Author + information
- Received August 2, 2001
- Revision received March 19, 2002
- Accepted April 5, 2002
- Published online July 3, 2002.
- George Hahalis, MD*,
- Christos Stathopoulos, MD*,
- Dimitrios Apostolopoulos, MD†,
- Pavlos Vasilakos, MD†,
- Dimitrios Alexopoulos, MD, FESC, FACC* and
- Antonis S Manolis, MD, FESC, FACC*,* ()
- ↵*Reprint requests and correspondence:
Dr. Antonis S. Manolis, Professor and Director of Cardiology, 41 Kourempana Street, Agios Dimitrios 173 43, Athens, Greece.
Objectives This study investigated whether ST-segment elevation and T-wave normalization (TWN) in Q-wave leads on pre-discharge exercise electrocardiogram (ECG) can contribute to patient management after a recent myocardial infarction (MI).
Background The clinical relevance of these exercise ECG changes remains controversial despite accumulating evidence of their association with myocardial viability. Because discrepancies of previous studies may depend on patient selection, the value of these ST/T abnormalities in the thrombolytic era should be better defined.
Methods One-hundred one patients, age 58 ± 11 years, with a recent, first Q-wave MI (57% thrombolyzed, ejection fraction 43 ± 7%) underwent pre-discharge, submaximal treadmill testing followed, in the absence of severe ischemia, by dobutamine stress echocardiography, thallium-201 single photon emission computed tomography, and coronary angiography.
Results ST elevation at peak exercise, but not TWN, was associated with more severe infarctions as indicated by higher peak creatine kinase (p < 0.05) and with a greater number of scarred segments both on echocardiography (p < 0.05) and scintigraphy (p < 0.01). However, the incidence of myocardial viability and ischemia did not differ between groups with or without these ST/T changes. Anterior infarction location and ≥3 echocardiographically scarred segments were among the independent predictors of ST elevation at peak ergometric exercise. During follow-up (31 ± 13 months), the rate of hard events was low (8%) and similar between the study groups.
Conclusions In patients after acute Q-wave MI without severe ischemia according to clinical and standard ECG criteria, exercise-induced ST elevation, but not TWN, is associated with larger infarctions. The contribution of these ST/T abnormalities toward identifying patients with myocardial viability or ischemia and determining risk stratification is poor. In-hospital management of such patients based on routine clinical practice is sufficient for selection of a population with a relatively low long-term risk.
Pre-discharge exercise testing (ET) is widely used in patients after acute myocardial infarction (MI) for early detection of residual myocardial ischemia and risk stratification (1). Electrocardiographic (ECG) changes that include new or further ST-segment elevation and/or T-wave normalization (TWN) of initially negative T-waves in Q-wave leads are frequently encountered during ET in patients after MI (2–21), but these findings are still controversially interpreted. In the pre-thrombolytic era, such changes were thought to represent left ventricular (LV) systolic dysfunction unrelated to ischemia (1,3,4,17,21). However, it has been recently demonstrated (8,10,11,13–15,18–20), albeit not unanimously (7,9,12), that they are indicative of myocardial viability. These discrepancies could be accounted for by patient selection, depending on the proportion of thrombolyzed patients (22)or of ischemia-driven ET (8,11,13,15,21)in previous studies (23).
Identification of viability has prognostic implications (24), and in this regard the ECG might greatly simplify the clinical management of these patients if it could offer reliable information. Thus, the aim of this study was to elucidate the role of myocardial viability and ischemia on the genesis of such exercise-induced ECG changes in patients with recent Q-wave MI. We analyzed ECGs both at low-intensity (6,10,20,25)and at peak pre-discharge ergometric exercise. In addition, the prognostic impact of these ECG changes was assessed.
Patients and study design
We prospectively screened consecutive patients presenting with first Q-wave MI from August 1995 until September 1999, verified by typical clinical, ECG, and enzymatic findings. Patients were eligible for participation if they fulfilled the following criteria: 1) stable clinical course without mechanical complications, recurrent angina, hemodynamic instability, congestive heart failure, or severe arrhythmias; 2) Q-waves of ≥0.04 s duration and ≥25% of the amplitude of the R-wave in depth in two or more contiguous leads; 3) absence of left bundle branch block, LV hypertrophy, and medications or conditions known to cause ST-segment/T-wave alterations; 4) ability to perform an ergometric ET; 5) acceptable acoustic window without contraindications to dobutamine stress echocardiography (DSE); and 6) informed consent to enter the study.
Eligible patients underwent first a pre-discharge, submaximal treadmill ET, according to the modified Bruce protocol 6 ± 2 days after admission, followed in random order and usually post discharge, by DSE, symptom-limited exercise thallium-201 single photon emission computed tomography (201Tl-SPECT) according to the standard Bruce protocol and coronary angiography. In every patient, heart rate (HR), blood pressure, 12-lead ECG, and symptoms were monitored throughout the ET. Two experienced cardiologists, blinded to the results of the undertaken studies, interpreted the examination findings by reaching a consensus. In case of disagreement a third opinion was sought.
Patients with severe ischemia on pre-discharge testing, defined as typical angina appearing in ≤6 metabolic equivalents, ST elevation in non-Q-wave leads ≥1 mm (≥2 mm for chest leads), or ST depression ≥1.5 mm lasting ≥2 min in the recovery phase were considered at high risk. These patients were catheterized before discharge, were revascularized when necessary, and were not included in the current study. In contrast, severe ST-segment elevation in Q-wave leads did not prompt an immediate invasive work-up, and these patients were permitted to undergo further investigation. End-points for ET termination included: 1) severe angina or very severe ischemia on the ECG with horizontal or downsloping ST-segment depression ≥3 mm or ST elevation as above described or 2) fulfillment of other, previously defined criteria (15,26).
Of the screened population, consisting of 220 eligible patients, 101 patients were finally included in this study and were divided into two groups: 1) patients with exercise-induced new or further ST-segment elevation (STel+) and 2) those without new or further ST-segment elevation (STel−). A separate analysis included patients with (TWN+) and those without TWN (TWN−) on pre-discharge treadmill ET. Furthermore, patients with and patients without rest ST elevation were also analyzed. We excluded 85 patients with severe ischemia and 34 patients for other reasons (Fig. 1).
Analysis of the ECG
Rest and exercise ECGs were obtained by using the computer-averaging technique. The ST segment was analyzed 80 ms after the J point and was considered as elevated or markedly elevated if it exceeded the isoelectric line by ≥1 mm or ≥2 mm, respectively. The lead exhibiting the greatest ST elevation, the sum of ST elevation in all Q-wave leads, and ST depression of ≥1 mm in non-Q-wave leads were analyzed. “Exercise TWN” was defined as negative T-wave at rest of ≥ −1 mm becoming positive by ≥ +2 mm during exercise. These ST/T changes had to occur in two or more Q-wave leads either at low-intensity treadmill stress (3 to 4 metabolic equivalents) or at peak ergometric exercise or maximal dobutamine dose as compared with the rest ECG.
Dobutamine stress echocardiography
Echocardiographic studies were performed 13 ± 5 days after acute MI as previously described (26). Analysis of the echocardiographic images, scoring of contractility of LV segments, and calculation of the wall motion score index (WMSI) were undertaken according to standard criteria (8,26). The differences of the WMSI between baseline and low dobutamine dose, between low and maximal dose, and between baseline and peak dose reflect the amount of viable tissue, the extent of ischemia, and predominantly ischemia of nondysynergic segments, respectively.
Viable segments were those hypokinetic or akinetic segments exhibiting improved contractility of ≥1 grade during DSE. Patients with viability and marked viability were those with contractile reserve in ≥2 or ≥3 contiguous segments, respectively. Homozonal and remote ischemia referred to one or more normally contracting segments at rest developing dysynergy during exercise in the infarct-related and non–infarct-related region, respectively. Furthermore, four echocardiographic responses of dysynergic segments to dobutamine were recognized (26). Intra- and inter-observer agreement values for hypokinetic and akinetic segments were 0.94 and 0.90, respectively, for rest studies, and 0.89 and 0.84 for studies at high dobutamine dose.
Thallium-201 SPECT imaging
201Tl-SPECT was performed 15 ± 6 days after acute MI with a large field-of-view rotating gamma camera equipped with an all-purpose collimator. At peak exercise 100 MBq of thallium-201 was injected intravenously. Exercise and redistribution images were acquired respectively, 10 to 15 min and 3 to 4 h later. On a separate day, an additional 60 MBq of thallium was injected and rest, and redistribution imaging was again performed (27).
Images were reconstructed and reoriented in the standard axes for interpretation and quantification of thallium uptake. Computerized two-dimensional polar maps of the three-dimensional radioactivity were generated, and a 16-segment model comparable to that of echocardiography was used. Images were normalized to the maximum count in each image set. Segments were scored semiquantitatively on a 3-point scale: 1) >75% to 100%; 2) 50% to 75%; and 3) <50% thallium uptake. By dividing the sum of the scores by 16, the total segment score index was derived. The score differences between exercise and redistribution and, in particular, between exercise and rest indicate the amount of ischemic and/or viable myocardium; the 4-h redistribution/rest difference reflects predominantly the amount of viable tissue, while the extent of scar is shown in the rest/redistribution score. Patients with viability and marked viability were those with ≥2 or ≥3 dysynergic segments, respectively, as judged by the rest echocardiographic images, exhibiting ≥50% thallium uptake. Ischemia was considered if one or more normally contracting segments exhibited a ≥25% reduction in thallium uptake.
Selective coronary angiography was performed in all patients 25 ± 7 days after MI. Severity of coronary artery stenosis was estimated visually and was considered significant if the lumen narrowing in one or more major epicardial arteries was ≥70%. The collateral circulation was semiquantitated on a 4-point scale and was present if the score was ≥2.
Follow-up data were obtained in all patients during routine follow-up visits in the outpatient cardiac clinic or from a telephone contact with the patient or the patient’s physician. Death from all causes, nonfatal MI, revascularization, and hospitalization for unstable angina were the events considered; but only the event that occurred first was taken into consideration for each patient.
Continuous data are expressed as mean value ± SD. Continuous and categorical variables were compared respectively, with the unpaired Student ttest (2-tailed) and chi-square or Fisher exact test, when appropriate. The relationship between WMSI or the number of scarred segments and the magnitude of ST elevation was examined with linear regression analysis. Agreement between modalities was assessed by kappa statistics. Multivariate logistic regression analysis was undertaken to identify independent predictors of ST elevation on rest and exercise ECG. Survival analysis and the time-dependent cumulative probability for cardiac events were calculated using Kaplan-Meier curves and differences between groups were tested with the log-rank chi-square statistics. Statistical analysis was performed with the SPSS 10.0 statistical package. Statistical significance was set at p < 0.05.
Clinical characteristics of the study patients, exercise results, and coronary angiographic findings
Most patients (97%) were in Killip class I or II, whereas almost half of them were not thrombolysed, had non-anterior MIs, a patent infarct-related artery, and multivessel disease.
Sixty-nine patients developed new (n = 20) or further (n = 49) ST elevation during peak ergometric exercise (STel+ group). These patients were compared with those without exercise ST elevation (STel− group, n = 32) (Table 1). Patients in the STel+ group had a higher percentage of anterior MIs and larger infarctions with higher creatine kinase (CK) levels and a tendency toward lower ejection fraction. Maximal HR achieved on the treadmill and the incidence of rest ST elevation were higher in the STel+ group compared with patients without ST elevation (Table 2). On DSE, eight patients developed angina, and one additional patient experienced reinfarction due to reocclusion of a severely stenotic right coronary artery.
Comparison of the two stress modalities demonstrated a lower incidence of exercise ST elevation (p < 0.001) as well as a lower peak HR (p < 0.05) and similar magnitude of ST elevation in Q-wave leads on ergometric ET, compared with DSE. The two tests were moderately concordant regarding the incidence of exercise-induced ST elevation (76%, kappa = 0.412, p = 0.001).
Echocardiographic and scintigraphic comparison of patients with and patients without ST elevation on pre-discharge exercise electrocardiogram
The STel+ group exhibited higher WMSI and thallium scores due to more extensive scar tissue, whereas the incidence of myocardial viability or ischemia did not differ between groups. In this respect, both the echocardiographic and scintigraphic findings were concordant (Table 3). The average number of segments per patient with biphasic response (Table 3), sustained improvement (0.79 ± 0.96 vs. 0.78 ± 1.01), and progressive worsening (0.32 ± 0.5 vs. 0.41 ± 0.62) on DSE, as well as the number of patients with scintigraphic-only viability were similar in the study groups. Moreover, the magnitude of exercise-induced ST elevation correlated neither with the WMSI or their differences, nor with the number of scarred segments on DSE or on 201Tl-SPECT.
Findings were similar in patients (n = 44) exhibiting ST elevation at low-level exercise. The accuracy of ST elevation during ergometric exercise in detecting the patients’ echocardiographic and scintigraphic response of dysynergic segments was low (Table 4and Fig. 2). Based on the results from the univariate analysis (Tables 1, 2, and 3), we included into the logistic model 10 variables (MI location; CK; HR at peak exercise; percent of maximal age-predicted HR on ET; rest ST elevation; WMSI at high dobutamine dose; ≥3 scarred segments on echocardiography; exercise thallium score; redistribution thallium score; and rest/redistribution thallium score). Table 5shows the independent favorable and unfavorable predictors of exercise-induced ST elevation on peak-exercise ECG according to the results of this multivariate analysis.
Comparison of patients with and patients without rest ST-segment elevation
Patients with rest ST elevation (n = 62) demonstrated, in comparison with those without rest ST elevation (n = 39), a higher percentage of anterior MIs (74% vs. 28%) and larger infarctions with higher CK levels, lower ejection fraction, higher WMSI, greater rest/redistribution score, and higher average number of scarred segments, both on DSE and on 201Tl-SPECT (all p < 0.01). On DSE, the presence of three or more scarred segments emerged as an independent predictor of rest ST elevation on multivariate analysis (p = 0.003). However, when the MI location was entered into the model, only anterior MI was independently associated with rest ST elevation (p = 0.009).
Comparison of patients with and patients without TWN
Patients exhibiting TWN at low-level (n = 25) or at peak exercise (n = 47) were not different from the patients without TWN in any variables (data not shown). The accuracy of TWN in detecting viability was low (Table 4).
During follow-up (mean 31 ± 13 months, range 10–63 months) a total of 43 events occurred: seven deaths, one nonfatal MI (hard events, n = 8), 26 revascularizations (only one in the first three months after MI), and nine episodes of unstable angina. Figure 3shows a life-table analysis of survival and event-free survival of the patient groups. The groups STel+/STel− and TWN+/TWN− were similar with regard to hard events (8.7% vs. 6.2% and 8.3% vs. 7.5%, respectively) and total events (41% vs. 47% and 37.5% vs. 47%, respectively) (p = NS).
In a relatively low-risk patient population with recent Q-wave MI, we found that exercise-induced ST elevation, but not TWN, was associated, among other factors, with larger MIs (Table 5). There was an independent contribution of MI location, which may be explained by the fact that anterior MIs are usually larger and in close proximity to the chest leads. Moreover, these ST/T changes were neither associated with higher incidence of myocardial viability or ischemia nor could they risk-stratify these patients.
Mechanisms of ST elevation and Twn in Q-wave leads
These ECG abnormalities may develop spontaneously (e.g., during reinfarction) or be provoked by ET. Exercise ST/T changes are thought to reflect either residual viability (13–15,18,19)or passive bulging of the scarred tissue (1–5)resulting in segmental dysynergy. However, the reported accuracy of ST/T changes at low-level stress in detecting viability (6,10,20)should indicate that improved regional contractility might rather underlie these ECG abnormalities. Experimentally, ST-elevation potential can be transmitted from the peri-infarction zone through the scarred region (28). In a clinical study, the time-dependency of residual myocardial viability could be shown (29). Furthermore, the temporal evolution of T-wave changes may reflect the underlying regional mechanical function of the myocardium (30). Thus, electrical, mechanical, and unknown factors or nonspecific ECG response, alone or in variable combination, may be responsible for such ST/T changes, and their relative importance may differ according to the clinical circumstances.
Some studies demonstrated an association between exercise-induced ST elevation and myocardial viability after acute MI (8,10,13–15), whereas other reports concurred with our results (7,9). Exercise-induced ST- elevation was found to predict recovery of LV function with a 67% to 82% accuracy (8,13,15). The revascularization rate ranged from 33% to 100% (8,13,15), mostly early post MI (8,15).
Furthermore, TWN was shown to be associated with recovery of myocardial function (6,11), viability (10), or the extent of wall motion abnormalities (12). With regard to prognosis, any ST shift (2), ST-elevation alone (3), or TWN (16)predicted cardiac events. Our results are concordant with those of Haines et al. (4), who showed no prognostic difference among non-thrombolyzed patients with and non-thrombolyzed patients without exercise ST elevation.
Potential explanation of the discordant results of previous studies in light of the current findings
Exercise ST/T response in Q-wave leads after acute MI may be influenced by a number of factors such as extent of exercise (4), chronicity of MI (29), dobutamine use (5–9,11–13,16), and most importantly, patient selection. Thus, the post-test probability that exercise-induced ST elevation/TWN will detect viability/ischemia depends on the prevalence of these characteristics (23). If thrombolyzed patients with extensive amounts of stunned myocardium undergo ET early after MI, and before the clinical picture has allowed the clinician to sort out high-risk patients, then the likelihood that these ST/T changes will detect viable or ischemic tissue should be high. Similarly, DSE studies may have included high-risk patients in low-dose-only DSE protocols (6)or enrolled patients unable to exercise. On the other hand, in stable patients with considerable myocardial damage and the attendantly high probability for severe contractile dysfunction without severe exercise-induced ischemia, exercise-provoked ST elevation/TWN should be more likely to indicate mechanical stretch of the infarcted myocardium (i.e., “aneurysmatic response”).
Our results are consistent with this interpretation. In the current study, rate of thrombolysis (5,7,10–13,15); ejection fraction (3–5,7,10–13); and incidence of exercise-induced angina (11,15), patent infarct-related artery (4,5,7,10–15), and revascularization (8,11,13,15)were among the lowest reported. In addition, the incidence of rest ST elevation, which is a recognized marker of extensive scar tissue (4,5,7–9,11,13), and WMSI, which is consistently greater in the STel+ groups (7–9,13,15,16), were in the highest reported range in this and other reports with findings similar to ours (5,7,9).
Aggressive management of high-risk patients who were filtered out of our study group accounts for the main differences in patient selection between our study and previous reports (Fig. 1). We believe that our study population probably represents the real world of usual clinical practice and patient selection after Q-wave MI.
Our study is limited by the lack of follow-up indices of LV myocardial status. Hereof, ECG prediction of functional recovery in patients with these ST/T abnormalities was associated with higher frequency of viability at baseline examination (6,8,13,15). Equally distributed incidence of viability in our study groups suggests that improvement of late myocardial function should be very unlikely in patients with exercise ST elevation or TWN.
Conclusions and clinical implications
In patients with recent MI but without severe ischemia, exercise-induced ST elevation in Q-wave leads are associated with larger MIs. In this relatively low-risk population, these exercise ST/T abnormalities are of little value in detecting myocardial viability or ischemia and in risk-stratifying such patients.
The authors acknowledge Dr. Ignatios Ikonomidis, who performed some of the echocardiographic studies and rendered advice on statistical analysis.
- creatine kinase
- dobutamine stress echocardiography
- exercise test/testing
- heart rate
- left ventricular
- myocardial infarction
- thallium-201 single photon emission computed tomography
- T-wave normalization
- wall motion score index
- Received August 2, 2001.
- Revision received March 19, 2002.
- Accepted April 5, 2002.
- American College of Cardiology Foundation
- Corbett J.R.,
- Dehmer G.J.,
- Lewis S.E.,
- et al.
- Sullivan J.D.,
- Davies D.W.,
- Sowton E.
- Haines E.D.,
- Beller G.A.,
- Watson D.D.,
- Kaiser D.L.,
- Sayre S.L.,
- Gibson R.S.
- Lanzarini L.,
- Fetiveau R.,
- Poli A.,
- Cavalott C.,
- Griffini M.,
- Previtali M.
- Ehlendy A.,
- Cornel J.H.,
- Roelandt J.R.T.C.,
- et al.
- Mobilia G.,
- Zanco P.,
- Desideri A.,
- et al.
- Pizzetti G.,
- Montorfano M.,
- Belotti G.,
- Margonato A.,
- Ballarotto C.,
- Chierchia S.L.
- Yamagishi H.,
- Akioka K.,
- Takagi M.,
- et al.
- Mezilis N.E.,
- Parthenakis F.I.,
- Kanakaraki M.K.,
- Marketou M.E.,
- Mayrakis H.E.,
- Vardas P.E.
- Bruce R.A.,
- Fisher L.D.,
- Pettinger M.,
- Weiner D.A.,
- Chaitman B.R.
- Margonato A.,
- Ballarotto C.,
- Bonetti F.,
- et al.
- Margonato A.,
- Chierchia S.L.,
- Xuereb R.G.,
- et al.
- Lombardo A.,
- Loperfido F.,
- Pennestri F.,
- et al.
- Bolli R.
- Hoffer E.P.,
- Dewe W.,
- Celentano C.,
- Pierard L.A.
- Smart S.,
- Wynsen J.,
- Sagar K.
- Ragosta M.,
- Beller G.A.,
- Watson D.D.,
- Kaul S.,
- Gimple L.W.
- Cinca J.,
- Bardaji A.,
- Carreno A.,
- et al.
- Bosimini E.,
- Giannuzzi P.,
- Temporelli P.,
- et al.