Prognostic Value of Exercise 201Tl Tomography in Patients Treated With Thrombolytic Therapy During Acute Myocardial Infarction
Background Although myocardial perfusion scintigraphy is of proven value in the risk stratification of patients with a recent myocardial infarction who receive conventional therapy, its value in patients undergoing thrombolytic therapy remains controversial.
Methods and Results Seventy-one patients who received thrombolytic therapy for acute myocardial infarction had exercise 201Tl tomography and coronary angiography before hospital discharge. Eleven (15%) of 71 patients had ischemic ST-segment depression during exercise, whereas 27 patients (38%) had scintigraphic ischemia. Twenty-five (37%) of 68 patients had a cardiac event consisting of either death (n=2), recurrent myocardial infarction (n=5), congestive heart failure (n=7), or unstable angina (n=11) during a follow-up of 26±18 months. Univariate predictors of cardiac events were as follows: Killip class (P=.04); left ventricular ejection fraction (P<.0005); total (P=.002) and ischemic (P<.0005) perfusion defect size; percent thallium lung uptake (P=.001); presence of infarct-zone redistribution (P=.02); and multivessel coronary artery disease (P=.01). By multivariate analysis, the significant joint predictors of risk were ejection fraction (P<.0005) and ischemic perfusion defect size (P=.005). The combination of ejection fraction and thallium tomography added significant incremental prognostic information to the clinical data, whereas angiography did not further improve a model that included clinical, ejection fraction, and tomographic variables.
Conclusions Quantitative exercise 201Tl tomography provides important incremental, long-term prognostic information in patients receiving thrombolytic therapy for acute myocardial infarction.
Studies from the prethrombolytic era1 2 3 4 have clearly established the important role of exercise 201Tl scintigraphy for predicting the extent and location of coronary artery stenoses after acute myocardial infarction and in stratifying patients for future cardiac events by risk. Thrombolytic therapy, however, has altered the clinical profile of patients after infarction, and particularly their propensity for developing subsequent cardiac events.5 6 7 8 9 Thus, it may not be appropriate to extrapolate into the current era the value of exercise 201Tl single photon emission computed tomography (SPECT) for stratifying risk. Accordingly, the present study was designed to investigate the utility of exercise 201Tl SPECT after thrombolytic therapy for (1) defining the extent and location of coronary artery stenoses, (2) detecting the presence of residual ischemia in the infarct and noninfarct zones, and (3) predicting risk on the basis of quantitatively derived SPECT variables.
Between January 1988 and December 1991, 425 patients received thrombolytic therapy for acute myocardial infarction in our institution; of these, 71 (66 men; mean age, 55±10 years) underwent predischarge exercise thallium SPECT, coronary angiography, and ventricular function assessment. Myocardial infarction was diagnosed if patients had prolonged chest pain (≥30 minutes), typical ECG changes of infarction, and a characteristic rise and fall in plasma creatine kinase–MB activity. ECG criteria used to diagnose acute Q-wave and non–Q-wave infarction were as previously reported.10 None of our patients had a left bundle branch block or other nonspecific intraventricular conduction defect in the baseline ECG. One patient had a right bundle branch block.
Patients were excluded from the study if they had (1) valvular, congenital, or cardiomyopathic heart disease; (2) cardiogenic shock; or (3) a life-threatening disease that might limit long-term follow-up.
Patients had treadmill exercise (Bruce protocol11 ) 201Tl SPECT at a median of 13 days after the index myocardial infarction. Twelve-lead ECGs were recorded each minute during exercise, at the onset of recovery, and thereafter for 5 minutes. Three leads were monitored continuously throughout recovery until the ECG returned to baseline. Criteria for premature termination of exercise were (1) the development of severe angina pectoris, dyspnea, or fatigue; (2) frequent (more than 10/min) multifocal or paired ventricular extrasystoles; (3) >3 mm ST-segment depression; or (4) a ≥10 mm Hg decrease in systolic blood pressure compared with the previous stage. Ischemic ST-segment changes were defined as ≥1 mm of horizontal or downsloping ST depression, measured 80 ms after the J-point in three consecutive beats during exercise, recovery, or both.
Quantitative 201Tl Tomography
201Tl (3.0 mCi) was administered intravenously and flushed with 10 mL of normal saline as the patient achieved target heart rate or developed limiting symptoms. SPECT was performed as previously reported from our laboratory.10 12 Imaging commenced 10 minutes after completion of exercise and was repeated 4 hours later.
The stress and redistribution SPECT images were quantified and compared with an exercise thallium normal data bank with standard computer-generated polar maps used to define the total left ventricular perfusion defect size and the extent of scintigraphic scar and ischemia.10 12 The patient's initial polar map was considered abnormal when a >3% focal perfusion defect existed within a specific vascular territory.12 Lung uptake was measured by placing a 5×5-pixel region of interest over the left lung and comparing the thallium activity in this region with that in a similar region of interest over the heart segment with the highest tracer activity (lung/heart ratio). The anterior image acquired as part of the tomographic study was used for this evaluation.
Coronary Angiography and Ventricular Function
Before hospital discharge, all patients underwent selective coronary angiography by use of standard techniques. The decision to perform coronary angiography rested with the patient's attending cardiologist. Because coronary angiography is performed in the majority of patients during admission for the index myocardial infarction whereas perfusion imaging is ordinarily reserved for a much smaller fraction of patients, we deliberately selected patients who had both tests so that we could compare the prognostic value of these two diagnostic techniques in the same patients. The coronary arteriograms were reviewed by an experienced angiographer who was blinded to the scintigraphic results. Coronary stenoses were measured with electronic calipers and expressed as percent luminal diameter stenosis. Stenosis severity was graded as insignificant (≤50%), moderate (51% to 69%), severe (70% to 99%), or total obstruction (100%). Significant stenoses of diagonal and marginal branches were assigned to the left anterior descending and circumflex coronary arteries, respectively. The infarct-related artery was inferred from the coronary angiogram and the ECG location of infarction. In patients who had coronary angioplasty of infarct- or non–infarct-related arteries before imaging, the residual percent stenosis after angioplasty was used to determine the sensitivity of SPECT for detecting significant coronary stenosis. The 7 patients who had coronary artery bypass surgery were excluded from the sensitivity analysis. The left ventricular ejection fraction (LVEF) was calculated by two-dimensional echocardiography in 55 patients and by contrast ventriculography in the remaining 16 patients.
Follow-up after hospital discharge was accomplished through phone interviews with patients and their cardiologists, independently of any scintigraphic or angiographic data. Medical records were carefully reviewed and all events confirmed by one of the investigators. Cardiac events were defined as (1) cardiac death, (2) nonfatal reinfarction, (3) unstable angina pectoris (ie, rest or worsening exertional pain requiring hospitalization and subsequent revascularization), and (4) signs and symptoms of severe congestive heart failure (ie, pulmonary edema) necessitating hospital admission. Revascularization per se was not considered a cardiac event.
Sensitivity of SPECT for detecting coronary stenosis was defined as the number of true-positive scans×100 divided by the true-positive plus the false-negative scans. Specificity was defined as the number of true-negative scans×100 divided by the true-negative plus the false-positive results.
The risk of cardiac events associated with possible prognostic predictors was investigated by use of Cox regression for censored data. These results are reported as relative risks (RRs) (hazard rate ratios) that incorporate the time to event or to censoring. For continuous predictors, the RR is the multiplicative increase in risk of a cardiac event associated with a unit change in the predictor. Actual probability values are reported for all tests; however, the interpretation of statistical significance was based on keeping the family-wise error rate ≤.05 for each group of related variables. These groups consisted of clinical, treadmill, angiographic, LVEF, and SPECT variables.
The clinical variables included patient's age, history of prior infarction, initial Killip classification on hospital admission, infarct location, infarct type (Q versus non-Q), and history of prior coronary revascularization. The treadmill exercise variables that we examined were the total exercise time, the rate-pressure product at maximal exercise, and whether the patient developed chest pain or ischemic ST-segment depression. The angiographic variables assessed were the extent of coronary artery disease and the patency of the infarct-related artery. The scintigraphic variables included in this analysis were the total and ischemic perfusion defect sizes and the percent lung uptake. All of these variables were obtained independently by separate investigators so as not to introduce a potential study bias. CIs were also adjusted for the multiple comparisons.13 Incremental improvements in the log-likelihood statistic were estimated as variables were sequentially added to a Cox regression model. Likelihood ratio tests indicated the significance of the improvement in the global χ2 statistic as each group of variables was added. Exploratory analyses evaluated cutpoints on several patient characteristics that would stratify patients into high- and low-risk groups for cardiac events. Log-rank tests and Kaplan-Meier product-limit estimates of survival curves were used in these analyses.14 Exploratory analyses were interpreted as hypothesis generating, not hypothesis testing.
All statistical tests were two-tailed. Stata software was used for the statistical analyses.15 Continuous data are reported as mean±SD.
The 71 study patients were predominantly men (66 men), 7 (10%) of whom had prior myocardial infarction and 5 (7%) of whom had previous coronary artery bypass surgery. Forty-nine percent had anterior infarction, and most patients (85%) developed pathological Q waves. At the time of hospital admission, 82% of patients were in Killip class I. The mean LVEF, determined a median of 2 days after infarction, was 50±13%. (See Table 1⇓.)
All patients received thrombolytic therapy according to standard protocols.5 16 The thrombolytic agent used was tissue plasminogen activator in 66 patients (93%) and streptokinase in the remaining 5 (7%). Thrombolytic therapy was begun 3.5±2.5 hours after the onset of chest pain.
All patients had coronary angiography early (median, 4 days) after infarction. Forty-five percent of patients had angioplasty of the infarct-related artery (n=28), both the infarct- and non–infarct-related arteries (n=2), or coronary artery bypass graft surgery (n=2) before scintigraphy. Coronary angioplasty and bypass surgery were performed a median of 5 and 7 days after infarction, respectively. In the 30 patients who had angioplasty of the infarct-related artery, 22 were successful, 3 were unsuccessful, and 5 were suboptimal with a residual coronary stenosis >50%. In the 2 patients who had bypass surgery, complete revascularization was performed as technically feasible. Exercise SPECT was performed a median of 13 days after infarction. Most patients (82%) performed the exercise test while taking antianginal medications: β-blockers (17%), calcium-channel antagonists (56%), or nitrates (45%).
As shown in Table 2⇓, on initial coronary angiography, 34 patients (48%) had one-vessel, 28 (39%) had two-vessel, and 7 (10%) had three-vessel disease. Two patients (3%) had no significant coronary stenoses. The infarct-related artery was occluded or severely stenosed in 19 (27%) and 36 (51%) of the patients, respectively. The extent and severity of residual coronary artery stenosis after revascularization are also shown in Table 2⇓.
Treadmill Exercise Results
The mean duration of treadmill exercise was 7.3±2.5 minutes. The mean peak heart rate and systolic blood pressure were 132±18 bpm and 149±18 mm Hg, respectively. Most patients (68%) performed a submaximal exercise test (ie, <85% predicted heart rate), and only 11 patients (15%) had exercise-induced ischemic ST-segment depression. Seven patients (10%) experienced angina during exercise testing.
Exercise SPECT Results
Most patients (92%) had at least one perfusion defect, which was fixed (54%) or showed either partial (35%) or complete (3%) reversibility. The 6 patients (8%) with normal scans had significantly smaller creatine kinase–MB levels than those with abnormal scans (52±50 versus 153±109 IU, respectively; P<.01). Of the 27 patients who had scintigraphic ischemia, 19 (70%) had ischemia solely localized to the infarct zone and 8 (30%) had ischemia in both the infarct and noninfarct zones.
Sensitivity and Specificity of 201Tl Tomography for Detecting Coronary Stenoses
In the 64 patients without prior bypass surgery, the infarct-related artery had minimal stenosis in 25 patients (22 after successful angioplasty) but significant (>50%) stenosis in the remaining 39 patients. The number of non–infarct-related arteries with significant stenosis was 35. The overall sensitivity was significantly higher for detecting infarct- (95%) compared with non–infarct-related (49%) coronary stenoses (P<.0005). A low percentage of non–infarct-related arteries with moderate (48%) or severe (46%) stenoses were detected by exercise SPECT. The specificity of exercise SPECT was 97% for the 93 non–infarct-related arteries with insignificant stenosis.
Of the 71 patients initially enrolled in the study, 68 were followed up for 26±18 months (range, 0.5 to 36 months). Three patients were lost to follow-up. Twenty-five (37%) of the 68 patients either died (n=2), had recurrent myocardial infarction (n=5), or were rehospitalized for unstable angina (n=11) or congestive heart failure (n=7).
Univariate Predictors of Events
As shown in Table 3⇓, the only clinical predictor of cardiac events was the initial Killip class (RR=2.42; 95% CI=0.80, 7.34; P=.04), although patients with anterior infarction tended to have a higher event rate (RR=2.23; 95% CI=0.74, 6.75; P=.06). None of the exercise variables contributed toward predicting risk. The estimated risk of a cardiac event was doubled for every 10% decrease in the LVEF (RR=2.03; 95% CI=1.49, 2.76; P<.0005). The LVEF was also a significant predictor of unstable angina and myocardial infarction (RR=1.66 for every 10% decrease in LVEF; 95% CI=1.15, 2.39; P=.006). The RR was 1.36 times higher for every 10% increase in the SPECT perfusion defect size (RR=1.36; 95% CI=1.06, 1.76; P=.002). Furthermore, the ischemic perfusion defect size, the presence of infarct-zone redistribution, and increased thallium lung uptake were also significant scintigraphic predictors of events. Patients with multivessel disease at the time of initial coronary angiography had a higher event rate than those with single-vessel involvement.
Kaplan-Meier curves depicting event-free survival were also generated for specific dichotomized variables. The LVEF (Fig 1A⇓) and quantified SPECT perfusion defect size (Fig 1B⇓) both effectively predicted high- and low-risk groups. The presence of scintigraphic ischemia was also a strong predictor of events (Fig 1C⇓), as was the presence of multivessel coronary artery disease (Fig 1D⇓). The absolute extent of left ventricle ischemia also defined risk when dichotomized at 10%. Fifty percent of patients with a >10% ischemic defect had a subsequent cardiac event versus 26% of patients with a defect ≤10% (P=.03). An example of SPECT of a patient who had recurrent infarction after hospital discharge is shown in Fig 2⇓.
The variables listed in Table 3⇑ were reanalyzed for predicting cardiac death and nonfatal reinfarction (n=7 patients). By univariate analysis, the only significant predictor of this combined end point was the LVEF; estimated risk doubled for every 10% decrease in LVEF (RR=2.06; 95% CI=1.17, 3.64; P=.01). The LVEF also predicted infarct-free survival by log-rank analysis (Fig 3A⇓). The estimated risk increased by 1.27-fold for every 10% increase in the SPECT perfusion defect size (RR=1.27; 95% CI=0.82, 1.99; P=.16) but did not reach statistical significance. However, the 20% cutpoint for defect size was useful for stratifying patients into high- and low-risk groups. Infarct-free survival was 97% in the 35 patients with a <20% perfusion defect size compared with 82% in the 33 patients with larger (≥20%) perfusion defects (P=.05) (Fig 3B⇓).
Multivariate Predictors of Cardiac Events
Stepwise Cox regression analysis was used to estimate the relative importance and overlap of the risk factors listed in Table 3⇑. The joint risk factors for cardiac events were low ejection fraction (RR=1.85 for every 10% decrease; 95% CI=1.34, 2.56; P<.0005) and ischemic defect size (RR=1.38 for every 5% increase; 95% CI=1.10, 1.74; P=.005). The positive predictive value of an ejection fraction <40% and scintigraphic ischemia for a subsequent cardiac event was 83%, with a negative predictive value of 73%. There were 2 instances of death or myocardial infarction (17%) among the 12 patients in this category. In comparison, among the 37 patients with an LVEF ≥40% and no scintigraphic ischemia, there were only 2 instances of death or myocardial infarction (5%). The LVEF was the only significant predictor of death or nonfatal reinfarction by multivariate analysis (RR=2.06 for every 10% decrease; 95% CI=1.17, 3.64; P=.01).
Incremental Prognostic Value of Ejection Fraction and Scintigraphic Variables
The incremental value of the ejection fraction and scintigraphic and angiographic variables for predicting risk above and beyond the clinical data is displayed in Fig 4⇓. As shown, the LVEF and SPECT variables significantly contributed toward predicting risk beyond the clinical data alone, whereas angiography did not. Incorporating the treadmill results did not significantly improve the predictive power of the clinical data (P=.19).
The optimal strategy for stratifying risk in patients after myocardial infarction has long been the subject of extensive research. Although ischemia detected by 201Tl scintigraphy is known to be a strong predictor of future events from studies published in the prethrombolytic era,1 2 3 4 17 18 19 its prognostic value in patients who receive thrombolytic therapy is controversial. It is clear that patients selected for thrombolytic therapy are generally less sick than those excluded from this therapy in that they tend to be younger, have a lower incidence of multivessel disease, and have a better LVEF. All of these facts contribute to a low cardiac event rate and better overall survival. This was also true in our patient cohort, in which only 10% of the patients had triple-vessel disease and the mean LVEF was 50%. Although few (10%) of our patients died or had recurrent myocardial infarction, exercise-induced scintigraphic ischemia, especially when combined with a low LVEF, remained an important predictor of both fatal and nonfatal cardiac events in patients receiving thrombolytic therapy for acute myocardial infarction.
Detection of Ischemia After Infarction
Studies from the prethrombolytic era2 3 4 17 18 19 report that ≈60% of patients have evidence of scintigraphic ischemia after infarction. Our data emphasize that a sizable percentage of patients (38%) still had detectable ischemia on exercise SPECT after thrombolytic therapy. In fact, this percentage is similar to the 43% reported in the literature20 21 22 and the 42% incidence we recently observed with adenosine SPECT in patients receiving thrombolytic therapy.23
The present study further demonstrates that exercise SPECT detects ischemia primarily within the infarct zone, supporting the concept that thrombolytic therapy limits infarction size but increases the potential for peri-infarct ischemia. The substantially lower incidence of exercise-induced ECG ischemia in patients evaluated in the thrombolytic (13%)21 22 23 24 25 compared with the prethrombolytic (31%)2 3 17 19 era may be explained by the inability of the ECG to accurately identify ischemia within the infarct zone and the current lower prevalence of multivessel coronary artery disease in patients presenting with acute infarction. ECG ischemia was observed in only 15% of patients in the present study. Moreover, although patients with subsequent events were slightly more likely to have lower exercise tolerance, exercise-induced angina, and ischemic ST-segment depression than patients without events, none of those differences was statistically significant (Table 3⇑), probably because of the small study cohort. However, because many patients continue to have scintigraphic ischemia after thrombolytic therapy for infarction, scintigraphic ischemia should remain an important substrate for the development of subsequent cardiac events.
Scintigraphic Predictors of Risk
In the present study, the total left ventricular perfusion defect size, the presence of scintigraphic ischemia, and increased thallium lung uptake were all important univariate predictors of risk. These findings are consistent with recent data emphasizing that the total stress-induced perfusion defect size predicts overall risk in a wide spectrum of patients with coronary artery disease.23 26 The Western Washington Streptokinase Trial27 demonstrated that a >20% left ventricular perfusion defect size with rest SPECT imaging predicted death. The perfusion defect size dichotomized at 20% also accurately defined overall risk both in the current study as well as in a recent report from our laboratory using adenosine 201Tl SPECT.23
Our data reaffirm that scintigraphic ischemia effectively predicts risk even in patients who have acute coronary reperfusion during infarction. Tilkemeier et al21 first studied the role of submaximal planar thallium scintigraphy in 171 patients who received either conventional care or reperfusion therapy during acute infarction. Although the positive predictive value of exercise-induced ischemia for predicting cardiac events was low (35%), it was similar in patients who did (36%) or did not (33%) receive thrombolytic agents. Furthermore, the presence of scintigraphic ischemia identified four (80%) of five patients who received intervention who later died or had recurrent infarction. In the present study, 58% of patients with scintigraphic ischemia detected by quantitative SPECT analysis had subsequent cardiac events. Similar results were reported by Brown et al20 using dipyridamole scintigraphy in a heterogeneous patient population of whom 50% received thrombolytic therapy during acute infarction. In that study, 55% of patients with infarct-zone ischemia had either recurrent chest pain or reinfarction before hospital discharge (40%) or a subsequent late cardiac event (15%). None of the 30 patients without infarct-zone ischemia had an event. Owing to the small sample size, a separate analysis of patients who did or did not receive thrombolytic therapy was not feasible. However, we recently reported23 that the quantified extent of scintigraphic ischemia detected with adenosine SPECT predicted future cardiac events, irrespective of whether patients received acute reperfusion therapy. More than 50% of patients with a >10% ischemic left ventricular perfusion defect had a cardiac event by 1 year compared with only 10% of patients with a ≤10% defect. Importantly, none of the latter patients died. These data and the results of the current study solidify an evolving concept that the resultant extent of residual myocardial ischemia, rather than the initial therapeutic strategy per se, best defines future cardiac risk.
Our results differ from those recently reported by Miller et al28 in a study assessing the value of exercise thallium scintigraphy after thrombolytic therapy. These authors found a limited value for this technique in predicting future events. However, most (68%) of their patients were taking β-blocking drugs, which presumably led to a very low peak heart rate during exercise (109 bpm compared with 132 bpm in our study), and most of the events consisted of late revascularization, with only one death occurring in the 3-year follow-up. In that study, 64% of the patients underwent coronary revascularization either before (38%) or after (26%) thallium scintigraphy. Another important difference from our study is that whereas we used computer quantification of defect sizes, Miller et al28 used a qualitative score to grade their scans.
Prognostic Value of LVEF
In the present study, the LVEF was also effective at stratifying cardiac risk. Our data are in agreement with previous studies reported in the prethrombolytic era.29 In the Western Washington Streptokinase Trial,27 the 20% of patients with an LVEF of ≤35% had a 3-year mortality of 22%. Similarly, Simoons et al30 studied 422 patients randomized to either streptokinase or placebo. In patients with an LVEF ≥40%, the mortality rate remained low (<1% per year) but increased to 25% at 3 years in those with an LVEF <40%. Survival was similar at various LVEF cutoff points in patients receiving streptokinase versus conventional management. Thus, the extent of residual left ventricular dysfunction remains an important predictor of long-term survival, irrespective of initial therapy.
Incremental Value of Ejection Fraction and Scintigraphic Variables
As shown in Fig 4⇑, both the LVEF and the SPECT variables had significant incremental value over the clinical findings alone. However, the addition of the angiographic variables did not significantly improve the predictive power of the model. Thus, patients can be optimally stratified with the use of a noninvasive strategy consisting of LVEF and SPECT imaging. Patients with an LVEF ≥40% and no scintigraphic ischemia have low risk for future events and could be managed conservatively. Conversely, patients with an LVEF <40% and scintigraphic ischemia should be preferentially targeted for invasive evaluation and possible revascularization.
This was a retrospective study that provides observational information but requires prospective validation. As is often the case with studies conducted in the thrombolytic era, our patients tended to have a low occurrence of “hard” events, namely, death and myocardial infarction. Moreover, the high utilization rate of coronary angiography and revascularization clearly modifies the natural history of these patients in that those with more severe stenoses or more abnormal exercise testing results undergo revascularization, thereby presumably reducing the frequency of subsequent hard cardiac events. Our results, nonetheless, are consistent with previous studies evaluating the utility of scintigraphy in patients after infarction. The frequency of early revascularization in our study (45%) is similar to that observed in current clinical practice31 and in other recently published trials evaluating acute therapy for myocardial infarction.24 32 The high frequency of early myocardial revascularization renders it far more difficult for any test to predict future events, because successful revascularization may decrease the occurrence of subsequent events. The inclusion of such patients, however, would appear to be clinically relevant, because this is the standard of practice in many tertiary referral institutions. Finally, because our patient sample was small, we combined several end points for analysis. Yet, the baseline characteristics of our study group were very similar to those observed in patients enrolled in the large TIMI II trial.16 24 Importantly, cardiac events still occurred in a substantial fraction of our patients. Death, myocardial infarction, or unstable angina during follow-up occurred in 26% of our patients compared with 28% in the TIMI II trial. Thus, despite our small sample size, the patients we enrolled are representative of those currently treated with thrombolytic therapy. A further limitation is that only five women were included in our cohort. This is fairly typical of the postinfarct population, but nonetheless, caution must be exercised in generalizing our findings to female patients.
Exercise 201Tl tomography continues to provide important prognostic information in patients receiving thrombolytic therapy for acute infarction, particularly when combined with LVEF. The noninvasive detection of high-risk patients after thrombolytic therapy should better identify those patients who might benefit most from aggressive medical therapy of ischemia and/or coronary revascularization. In this era of heightened awareness of cost-benefit ratios regarding testing and therapeutic options, one may question whether the routine use of myocardial perfusion imaging in patients with a myocardial infarction treated with (or without) thrombolytic therapy would be cost-effective. Although we did not perform a cost analysis in the present study, the TIMI-II study clearly demonstrated that a conservative approach (ie, performing coronary angiography only in patients with spontaneous or exercise-induced ischemia) was far more cost-effective than generalizing coronary angiography to everyone. Because perfusion imaging is a more sensitive and specific noninvasive modality for detecting ischemia than the exercise ECG, one would anticipate even greater cost-effectiveness with perfusion scintigraphy. This, however, remains speculative and warrants further investigation.
Whether exercise or pharmacological stress should be preferred is not clear at the present time. Our group has recently demonstrated the high prognostic accuracy of adenosine thallium tomography and the LVEF in patients with a recent myocardial infarction. Thus, at the present time, both exercise and pharmacological stress are good options for these patients. Pharmacological stress may be a more attractive alternative early after the acute event, but exercise testing provides additional important information regarding exercise tolerance and presence of cardiac arrhythmias. Although the choice of stressor agent should be individualized, exercise scintigraphy may still be preferable in many patients, with pharmacological scintigraphy reserved for those who are unable to exercise or when very early (2 to 5 days) risk stratification is sought.
- Received February 7, 1996.
- Revision received June 17, 1996.
- Accepted July 8, 1996.
- Copyright © 1996 by American Heart Association
Gibson RS, Watson DD, Craddock GB, Crampton RS, Kaiser DL, Denny MJ, Beller GA. Prediction of cardiac events after uncomplicated myocardial infarction: a prospective study comparing predischarge exercise thallium-201 scintigraphy and coronary angiography. Circulation. 1983;68:321-336.
Gibson RS, Beller GA, Gheorghiade M, Nygaard TW, Watson DD, Huey BL, Sayre SL, Kaiser DL. The prevalence and clinical significance of residual myocardial ischemia 2 weeks after uncomplicated non-Q wave infarction: a prospective natural history study. Circulation. 1986;73:1186-1198.
Yusuf S, Collins R, Peto R, Furberg C, Stampfer MJ, Goldhaber SZ, Hennekens CH. Intravenous and intracoronary fibrinolytic therapy in acute myocardial infarction: overview of results on mortality, reinfarction and side effects from 33 randomized controlled trials. Eur Heart J. 1985;6:556-585.
ISIS 2 Collaborative Group. Randomized trial of intravenous streptokinase, oral aspirin, both or neither among 17,187 cases of suspected acute myocardial infarction. Lancet. 1987;2:349-360.
Mahmarian JJ, Pratt CM, Borges-Neto S, Cashion WR, Roberts R, Verani MS. Quantification of infarct size by thallium-201 single-photon emission computed tomography during acute myocardial infarction in man: comparison with enzymatic estimates. Circulation. 1988;78:831-839.
Lee ET. Statistical Methods for Survival Data Analysis. 2nd ed. Belmont, Calif: Lifetime Learning Publications; 1992.
Stata Corporation. Stata Reference Manual: Release 3.1. 6th ed. College Station, Tex: Stata Corp; 1993.
Gimple LW, Hutter AM, Guiney TE, Boucher CA. Prognostic utility of predischarge dipyridamole-thallium imaging compared to predischarge submaximal exercise electrocardiography and maximal exercise thallium imaging after uncomplicated acute myocardial infarction. Am J Cardiol. 1989;64:1243-1248.
Wilson WW, Gibson RS, Nygaard TW, Craddock GB Jr, Watson DD, Crampton RS, Beller GA. Acute myocardial infarction associated with single vessel coronary artery disease: an analysis of clinical outcome and the prognostic importance of vessel patency and residual ischemic myocardium. J Am Coll Cardiol. 1988;11:223-234.
Mahmarian JJ, Mahmarian AC, Marks GF, Pratt CM, Verani MS. Role of adenosine thallium-201 tomography for defining long-term risk in patients after acute myocardial infarction. J Am Coll Cardiol. 1995;25:133-140.
Williams DO, Braunwald E, Knatterud G, Babb J, Bresnahan J, Greenberg MA, Raizner A, Wasserman A, Robertson T, Ross R. One-year results of the Thrombolysis In Myocardial Infarction investigation (TIMI) phase II trial. Circulation. 1992;85:533-542.
Simoons ML, Vos J, Tijssen JGP, Vermeer F, Verheugt FWA, Krauss XH, Cats VM. Long-term benefit of early thrombolytic therapy in patients with acute myocardial infarction: 5 year follow-up of a trial conducted by the interuniversity cardiology institute of the Netherlands. J Am Coll Cardiol. 1989;14:1609-1615.
Mark DB, Naylor CD, Hlatky MA, Califf RM, Topol EJ, Granger CB, Knight JD, Nelson CL, Lee KL, Clapp-Channing NE, Sutherland W, Pilote L, Armstrong PW. Use of medical resources and quality of life after acute myocardial infarction in Canada and the United States. N Engl J Med. 1994;331:1130-1135.
Pilote L, Califf RM, Sapp S, Miller DP, Mark DB, Weaver WD, Gore JM, Armstrong PW, Ohman EM, Topol EJ. Regional variation across the United States in the management of acute myocardial infarction: GUSTO-1 investigators—Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded coronary arteries. N Engl J Med. 1995;333:565-572.