraction was obtained from either coronary angiography with left ventricular angiography or radionuclide angiography, using previously published techniques.16 Statistks and data madpdatbm Becausethe incidenceof late potentials has been shown to be associated with arrhythmic death or ventricular tachycardia,1-7the follow-up end point consideredin this study was an arrhythmic event defined as either sudden cardiac death or symptomatic sustainedventricular tachycardia. VALWS OFSIGNAL-AVERAGED ELECTROCARDIOGRAPHIC VARIABLES: The numeric values of individual signal-av-

cardiographic recording of eachpatient was classifiedas normal or abnormal, and the associationof this diagnosis with the arrhythmic eventswas examinedusing Fisher’s exact test. This examination was again performed for the total population and separately for subgroupsof patients who did and did not receivethrombolytic theraPY* This whole procedure was performed twice, using 2 previously published criteria for the abnormality of signal-averagedelectrocardiography: (1) tQRS I1 14 ms; LAS 138 ms; and RMS-40 220 pv17;and (2) tQRS 1120 ms; LAS 140 ms; and RMS-40 225 ~v.~ With both criteria, the result was categorizedabnormal if any 2 of the 3 variables had an abnormal value, and the incidence of abnormal signal-averaged electrocardiographic findings in the groups with and without thrombolysis was compared using Fisher’s exact test. STRATIFICATION OF ARRHYTHMIC EVENTS: The socalled receiver-operatorcharacteristic curves that associate specificity with sensitivity were computed for the multivariate prediction of arrhythmic events, using all 3 signal-averaged electrocardiographic variables. For the computation of these curves, the dichotomy points of individual signal-averagedelectrocardiographic variables were varied to achieve all possiblevalues of sensitivity, and for each of these values, the maximum specificity was computed using a previously published algorithm.i8 The computation of receiver-operatorcharacteristic curveswas performed for the prediction of arrhythmic eventsin the total population and in the 2 sub groups (thrombolytic therapy used and not used). In all cases,>2 signal-averagedelectrocardiographic variables were neededto exceedtheir dichotomy points for a positive result. The samemethodswere used to compute the curves expressingthe dependenceof the maximum achievable positive predictive accuracy on sensitivity. RESULTS Basic characteristics of the patient population are presented in Table I. The incidence of arrhythmic events in the total population (7.6%), and in the subgroups of patients who did (6.2%) and did not (8.5%) receivethrombolytic therapy is listed in Table II. In the group with thrombolysis, all arrhythmic eventsoccurred in patients who received streptokinase <6 hours after the onsetof symptoms.There were no significant differencesbetweenthe groups with and without thrombolysis regarding age (56 f 10 vs 57 f 9 years), left ventricular ejection fraction (49 f 15 vs 50 f 140/o),infarct site (47 vs 52% with anterior AMI), treatment with b blockers (28 vs 32%), or history of AM1 (7 vs 14%). Comparison

of signal-averaged

ekdrocardiograph-

eraged electrocardiographic quantitative variables were ie variables: The statistical comparison of individual compared for patients with and without arrhythmic signal-averaged electrocardiographic variables in pa22

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70

JULY 1. 1992

tients with and without arrhythmic events is shown in Table III. Although the differences were highly signilicant in patients who did not receive thrombolytic therapy, only the RMS-40 values were significantly different (p value close to nonsignificance) in patients with thrombolysis. Pehmmnw of dhgnodk categoric criteda: The performance of the categoric criteria for signal-averaged electrocardiographic diagnosis(listed in the Methods section) is shown in Table IV. Roth criteria performed better in the subgroup of patients who did not receive streptokinase; the positive predictive accuracy was >3 times higher in this group than in those with thrombolysis. With both criteria, there was no significant differencebetweenthe incidence of abnormal signal-averaged electrocardiographic diagnosis between patients with and without thrombolysis (criterion 1: 19.2 vs 11.4%; and criterion 2: 18.5 vs 11.9%). Rebwpedh risk straHeatkm With hindsight, dichotomy points can be chosenthat optimize the sensitivity for any given specificity. The relation between the optimum positive predictive accuracy and sensitivity is shown in Figure 1. Although it was possibleto adjust the diagnostic criteria in order to achieve lOO?&positive predictive accuracy for up to 30% sensitivity in the subgroup of

TABLE III Values of Signal-Averaged Electrocardiographic Variables in All Patients, and in Those With and Without 1 Thrombolysis Variable tQRS

Group

VTISD

All lT not used

115 2 18.4 116 -1- 19.4

TTusad RMS-40

All TT l7 All ll TT

LAS

Others

102 100 104 67.7 72.6 60.4 25.3

114 -+ 17.3 29.5 k 23.8

not used used

29.5 2 27.1 29.4 f 16.1

not used used

37.8 f 17.3 40.2 ” 19.0 32.8 f 12.7

+ T f k i: f f

p Value

17.1 16.4 18.0 51.4 54.0 46.4 11.8

0.0002 0.0008 0.0699 0.0000 0.0001 0.0419 0.0001 0.0001 0.1186

23.7 k 10.3

27.6 f 13.4

For each signal-averaged electrocardiographic variable, table shows mean value k standard deviation found in patients with arrhythmic events (VT/SD) and in other patients @then). These values are shown for total patient pulation (All). and for those who did and did not receive thmmbolytic therapy f R 1. Values of statistical significance (Mann-Whitney U test) of differences betwean patients with arrhythmic events and others are shown. LAS = duration of low-amplitude signals <40 pv; RMS-40 = root-mean-square voltage of terminal 40 rns of signal-averaged QRS complex; SD = sudden death; tQRS = total duration of signal-averaged QRS complex; VT = ventricular tachycardia.

patients without thrombolysis, the maximum achievable positive predictive accuracy in the subgroup with thrombolysis was <30%. DISCUSSION Sturdyfindings: Several studies reported the relation betweenthrombolysis and patency of the infarct-related

TABLE IV Performanceof Categoric Criteria for the Diagnosis of Signal-Averaged Electrocardiography in All Patients, and in Those With and Without Thrombolysis PPA

p Value

88

25

92

39

12

4.9 x 10-G 7.4 x 10-5 0.01

25 38 12

4.9 x 10-e 0.0001 0.008

Criterion

Population

TP

TN

FN

FP

Sen

(1)

All TT not used TT used

12 9

270

13

36

48

170

8

14

53

3

100

5

22

38

82

All TTnotusad TT used

12 9

270 169 101

13 8 5

36 15 21

48 53 38

88 92 83

(2)

3

Spc

For categodc criteria (1) and (2) (see text), numbers of true positive (TP), true negative (TN), false negative (FN) and false positive (FP) patients when using the criterion for the prediction of arrhythmic events in the complete patient population (All), and in gmups of those who did and did not receive thmmbolytic therapy 0-T) are shown. Values of statistical significance of association of positive outcome of criterion with arrhythmic events (Fisher exact test) are shown. PPA = positive predictive accuracy (%I; Sen = sensitivity f%); Spc = specificity f%).

Positive predictive accuracy w]

0 0

/ 20





I 40

a

8

Sensitivity

I, 60

I,

I 80

*

r

[“h]

THROMBOLYSISAND LATE POTENTIALSAFTERAMI

23

artery,12J9and showedthe reduced incidenceof late potentials in patients with thrombolysis.10-12It was suggestedthat reperfusion of the infarct-related artery has a beneficial effect on the electrophysiologic substrate for serious ventricular arrhythmias, becausesuccessful thrombolysis reducesthe probability of the development of a large area of slow conduction with arrhythmogenic properties, which may manifest as a detectablelate potential. However, with both categoric criteria for signal-averaged electrocardiographic diagnosis, there was no significant difference between the incidence of abnormal findings in the groups with and without thrombolysis. This study examined a large post-AM1 population compared with previous studies that reported a lower incidence of late potentials among patients with thrombolysis. The group with thrombolysis included all patients who receivedstreptokinase,and we did not verify the successof thrombolysis by cardiac catheterization. Thus, the results are consistent with those reported by Turitto et al9 who found significant signal-averaged electrocardiographic differences only between patients with and without successfulthrombolysis, but not between groups with and without thrombolysis. Our protocol also included administration of streptokinaseup to 24 hours after the onset of symptoms. Hence, the impact of thrombolysis in our study group may have been different compared with that in studies restricted to much earlier administration.9g20Although all patients with thrombolysis who had arrhythmic events received streptokinase within 6 hours after the onset of symp toms, the inclusion of those receiving thrombolytic therapy late may have influenced the predictive importance of signal-averagedelectrocardiography. The reduction of arrhythmic events in AM1 survivors with successfulthrombolysis1° may be solely explained by the reduction in the developmentof arrhythmogenic regions with slow conduction. Some patients may have a different arrhythmogenic substrate that does not manifest on signal-averagedelectrocardiography and that may be more or less affected by reperfusion. A hypothesis of a complex and indirect effect of thromboiytic intervention on the substrate for arrhythmic eventsmay also be supported by the observationof newly developedlate potentials after successfulrecanalization.l l The reduction of abnormalities on signal-averaged electrocardiography after successfulreperfusion has also been observedwithout the reduction of complex acute ventricular arrhythmias.20This, together with the observation of new late potentials after recanalization,ll suggests that the therapeutic mechanismsof thrombolysis are more complex than were originally believed. This agrees with our finding that both the sensitivity and specificity of categoric criteria for late potential diagnosis were reduced in patients with thrombolysis. Although the relation between positive signal-averaged electrocardiographic findings and arrhythmic eventsafter AM1 has been clearly demonstrated-’ not all mechanisms involved are clearly understood. It is even possible that by remodeling the infarcted area, thrombolytic therapy may be proarrhythmic in some 24

THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 70

casesand be antiarrhythmic in others, especially when not usedimmediately after the onset of symptoms(e.g., >4 hours, as was the case with a large portion of our population). A combination of such mechanismswith a complex influence of thrombolytic therapy on the incidenceof late potentials may explain the findings of this study, and the difference from studies that administered thrombolytic agents much earlier.‘* Our investigation was confined to the standard time domain signal-averaged electrocardiographic variables and the 25 to 250 Hz filter setting. Spectral analysis, especiallywhen applied to the complete signal-averaged QRS complex,*l may provide useful independent information. Other innovative techniques of signal-averaged electrocardiographic analysis** may also provide progno&ally useful information in patients with thrombolysis. Study limitationr: The most important limitation of this study is the small number of arrhythmic events(especially in the group with thrombolysis). However, the differences between the predictive value of signal-averaged electrocardiography in the groups with and without were marked (Figure l), and therefore, we believe that they are not causedby the lack of data. The group without thrombolysis included both patients admitted before the secondhalf of 1988 and those admitted later who were not eligible for treatment with streptokinase. Thus, the data of the groups with and without thrombolysis were not collected exactly contemporarily. This may have had a bearing on the incidence of late potentials. We did not relate the findings of this study to other recognizedrisk stratification factors (e.g., left ventricular ejection fraction, ventricular ectopy or heart rate variability). Although the effect of thrombolytic therapy on novel stratification factors needs investigation, the reperfusion of an infarct-related artery has been reported not to have a significant effect on left ventricular ejection fraction. I9 This is in agreement with our observationsand suggeststhat the predictive value of late potential diagnosis in the group with thrombolysis would not be substantially improved by its combination with left ventricular ejection fraction. However, multivariate analysisof a large cohort of patients with thrombolysis is neededto examine this properly. All recordingswere obtained when patients were not treated with antiarrhythmic drugs that are likely to influence signal-averagedelectrocardiographic variables; we did not investigate differencesin the predictive power of late potentials in patients with thrombolysis who were and were not treated with @blockers. However, other reports showedthat thesedrugs do not significantly influence signal-averagedelectrocardiographic analysis.23Becausethe incidenceof treatment with @blockers was similar in the groups with and without thrombolysis, the effect of p blockers could not have influenced our major findings. The 2 categoric criteria usedfor the diagnosisof signal-averaged electrocardiography were selected arbitrarily from many different criteria published by other investigators. Furthermore, criterion (l), which was proposedby the joint American College of Cardiology,

JULY 1, 1992

American Heart Association and European Society of Cardiology task force committee,17was suggestedfor signal-averaged electrocardiographic variables derived at filter settings of 40 to 250 Hz, whereas our study used settings of 25 to 250 Hz. However, the findings obtained with both criteria were consistent,and the retrospectively computed receiver-operatorcharacteristics, as well as the curves presentedin Figure 1, showedthat any categoric criterion was likely to perform better in the subgroup of patients who did not receive streptokinase. REFERENCES 1. Breithardt G, Schwarzmaier J, Borggrefe M, Haerten K, Seipel L. Prognostic significance of late ventricular potentials after acute myocardial infarction. Eur Heart J 1983;4:487-495. 2. Kanovsky MS, Falcone RA, Dresden CA, Josephson ME, Simpson MB. Identification of patients with ventricular tachycardia after myocardial infarction: signal-averaged electrocardiogram, Halter monitoring, and cardiac catheterization. Circulation 1984;70:264-270, 3. Denniss AR, Richards DA, Cody DV, Russell PA, Young AA, Cooper MJ, Ross DL, Uther JB. Prognostic significance of ventricular tachycardia and Iibrillation induced at programmed stimulation and delayed potentials detected on the signal-averaged electrocardiograms of survivors of acute myocardial infarction. Circulation 1986;74:131-745. 4. Comes JA, Winters SL, Stewart D, Horowitz S, Milner M, Barreca P. A new noninvasive index to predict sustained ventricular tachycardia and sudden death in the first year after myocardial infarction: based on signal-averaged electrocardiogram, radionuclide ejection fraction and Holter monitoring. J Am Co11 Cardiol 1987;10:349-357. 5. Kuchar DL, Thorburn CW, Sammel NL. Prediction of serious arrhythmic events after myocardial infarction: signal-averaged electrocardiogram, Holter monitoring and radionuclide ventriculography. J Am Coil Cordial 1987;9: 531-538. 6. Cripps T, Bennett ED, Camm AJ, Ward DE. High gain signal averaged electrocardiogram combined with 24 hour monitoring in patients early after myocardial infarction for bedside prediction of arrhythmic events. Br Heart J 1988;60:181-187. 7. El-Sherif N, Ursell SN, Bekheit S, Fontaine J, Turitto G, Henkin R, Caref EB. Prognostic significance of the signal-averaged ECG depends on the time of recording in the postinfarction period. Am Heart J 1989;118:256-264. 8. Gang Es, Lew AS, Hong M, Wang FZ, Siebert CA, Peter T. Decreased incidence of ventricular late potentials after successful thrombolytic therapy for acute myocardial infarction. N Engl J Med 1989;321:712-716. 9. Turitto G, Risa AL, Zanchi E, Prati PL. The signal-averaged elcctrocardiogram and ventricular arrhythmias after thrombolysis for acute myocardial infarc-

tion. J Am Coil Cardiol 1990;15:1270-1276. 10. Chew E-W, Morton P, Murtagh JG, Scott ME, O’Keeffe DB. Intravenous streptokinase for acute myccardial infarction reduces the occurrence of ventricular late potentials. Br Heart J 1990;64:5-8. 11. Tranchesi B, Verstraete M, Van de Werf F, de Albuquerque CP, Caramelli B, Gebara OC, Pereire WI, Moffa P, Bellotti G, Pileggi F. Usefulness of highfrequency analysis of signal-averaged surface electrocardiograms in acute myocardial infarction before and after coronary thrombolysis for assessing coronary reperfusion. Am J Cardiol 1990;66: I 196- 1198. 12. Zimmermann M, Adamec R, Ciaroni S, Malbois F, T&he R. Reduction in the frequency of ventricular late potentials after acute myocardial infarction by early thrombolytic therapy. Am J Cardiol 1991;67:697-703. 13. Farrell T, Bashir Y, Poloniecki J, Ward D, Camm AJ. The effect of thrombolysis on risk stratification for arrhythmic events in post infarction patients (abstr). J Am Co// Cardiol 1991;17:17A. 14. Malik M, Odemuyiwa 0, Poloniecki J, Kulakowski P, Farrell T, Staunton A, Camm AJ. Age-related normal values of signal averaged electrocardiogram variables after acute myocardial infarction. Am J Cardiol 1991;68:440-445. 15. Greene HL, Richardson DW, Barker AH, Roden DM, Capone RJ, Echt DS, Friedman LM, Gillespie MJ, Hallstrom AP, Verter J. Classification of deaths after myocardial infarction as arrhythmic or nonarrhythmic. Am J Cardiol 1989;63: I-6. 16. Odemuyiwa 0, Malik M, Farrell T, Bashir Y, Poloniecki J, Camm AJ. A comparison of the predictive characteristics of heart rate variability index and left ventricular ejection fraction for all-cause mortality, arrhythmic events and sudden death after acute mywardial infarction. Am J Cardiol 1991;68:434-439. 17. Breithardt G, Cain ME, El-Sherif N, Flowers N, Hombach V, Janse M, Simpson MB, Steinbeck G. Standards for analysis of ventricular late potentials using high resolution or signal-averaged electrocardiography. Eur Heart J 1991;12:473-480. 18. Malik M, Farrell T, Poloniecki J, Camm AJ. Evaluation of receiver operator characteristics - optimum time of day for the assessment of heart rate variability after acute myocardial infarction. Int J Biomed Comput 1991;27:175-192. 19. Vatterott PJ, Hammill SC, Bailey KR, Wiltgen CM, Gersh BJ. Late potentials on signal-averaged electrocardiograms and patency of the infarct-related artery in survivors of acute myocardial infarction. J Am Coil Cardiol 1991;17:

330-337. 20. Leor J, Hod H, Rotstein Z, Truman S, Gansky S, Goldbourt U, Abboud S, Kaplinsky E, Eldar M. Effects of thrombolysis on the Itlead signal-averaged ECG in the early postinfarction period. Am Heart J 1990;120:495-502. 21. Worley SJ, Mark DB, Smith WM, Wolf P, Califf RM, Strauss HC, ManWaring MG, Ideker RE. Comparison of time domain and frequency domain variables from the signal-averaged electrocardiogram: a multivariable analysis. J Am Coil Cardiol 1988;11:1041-1051. 22. Morlet D, Peyrin F, Desseigne, Touboul P, Rubel P. Time-scale analysis of high-resolution signal-averaged surface ECG using wavelet transformation. In: Computers in Cardiology 1991, proceedings. IEEE Computer Society Press, in press. 23. Denniss AR, Ross DL, Richards DA, Cody DV, Russell PA, Young AA, Uther JB. Effect of antiarrhythmic therapy on delayed potentials detected by the signal-averaged electrocardiogram in patients with veniricular tachycardia after acute myocardial infarction. Am J Cardiol 1986;58:261-265.

THROMBOLYSIS AND LATE POTENTIALS AFTER AMI 25