- Email: [email protected]
Accuracy of exercise testing in the assessment of the severity of myocardial ischemia as determined by means of technetium-99m tetrofosmin SPECT scintigraphy
Accuracy of exercise testing in the assessment of the severity of myocardial ischemia as determined by means of technetium-99m tetrofosmin SPECT scintigraphy Alfredo R. Galassi, MD, FACC, FESC,a Salvatore Azzarelli, MD,a Lorenzo Lupo, MD,b Carmelo Mammana, MD,a Rosario Foti, MD,a Corrado Tamburino, MD,a Salvatore Musumeci, MD,c and Giuseppe Giuffrida, MD, FACCa Background. The separation of patients with suspected or known coronary artery disease into low- and high-risk subgroups by means of noninvasive testing is highly relevant in the selection of patients who require further diagnostic or therapeutic investigation. We evaluated whether exercise electrocardiographic variables during exercise testing might be a means of predicting the severity of myocardial ischemia as assessed with myocardial scintigraphy. Methods and Results. We retrospectively reviewed 816 consecutive patients (mean age, 57 ± 10 years) who underwent exercise technetium-99m tetrofosmin single photon emission computed tomography (SPECT) for the assessment of suspected or known coronary artery disease. Eight independent significant predictors of the extent and severity of reversible perfusion defects (ischemic perfusion score), which when integrated in a diagnostic algorithm satisfactorily discriminated patients with no reversible perfusion defects (sensitivity, 75%; specificity, 80%) and patients with severe impaired myocardial perfusion (≥11 ischemic perfusion score; sensitivity, 77%; specificity, 82%), were identified by means of stepwise discriminant analysis. However, patients with mildly to moderately impaired myocardial perfusion (≥1 but <11 ischemic perfusion score) were poorly discriminated (sensitivity, 50%; specificity, 78%). The set of variables that were significant (P < .0001) for prediction included sex, myocardial infarction, exercise angina, the maximal amount of ST segment depression, rate-pressure product threshold criteria, slope of ST segment depression, ST/heart rate index, and peak exercise heart rate. Conclusions. The results of the use of clinical and electrocardiographic exercise variables satisfactorily agrees with the results from scintigraphy only for patients with no reversible perfusion defects and with severely impaired myocardial perfusion. However, it fails as an approach with universal applicability. (J Nucl Cardiol 2000;7:575-83.) Key Words: Exercise testing • technetium-99m tetrofosmin single photon emission computed tomography myocardial scintigraphy • rate-pressure product • ST/heart rate index
The separation of patients with suspected or known coronary artery disease into low- and high-risk subgroups by means of noninvasive testing is highly relevant in the selection of patients who require further diagnostic or From the Institute of Cardiology, Ferrarotto Hospital,a Lecturership in Medical Statisticsb and Chair of Nuclear Medicine,c University of Catania, Italy. Received for publication Mar 29, 1999; final revision accepted May 16, 2000. The diagnostic algorithm described in this article has a patent pending. Reprint requests: Alfredo R. Galassi, MD, FACC, FESC, Via Antonello da Messina 75, Acicastello, 95021, Catania, Italy; [email protected]. Copyright © 2000 by the American Society of Nuclear Cardiology. 1071-3581/2000/$12.00 + 0 43/1/108731 doi:10.1067/mnc.2000.108731
therapeutic investigation. Exercise electrocardiography remains the most widely used noninvasive means for detecting coronary artery disease.1-3 However, the prediction of disease severity may be difficult. Several electrocardiographic variables have been used as a means of improving exercise test accuracy. Various authors have shown that a greater ST-segment depression during exercise testing is predictive of a higher severity of coronary artery disease,4,5 especially when it is observed at low workload.6,7 However, results obtained in previous and recent works have challenged this concept.8-10 Furthermore, computer-based processing of ST-segment depression such as ST/heart rate (HR) index and ST/HR slope have been compared with standard ST-segment 575
576
Galassi et al Accuracy of exercise testing for myocardial ischemia
analysis. Conflicting data also have been reported that either support11-14 or undermine15-17 the improved accuracy obtained by means of these criteria as compared with coronary angiography. Good correlations between the rate-pressure product measured during exercise testing and myocardial oxygen consumption have been shown by other authors.18-20 The measurement of such variables has been used as a means of assessing therapeutic interventions21 and comparing different exercise protocols.22-24 Coronary angiography generally has been the gold standard for determining the significance of coronary artery lesions, usually expressed in terms of percent diameter stenosis. However, this approach correlates only sufficiently with estimates obtained from quantitative myocardial perfusion.25,26 Indeed, coronary flow reserve is greatly influenced by other lesion characteristics, such as length, shape, and absolute cross-sectional luminal area, which all affect impedance to blood flow. Furthermore, coronary angiography is affected by the inherent verification bias of all invasive techniques (only patients with a high probability of coronary artery disease produced by means of the given test will undergo coronary angiography). Myocardial scintigraphy, however, is a means of providing an estimate of regional myocardial perfusion that accurately reflects the amount of ischemic myocardium,27,28 and it can be performed easily in all patients with suspected or known coronary artery disease. When tested against coronary angiography, myocardial tomographic scintigraphy has proved to be highly sensitive and specific.29-32 Therefore, the aim of the present study was to assess whether easily measured exercise testing variables might be a means of predicting the magnitude of the extent and severity of reversible perfusion defects as assessed by means of myocardial scintigraphy with Tc-99m tetrofosmin single photon emission computed tomography (SPECT).
METHODS Study Population The study group consisted of 816 consecutive patients (628 men and 188 women; mean age, 57 ± 10 years) retrospectively selected from 1140 patients who underwent technetium99m tetrofosmin SPECT scintigraphy in our laboratory in a 28month period. Patients were referred for myocardial perfusion scintigraphy to assess suspected coronary artery disease (259 patients, 32%), earlier coronary revascularization (60 patients, 7%), earlier myocardial infarction (411 patients, 50%), and coronary artery disease documented angiographically (86 patients, 11%). We excluded patients who had a recent myocar-
Journal of Nuclear Cardiology November/December 2000
dial infarction (≤6 weeks); had hypertrophic, restrictive, and idiopathic dilated cardiomyopathy; had electrocardiographic signs of left ventricular hypertrophy and valvular disease (including mitral valve prolapse); were receiving anti-anginal treatment because of inappropriate pharmacological washout or digitalis; or had uninterpretable exercise electrocardiograms, including left bundle branch block, ventricular preexcitation, paced rhythm, and resting ST-segment depression >1 mm.33 An earlier myocardial infarction (>6 weeks) was diagnosed by means of documented history (chest pain, characteristic electrocardiographic and creatinine kinase enzymes changes) or electrocardiographic evidence (pathologic Q waves, 0.04 second in duration in 2 or more leads) at the time of the study. There were 312 patients (38%) with a Q-wave myocardial infarction and 99 patients (12%) with a non–Q-wave myocardial infarction; in 5 patients (1%), a Q-wave myocardial infarction coexisted with a non–Q-wave myocardial infarction. The myocardial infarction location was anterior in 134 patients (33%), anteroseptal in 41 patients (10%), anterolateral in 8 patients (2%), inferior in 158 patients (39%), inferolateral in 27 patients (6%), inferoposterolateral in 5 patients (1%), inferoposterior in 8 patients (2%), posterior in 3 patients (0.5%), lateral in 6 patients (1.3%), posterolateral in 1 patient (0.2%), and undetermined in 18 patients (5%). In 3 patients, an anterior myocardial infarction was associated with an inferior myocardial infarction.
Exercise Testing After the washout of all antianginal medications, patients underwent treadmill exercise testing with the modified Bruce protocol.34 Test-termination criteria included ST-segment depression of 3 mm or greater, progressive angina pectoris, limiting dyspnea, more than 3 consecutive premature ventricular contractions, and a significant decrease (10 mm Hg or higher) in systolic blood pressure.3 Exercise tests were considered maximal when the HR achieved was 85% or more of maximal HR predicted for age and sex. Twelve-lead electrocardiograms and blood pressure measurements were recorded during the control pre-exercise period, at the end of each stage during exercise, and at each minute during recovery. Twelve electrocardiographic leads were continuously monitored throughout the test and for 5 minutes after exercise. The level of the ST segment was calculated, after signal averaging for 12-second periods, in each lead every minute by means of a computer-assisted system (Cardioline ECT stress) and confirmed in all cases by means of visual analysis. The maximal amount of ST segment depression in any lead was measured. Ischemic threshold was defined as 1 mm horizontal or downward or 1.5 mm upward sloping ST-segment depression at 0.06 second after the J point in 3 or more consecutive complexes of the same lead. When the ST segment level at rest standing was above the isoelectric line, the resting baseline ST segment level was considered to be 0; when it was below the isoelectric line, the negative value was used. HR, systolic blood pressure, and HRsystolic blood pressure (rate-pressure product) were measured at rest with the patient in an upright position, at the peak of exercise in all cases and at the ischemic threshold. Rate-pressure product
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
at the ischemic threshold or in the case of negative tests, at the peak of exercise (threshold criteria), was evaluated in all patients. Exercise time was measured at the peak of exercise in all cases and at the ischemic threshold. Similarly, exercise time was evaluated at the threshold criteria. ST/HR index (reported in µV/bpm) was calculated by dividing the maximal additional ST segment depression at end-exercise (corrected for any ST segment depression at rest in that lead on the upright pre-exercise control electrocardiogram) by the exercise-induced change in heart rate times 100.11
Technetium-99m Tetrofosmin SPECT Image acquisition. Tc-99m tetrofosmin was prepared from a freeze-dried kit (Myoview, Amersham International, Buckinghamshire, UK)35 by reconstitution with approximately 6 mL of a sterile sodium pertechnetate solution containing 20 to 24 mCi. Patients were injected from 60 to 90 seconds before the termination of exercise with 20 mCi of Tc-99m tetrofosmin, and SPECT imaging was obtained 1 hour after with an IGE Starcam 3000 AC tomographic gamma camera equipped with a low-energy high-resolution parallel hole collimator centered on the 140 keV-photopeak with a 20% window. Thirty-two projections were acquired over a 180degree arc, starting from 45 degrees right anterior oblique to 45 degrees left posterior oblique in a 40 cm field of view. Each projection was acquired for 30 seconds. Forty-eight hours later, patients were injected at rest with 24 mCi of Tc-99m tetrofosmin, and imaging was performed in the same manner as for the stress study. To minimize gall bladder activity, all patients were instructed to consume a light fatty meal after tracer injection and before imaging. Image processing and analysis. From the raw transaxial tomograms, orthogonal images were generated by means of oblique angle reconstruction, producing short-axis, horizontal long-axis, and vertical long-axis views. Representative sections composed of 3 short-axis sections (apical, midventricular, and basal) and the midventricular vertical long-axis combined to assess apical segments were selected for analysis, as described by Garcia et al.36 Myocardial segments were also grouped according to anterior, inferior, and lateral regions.29 Defect extension and severity were separately analyzed by 2 blinded observers. A 4-point scoring system for segmental uptake of Tc99m tetrofosmin (in which 0 = normal, 1 = mildly reduced, 2 = moderately reduced, and 3 = severely decreased uptake) was used. Patterns of perfusion defects were separately determined in each segment for rest and stress studies. The reversibility of radiotracer uptake was defined according to the change in segmental score between the stress and rest studies: nonreversible (3-3, 2-2, 2-3), partially reversible (3-2, 2-1), and completely reversible (3-0, 2-0, 3-1). A reversible defect was considered to be present when a completely reversible or partially reversible uptake defect was present. A nonreversible uptake defect was scored when uptake was worse in the rest image, as compared with the stress image (1-2, 0-3, 1-3, 0-2). Patterns of completely reversible perfusion defects (1-0) and nonreversible perfusion
Galassi et al Accuracy of exercise testing for myocardial ischemia
577
defects (1-1, 0-1) were not considered to be significant for analysis and were grouped as normal.37 A total ischemic perfusion score that takes into account both the extent and severity of reversible perfusion defects was determined by adding the difference in the perfusion score from stress to rest for those segments in which the defects proved to be partially or completely reversible. Of the total 13.056 segments scored, agreement on the total ischemic perfusion score was reached in 12.142 (93%); whereas, in the remaining segments, a split decision was resolved by means of consensus.
Statistical Analysis A stepwise discriminant analysis was performed to identify variables that were a means of predicting patients classified according to the ischemic perfusion score. These 23 variables were selected as possible predictors for discrimination: age (years), sex (male and female), risk factors (family history of coronary artery disease, smoking, hypertension, diabetes mellitus, hyperlipidemia; each at 2 levels, yes or no), test referral (suspected and known coronary artery disease), previous myocardial infarction (at 3 levels, Q wave, non-Q wave, absent), resting HR (beats/min), resting rate-pressure product (beats/min × mm Hg), peak exercise HR (beats/min), peak exercise rate-pressure product (beats/min × mm Hg), peak exercise time (minutes), threshold criteria HR (beats/min), threshold criteria rate-pressure product (beats/min × mm Hg), threshold criteria exercise time (minutes), exercise angina (yes or no), percent of maximal heart rate achieved during exercise (%), location of myocardial ischemia (anterior, inferior, lateral), maximal amount of ST segment depression (mm), ST/HR index (µV/beats/min), and slope of ST segment depression (at 3 levels, up-sloping, not up-sloping, negative). The BMDP statistical package was used with equal prior probabilities assigned to groups. A logistic regression analysis, with GLIM, was performed as a means of analyzing differences among results given by means of the jackknife classification. The latter is a validation method by which each case is in turn classified by discrimination functions derived by performing discriminant analysis on the remaining n-1 cases. Thus sensitivity and specificity rates can be calculated with higher validity. Because of the classification of extent and severity of ischemia in more than 2 groups, sensitivity was expressed as the percent of cases correctly classified into each group; whereas specificity was expressed as the correct classification of cases classified into either of the remaining groups. BMDP is a registered trademark of BMDP Statistical Software. Glim is a registered trademark of the Royal Statistical Society. The chi-square test and analysis of variance were used as a means of testing and assessing univariate significant differences for variables among patients with different ischemic perfusion scores. To correct increases of type-I error caused by multiple comparisons, a more conservative P value of .01 replaced the traditional .05 level of significance. Quantitative data are expressed as the mean ± SD. For categorical data, relative frequencies are given.
578
Galassi et al Accuracy of exercise testing for myocardial ischemia
Journal of Nuclear Cardiology November/December 2000
Table 1. Clinical variables in patients with an ischemic perfusion score of 0 (group 1), with an ischemic perfusion score of 1 or greater but less than 11 (group 2), and an ischemic perfusion score of 11 or greater (group 3)
Reversible perfusion defects Group 1 Patients 284 (35%) Age (y) 55 ± 10 Sex (men/women) 172/112 (61%/39%) Test referral (suspected 114/170 (40%/60%) CAD/known CAD) Family history of CAD 125/159 (44%/56%) (yes/no) Smoker (yes/no) 124/160 (44%/56%) Hypertension (yes/no) 144/140 (51%/49%) Diabetes mellitus (yes/no) 43/241 (15%/85%) Hyperlipidemia (yes/no) 134/150 (47%/53%) MI: Q-wave/ 50/23/211 (18%/8%/74%) non–Q-wave/absent
Group 2
Group 3
P value
457 (51%) 57 ± 10 388/69 (85%/15%) 51/406 (11%/89%)
75 (9%) 59 ± 10 68/7 (91%/9%) 6/69 (8%/92%)
<.0001 <.0001 <.0001
229/228 (50%/50%)
32/43 (43%/57%)
NS
254/203 (56%/44%) 195/262 (43%/57%) 90/367 (20%/80%) 210/247 (46%/54%) 229/62/166 (50%/14%/36%)
37/38 (49%/51%) NS 38/37 (51%/49%) NS 23/52 (31%/69%) NS 36/39 (48%/52%) NS 31/16/28 (41%/21%/37%) <.0001
Data are presented as mean value ± SD or number (%) of patients. MI, Myocardial infarction; CAD, coronary artery disease; NS, not significant.
RESULTS There were 614 maximal tests (75%) and 202 submaximal tests (25%). Exercise tests were stopped because of ST-segment depression 3 mm or greater in only 8 cases (1%). An exercise ischemic threshold was achieved in 326 cases (40%). Prediction of Impaired Myocardial Perfusion According to discriminant analysis, differences among patients were maximized when they were allocated in 3 groups according to ischemic perfusion score: patients without reversible perfusion defects (group 1), patients with an ischemic perfusion score of 1 or higher but less than 11 (group 2), and patients with an ischemic perfusion score of 11 or higher (group 3). This classification was confirmed on the basis of previous reports, which suggested that the severity of impaired myocardial perfusion by means of scintigraphy is a means of stratifying patients with different prognoses.38 An exercise ischemic threshold was achieved in 27 of 284 cases of patients in group 1 (10%), in 231 of 457 cases of patients in group 2 (51%), and in 68 of 75 cases of patients in group 3 (91%; P < .0001). Tables 1 and 2 summarize all electrocardiographic parameters measured in the 3 groups of patients and P values for differences at univariate analysis. Variables that were altogether significant (P < .001) by means of the U statistic for discrimination were hier-
archically moved by means of this stepwise method: maximal amount of ST-segment depression, presence and type of myocardial infarction, rate-pressure product threshold criteria, slope of ST-segment depression, sex, angina, ST/HR index, peak exercise HR. Exercise variables were also combined into a new global exercise index according to this formula: [Rate-pressure product threshold criteria / 100 + Maximal amount of ST segment depression / Peak exercise heart rate + (1 / ST / HR index)]1/2 in which if ST/HR index equals 0, then 1/ST/HR index should equal 0.01, to discriminate patients in the 3 groups according to the ischemic perfusion score (Figure 1A and B). Discriminant analysis was then repeated and, by using the jackknife classification to validate the classification functions for the 3 groups of patients, provided a sensitivity rate of 75% (214/284), 50% (229/457), and 77% (58/75), respectively, and a specificity rate of 80% (425/532), 78% (281/359), and 82% (611/741), respectively. According to this method, among the misclassified patients of group 1, 65 (93%) were allocated in group 2, and the remaining 5 patients (7%) were allocated in group 3. Similarly, among the misclassified patients of group 3, 13 (76%) were allocated in group 2, and the remaining 4 patients (24%) were allocated in group 1. Conversely, among misclassified patients of group 2, 103 (45%) were allocated in group 1, and 125 (55%) were allocated in group 3.
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
Galassi et al Accuracy of exercise testing for myocardial ischemia
A
B Figure 1. A, Diagnostic clinical and exercise algorithm in male patients. B, Diagnostic clinical and exercise algorithm in female patients. MI, Myocardial infarction; group 1, patients without impaired myocardial perfusion; group 2, patients with mild to moderate impaired myocardial perfusion; group 3, patients with severe impaired myocardial perfusion; global exercise index, [rate-pressure product threshold criteria/100 + maximal amount of ST-segment depression/peak exercise HR + (1/ST/HR index)*]1/2, (*when ST/HR index = 0, then 1/ST/HR index = 0.01).
579
580
Galassi et al Accuracy of exercise testing for myocardial ischemia
Journal of Nuclear Cardiology November/December 2000
Table 2. Exercise variables in patients with an ischemic perfusion score of 0 (group 1), with an ischemic perfusion score of 1 or greater but less than 11 (group 2) and an ischemic perfusion score of 11 or greater (group 3)
Reversible perfusion defects
Maximal ST ↓(mm) ST/HR index (µV/beats/min) Rest HR (beats/min) Rate-pressure product (beats/min × mm Hg) Peak exercise HR (beats/min) Rate-pressure product (beats/min × mm Hg) Time (min) Threshold criteria HR (beats/min) Rate-pressure product (beats/min × mm Hg) Time (min) Maximal HR (%) Exercise angina (no/yes) Slope of ST↓ (↑/non-↑/negative) Location of ischemia (ant/lat/inf/ant+inf/ ant+lat/inf+lat)
Group 1
Group 2
0.2 ± 0.5 0.4 ± 0.9
1.0 ± 0.9 1.8 ± 2.1
Group 3 1.9 ± 0.9 4.7 ± 4.7
P value <.0001 <.0001
82 ± 14 11.362 ± 2.588
78 ± 14 10.777 ± 2.344
78 ± 13 10.806 ± 2.096
<.003 <.003
150 ± 15 27.022 ± 3.473
141 ± 17 25.325 ± 3.789
131 ± 19 22.644 ± 4.089
<.0001 <.0001
16 ± 3
15 ± 3
14 ± 5
<.0001
149 ± 15 26.725 ± 3.483
135 ± 19 23.506 ± 4.339
115 ± 17 18.453 ± 3.738
<.0001 <.0001
16 ± 3 95 ± 9 265/19 (93%/7%) 69/23/192 (24%/8%/68%) —
14 ± 4 88 ± 10 306/151 (67%/33%) 157/169/131 (34%/37%/29%) 156/97/175/19/7/3 (34%/21%/38%/4%/2%/1%)
10 ± 4 82 ± 10 21/54 (28%/72%) 12/59/4 (16%/79%/5%) 34/13/10/11/7/0 (45%/17%/13%/15%/10%/0%)
<.0001 <.0001 <.0001 <.0001 <.0001
Data are presented as mean value ± SD or number (%) of patients. ST ↓ , ST segment depression; ↑, up-sloping; non-↑, non-up-sloping; ant, anterior; lat, lateral; inf, inferior; NS, not significant.
An allocation tree was finally built up by using significant clinical and exercise variables to discriminate among patients in the 3 groups. According to each combination of clinical variables (sex and presence and type of myocardial infarction) and exercise variables (maximal amount of ST-segment depression, rate-pressure product threshold criteria, slope of ST-segment depression, angina, ST/HR index, peak exercise HR), cutoff limits of the global exercise index for allocation into each of the 3 groups were given (Figure 1A and B). We also performed the same analysis including clinical variables only. The number of correctly classified patients was 216 of 284 (76%) in group 1, 229 of 457 (50%) in group 2, and 16 of 75 (21%) in group 3.
DISCUSSION This study demonstrates that clinical and electrocardiographic exercise variables may satisfactorily discrim-
inate patients with no reversible perfusion defects or with severe impaired myocardial perfusion at scintigraphy, but is far less satisfactory in those patients with mild to moderate perfusion defects. We assessed whether easily measured exercise testing variables might be able to predict the severity of myocardial ischemia. Although coronary angiography has traditionally been used as the reference standard for the evaluation of coronary artery disease, myocardial perfusion is an integrated response of an anatomic-hemodynamic system in which the coronary stenosis is only one component.25,26 Coronary angiography may not accurately indicate the physiologic significance of a coronary artery stenosis and may generally suffer from the inherent methodological verification bias. Only patients with noninvasive positive test results are more likely to have the results verified by means of coronary angiography, whereas patients with negative exercise test results are rarely referred for subsequent invasive studies.39 This
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
increases sensitivity, because false-negative results are unlikely to be discovered, and decreases specificity. To overcome such limitations, we compared exercise testing variables with myocardial SPECT. Although some technical disadvantages, such as variable tracer myocardial extraction fraction at different flow rates and lack of attenuation correction factors, may prevent this technique from being a quantitative tool for measuring myocardial perfusion, myocardial imaging with Tc-99m–labeled compound with SPECT provides a sensitive means for detecting exercise-induced regional hypoperfusion and can be easily used as a means of evaluating all patients with suspected or known coronary artery disease.27-30 In the present study, among all clinical and exercise variables tested, multivariate discriminant analysis identified 8 independent predictors for the discrimination of patients without impaired myocardial perfusion and with different severity of impaired myocardial perfusion. By using a diagnostic algorithm that included clinical and exercise variables, we were able to discriminate patients with no reversible perfusion defects and patients with severe impaired myocardial perfusion with a sensitivity rate of 75% and 77%, respectively. However, a less satisfactory sensitivity rate of 50% was obtained for the identification of patients with mild to moderate impairment of myocardial perfusion. Among all electrocardiographic variables, maximal amount of ST-segment depression, rate-pressure product threshold criteria, ST/HR index, and peak exercise HR were significant. These results are concordant with those obtained by other authors5,6 and confirm the usefulness of ST-segment analysis for the discrimination between patients with severe disease and healthy patients. Furthermore, the rate-pressure product that best correlates with myocardial oxygen consumption represents an index of myocardial metabolic demand.1820 It has been used by different authors as an objective and accurate estimate of the severity of myocardial ischemia. Waters et al showed that considerable variability of exercise tolerance occurs even when ratepressure product at the ischemic threshold remains constant.22 Similarly, Kaski et al21 have shown that a higher rate-pressure product at the ischemic threshold after sublingual nitrates is a means of identifying those patients in whom the prevailing mechanism of action is an increase in coronary flow reserve. Although myocardial wall tension and contractility should also have been taken into account, because they might contribute to myocardial oxygen consumption, they cannot be directly measured during exercise testing and are not even indirectly measured by means of rate-pressure product. Recently, computer-based processing of ST-segment depression such as ST/HR index and ST/HR slope has
Galassi et al Accuracy of exercise testing for myocardial ischemia
581
been compared with standard ST-segment analysis to improve method sensitivity. Various authors have shown an improved sensitivity by using the ST/HR index,11-14 whereas other authors have not.15-17,40,41 The degree of increase in accuracy of ST/HR index may depend on the population being studied42 and on referral bias for coronary angiography.43 Indeed, our study showed that the ST/HR index may be of value when applied to an unselected population. As compared with our method, the added value of myocardial scintigraphy is as large as 24.6% of patients with false-positive results, in whom a perfusion defect would have been diagnosed. Conversely, the added value of myocardial scintigraphy is as great as 50% for those patients with mild to moderate perfusion defects, in whom a perfusion defect would not be diagnosed (22.5%) or in whom a severe perfusion defect would be diagnosed (27.3%). For those patients with a severe perfusion defect, myocardial scintigraphy proved to have the lowest added value of 22.7%, because only 5.3% of patients would have been found to not have any perfusion defects, whereas the remaining 17.3% would have been found to have mild to moderate perfusion defects. Conversely, the added value of our method, as compared with clinical variables alone, was highest for those patients with severe perfusion defects, in whom the added value was as great as 56%. Thus these findings suggest that, as a first investigation, clinical variables and exercise testing may be a reasonable method for discriminating patients with no reversible perfusion defects and with severe impaired myocardial perfusion, but it is of little value in identifying patients with mild to moderate myocardial perfusion abnormalities. Limitations of the Study Among all exercise variables measured, we used the rate-pressure product threshold criterion. This electrocardiographic index results from the rate-pressure product measured at the ischemic threshold in the case of positive exercise test results and from rate-pressure product value measured at peak exercise in the case of negative test results. Although it represents a measurement that differs among patients, it allows for the discrimination of the severity of myocardial hypoperfusion in all exercise tests performed. Our study did not take into account the ST/HR slope, a method that is much more cumbersome to measure and is only marginally more accurate than the ST/HR index.11-13 Furthermore, given the reported differences in the extraction between Tc-tetrofosmin, Tc-sestamibi, and thallium-201 imaging, findings obtained with tetrofosmin might be marginally
582
Galassi et al Accuracy of exercise testing for myocardial ischemia
different than those obtained with other tracers, especially when using pharmacologic coronary vasodilators. Conclusions The exercise test continues to play an important role in the assessment of patients with ischemic heart disease and, in many cases, continues to represent the first investigation for those patients who can exercise maximally and in whom exercise is interpretable. However, subset selection of patients will continue to be a problem if the intention is to identify a particular approach that will have universal applicability. We thank Iain Halliday, MA, for constructive comments and editorial assistance.
References 1. Ellestad MH. Stress testing: Principles and practice. New York: Davis; 1996. p. 111-38. 2. Bruce RA, Hossack KF, Rouen TA, Hofer V. Enhanced risk assessment for primary coronary heart disease events by maximal exercise testing: 10 years’ experience of Seattle Heart Watch. J Am Coll Cardiol 1983;2:565-73. 3. Detrano R, Gianrossi R, Froelicher V. The diagnostic accuracy of the exercise electrocardiogram: A meta-analysis of 22 years of research. Prog Cardiovasc Dis 1989;3:173-206. 4. Martin CM, McConahay DR. Maximal treadmill exercise electrocardiography: Correlations with coronary arteriography and cardiac hemodynamics. Circulation 1972;26:956-62. 5. Weiner DA, McCabe CH, Ryan TJ. Identification of patients with left main and three vessel coronary disease with clinical and exercise test variables. Am J Cardiol 1980;46:21-7. 6. Goldschlanger N, Selzer A, Cohn K. Treadmill stress tests as indicators of presence and presence and severity of coronary artery disease. Ann Int Med 1976;85:277-86. 7. Podrid PJ, Graboys TB, Lown B. Prognosis of medically treated patients with coronary artery disease with profound ST-segment depression during exercise testing. N Engl J Med 1981;305:1111-6. 8. Borer JS, Brensike JF, Redwood DR, et al. Limitations of electrocardiographic response to exercise in predicting coronary artery disease. N Eng J Med 1975;293:367-71. 9. Weiner DA, Ryan TJ, McCabe CH, et al. Exercise stress testing: Correlations among history of angina, ST segment response and prevalence of coronary artery disease in the Coronary Artery Surgery Study (CASS). N Engl J Med 1979;301:230-5. 10. Bogaty P, Guimond J, Robitaille NM, et al. A reappraisal of exercise electrocardiographic indexes of the severity of ischemic heart disease: Angiographic and scintigraphic correlates. J Am Coll Cardiol 1997;29:1497-504. 11. Kligfield P, Ameisen O, Okin PM. Heart rate adjustment of ST segment depression for improved detection of coronary artery disease. Circulation 1989;79:245-55. 12. Okin PM, Kligfield P. Heart rate adjustment of segment depression and performance of the exercise electrocardiogram: A critical evaluation. J Am Coll Cardiol 1995;25:1726-35. 13. Okin PM, Lauer MS, Kligfield P. Chronotropic response to exercise. Improved performance of ST-segment depression criteria after adjustment for heart rate reserve. Circulation 1996;94:3226-31.
Journal of Nuclear Cardiology November/December 2000
14. Elamin MS, Boyle R, Kardash MM, et al. Accurate detection of coronary heart disease by new exercise test. Br Heart J 1982;48: 311-20. 15. Lachterman B, Lehmann KG, Detrano R, Neutel J, Froelicher VF. Comparison of ST segment/heart rate index to standard ST criteria for analysis of exercise electrocardiogram. Circulation 1990;82:44-50. 16. Bobbio M, Detrano R, Schmid J-J, et al. Exercise-induced ST depression and ST/heart rate index to predict triple-vessel or left main coronary disease: A multicenter analysis. J Am Coll Cardiol 1992; 19:11-8. 17. Morise AP, Duval RD. Accuracy of ST/heart rate index in the diagnosis of coronary artery disease. Am J Cardiol 1992;69:603-6. 18. Gobel FL, Nordstrom LA, Nelson RR, Jorgensen CR, Wang Y. The ratepressure product as an index of myocardial oxygen consumption during exercise in patients with angina pectoris. Circulation 1978;57: 549-56. 19. Kitamura K, Joregensen CR, Gobel FL, Taylor HL, Wang Y. Hemodynamic correlates of myocardial oxygen consumption during upright exercise. J Appl Physiol 1972;32:516-20. 20. Nelson RR, Gobel FL, Jorgensen CR, Wang Y, Wang Y, Taylor HL. Hemodynamic predictors of myocardial oxygen consumption during static and dynamic exercise. Circulation 1974;50:1179-84. 21. Kaski JC, Plaza LR, Meran DO, Araujo L, Chierchia S, Maseri A. Improved coronary supply: Prevailing mechanism of action of nitrates in chronic stable angina. Am Heart J 1985;110:238-45. 22. Waters DD, McCans JL, Crean PA. Serial exercise testing in patients effort angina: Variable tolerance, fixed threshold. J Am Coll Cardiol 1985;6:1011-5. 23. Panza JA, Arshed AQ, Diodati JG, Callahan TS, Bonow RO, Epstein SE. Long-term variation in myocardial ischemia during daily life in patients with stable coronary artery disease: Its relation to changes in the ischemic threshold. J Am Coll Cardiol 1992;19:500-6. 24. Garber CE, Carleton RA, Camaione DN, Heller GV. The threshold for myocardial ischemia varies in patients with coronary artery disease depending on the exercise protocol. J Am Coll Cardiol 1991;17: 1256-62. 25. Wilson RF, Marcus ML, Christensen BV, Talman C, White CW. Accuracy of exercise electrocardiography in detecting physiologically significant coronary arterial lesions. Circulation 1991;83:412-21. 26. Folland ED, Vogel RA, Hartigan P, et al. Relation between coronary artery stenosis assessed by visual, caliper, and computer methods and exercise capacity in patients with single-vessel coronary artery disease. Circulation 1994;89:2005-14. 27. Gould KL. Identifying and measuring severity of coronary artery stenosis: Quantitative coronary arteriography and positron emission tomography. Circulation 1988;78:237-45. 28. Di Carli M, Czernin J, Hoh CK, et al. Relation among stenosis severity, myocardial blood flow, and flow reserve in patients with coronary artery disease. Circulation 1995;91:1944-51. 29. Kahn JK, McGhie I, Akers MS, et al. Quantitative rotational tomography with 201Tl and 99mTc 2-methoxy-isobutyl-isonitrile; a direct comparison in normal individuals and patients with coronary artery disease. Circulation 1989;79:1282-93. 30. Zaret BL, Rigo P, Wackers FJT, et al. Myocardial perfusion imaging with 99mTc Tetrofosmin: Comparison to 201Tl imaging and coronary angiography in a phase III multicenter trial. Circulation 1995;91:313-9. 31. Tamaki N, Yonekura Y, Mukai T, et al. Stress thallium-201 transaxial emission computed tomography: Quantitative versus qualitative analysis for evaluation of coronary artery disease. J Am Coll Cardiol 1984;4:1213-20. 32. DePasquale EE, Nody AC, DePuey EG, et al. Quantitative rotational thallium-201 tomography for identifying and localizing coronary artery disease. Circulation 1988;77:316-21.
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
Galassi et al Accuracy of exercise testing for myocardial ischemia
33. Gibbons RJ, Balady GJ, Beasley JW et al. ACC/AHA Guidelines for exercise testing: A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee on exercise testing). J Am Coll Cardiol 1997;30(1):260-315. 34. Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake and nomografic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 1973;85:546-62. 35. Higley B, Smith FW, Smith T, et al. Technetium-99m-1, 2bis [bis(2ethoxyethyl) phosphino] ethane: Human biodistribution, dosimetry and safety of a new myocardial perfusion imaging agent. J Nucl Med 1993;34:30-8. 36. Garcia EV, Van Train K, Maddahi J, et al. Quantification of rotational thallium-201 myocardial tomography. J Nucl Med 1985;26:17-26. 37. Chouraqui P, Maddahi J, Ostrzega E, et al. Quantitative exercise thallium-201 rotational tomography for evaluation of patients with prior myocardial infarction. Am J Cardiol 1990;66:151-7. 38. Hachamovitch R, Berman DS, Shaw LJ, et al. Incremental prognostic value of myocardial perfusion single photon emission computed
O
39. 40.
41.
42.
43.
583
tomography for the prediction of cardiac death: Differential stratification for risk of cardiac death and myocardial infarction. Circulation 1998;97:535-43. Detrano R, Froelicher VF. Exercise testing: Uses and limitation considering recent studies. Prog Cardiovasc Dis 1988;31:173-204. Okin PM, Kligfield P, Ameisen O, Goldberg HL, Borer JS. Improved accuracy of exercise electrocardiogram: Identification of three-vessel coronary disease in stable angina pectoris by analysis of peak raterelated changes in ST segments. Am J Cardiol 1985;55:271-6. Bobbio M, Detrano R. A lesson from the controversy about heart rate adjustment of ST segment depression. Circulation 1991;84: 1410-3. Okin PM, Kligfield P. Population selection and performance of the exercise ECG for the identification of coronary artery disease. Am Heart J 1994;127:296-304. Morise A, Duval R. Diagnostic accuracy of heart rate-adjusted ST segment compared with standard ST-segment criteria. Am J Cardiol 1995;75:118-21.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS:
NEW ADDRESS:
Affix the address label from a recent issue of the journal here.
Clearly print your new address here. Name Address City / State / ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 1-800-654-2452 Outside the U.S., call 407-345-4000
The separation of patients with suspected or known coronary artery disease into low- and high-risk subgroups by means of noninvasive testing is highly relevant in the selection of patients who require further diagnostic or From the Institute of Cardiology, Ferrarotto Hospital,a Lecturership in Medical Statisticsb and Chair of Nuclear Medicine,c University of Catania, Italy. Received for publication Mar 29, 1999; final revision accepted May 16, 2000. The diagnostic algorithm described in this article has a patent pending. Reprint requests: Alfredo R. Galassi, MD, FACC, FESC, Via Antonello da Messina 75, Acicastello, 95021, Catania, Italy; [email protected]. Copyright © 2000 by the American Society of Nuclear Cardiology. 1071-3581/2000/$12.00 + 0 43/1/108731 doi:10.1067/mnc.2000.108731
therapeutic investigation. Exercise electrocardiography remains the most widely used noninvasive means for detecting coronary artery disease.1-3 However, the prediction of disease severity may be difficult. Several electrocardiographic variables have been used as a means of improving exercise test accuracy. Various authors have shown that a greater ST-segment depression during exercise testing is predictive of a higher severity of coronary artery disease,4,5 especially when it is observed at low workload.6,7 However, results obtained in previous and recent works have challenged this concept.8-10 Furthermore, computer-based processing of ST-segment depression such as ST/heart rate (HR) index and ST/HR slope have been compared with standard ST-segment 575
576
Galassi et al Accuracy of exercise testing for myocardial ischemia
analysis. Conflicting data also have been reported that either support11-14 or undermine15-17 the improved accuracy obtained by means of these criteria as compared with coronary angiography. Good correlations between the rate-pressure product measured during exercise testing and myocardial oxygen consumption have been shown by other authors.18-20 The measurement of such variables has been used as a means of assessing therapeutic interventions21 and comparing different exercise protocols.22-24 Coronary angiography generally has been the gold standard for determining the significance of coronary artery lesions, usually expressed in terms of percent diameter stenosis. However, this approach correlates only sufficiently with estimates obtained from quantitative myocardial perfusion.25,26 Indeed, coronary flow reserve is greatly influenced by other lesion characteristics, such as length, shape, and absolute cross-sectional luminal area, which all affect impedance to blood flow. Furthermore, coronary angiography is affected by the inherent verification bias of all invasive techniques (only patients with a high probability of coronary artery disease produced by means of the given test will undergo coronary angiography). Myocardial scintigraphy, however, is a means of providing an estimate of regional myocardial perfusion that accurately reflects the amount of ischemic myocardium,27,28 and it can be performed easily in all patients with suspected or known coronary artery disease. When tested against coronary angiography, myocardial tomographic scintigraphy has proved to be highly sensitive and specific.29-32 Therefore, the aim of the present study was to assess whether easily measured exercise testing variables might be a means of predicting the magnitude of the extent and severity of reversible perfusion defects as assessed by means of myocardial scintigraphy with Tc-99m tetrofosmin single photon emission computed tomography (SPECT).
METHODS Study Population The study group consisted of 816 consecutive patients (628 men and 188 women; mean age, 57 ± 10 years) retrospectively selected from 1140 patients who underwent technetium99m tetrofosmin SPECT scintigraphy in our laboratory in a 28month period. Patients were referred for myocardial perfusion scintigraphy to assess suspected coronary artery disease (259 patients, 32%), earlier coronary revascularization (60 patients, 7%), earlier myocardial infarction (411 patients, 50%), and coronary artery disease documented angiographically (86 patients, 11%). We excluded patients who had a recent myocar-
Journal of Nuclear Cardiology November/December 2000
dial infarction (≤6 weeks); had hypertrophic, restrictive, and idiopathic dilated cardiomyopathy; had electrocardiographic signs of left ventricular hypertrophy and valvular disease (including mitral valve prolapse); were receiving anti-anginal treatment because of inappropriate pharmacological washout or digitalis; or had uninterpretable exercise electrocardiograms, including left bundle branch block, ventricular preexcitation, paced rhythm, and resting ST-segment depression >1 mm.33 An earlier myocardial infarction (>6 weeks) was diagnosed by means of documented history (chest pain, characteristic electrocardiographic and creatinine kinase enzymes changes) or electrocardiographic evidence (pathologic Q waves, 0.04 second in duration in 2 or more leads) at the time of the study. There were 312 patients (38%) with a Q-wave myocardial infarction and 99 patients (12%) with a non–Q-wave myocardial infarction; in 5 patients (1%), a Q-wave myocardial infarction coexisted with a non–Q-wave myocardial infarction. The myocardial infarction location was anterior in 134 patients (33%), anteroseptal in 41 patients (10%), anterolateral in 8 patients (2%), inferior in 158 patients (39%), inferolateral in 27 patients (6%), inferoposterolateral in 5 patients (1%), inferoposterior in 8 patients (2%), posterior in 3 patients (0.5%), lateral in 6 patients (1.3%), posterolateral in 1 patient (0.2%), and undetermined in 18 patients (5%). In 3 patients, an anterior myocardial infarction was associated with an inferior myocardial infarction.
Exercise Testing After the washout of all antianginal medications, patients underwent treadmill exercise testing with the modified Bruce protocol.34 Test-termination criteria included ST-segment depression of 3 mm or greater, progressive angina pectoris, limiting dyspnea, more than 3 consecutive premature ventricular contractions, and a significant decrease (10 mm Hg or higher) in systolic blood pressure.3 Exercise tests were considered maximal when the HR achieved was 85% or more of maximal HR predicted for age and sex. Twelve-lead electrocardiograms and blood pressure measurements were recorded during the control pre-exercise period, at the end of each stage during exercise, and at each minute during recovery. Twelve electrocardiographic leads were continuously monitored throughout the test and for 5 minutes after exercise. The level of the ST segment was calculated, after signal averaging for 12-second periods, in each lead every minute by means of a computer-assisted system (Cardioline ECT stress) and confirmed in all cases by means of visual analysis. The maximal amount of ST segment depression in any lead was measured. Ischemic threshold was defined as 1 mm horizontal or downward or 1.5 mm upward sloping ST-segment depression at 0.06 second after the J point in 3 or more consecutive complexes of the same lead. When the ST segment level at rest standing was above the isoelectric line, the resting baseline ST segment level was considered to be 0; when it was below the isoelectric line, the negative value was used. HR, systolic blood pressure, and HRsystolic blood pressure (rate-pressure product) were measured at rest with the patient in an upright position, at the peak of exercise in all cases and at the ischemic threshold. Rate-pressure product
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
at the ischemic threshold or in the case of negative tests, at the peak of exercise (threshold criteria), was evaluated in all patients. Exercise time was measured at the peak of exercise in all cases and at the ischemic threshold. Similarly, exercise time was evaluated at the threshold criteria. ST/HR index (reported in µV/bpm) was calculated by dividing the maximal additional ST segment depression at end-exercise (corrected for any ST segment depression at rest in that lead on the upright pre-exercise control electrocardiogram) by the exercise-induced change in heart rate times 100.11
Technetium-99m Tetrofosmin SPECT Image acquisition. Tc-99m tetrofosmin was prepared from a freeze-dried kit (Myoview, Amersham International, Buckinghamshire, UK)35 by reconstitution with approximately 6 mL of a sterile sodium pertechnetate solution containing 20 to 24 mCi. Patients were injected from 60 to 90 seconds before the termination of exercise with 20 mCi of Tc-99m tetrofosmin, and SPECT imaging was obtained 1 hour after with an IGE Starcam 3000 AC tomographic gamma camera equipped with a low-energy high-resolution parallel hole collimator centered on the 140 keV-photopeak with a 20% window. Thirty-two projections were acquired over a 180degree arc, starting from 45 degrees right anterior oblique to 45 degrees left posterior oblique in a 40 cm field of view. Each projection was acquired for 30 seconds. Forty-eight hours later, patients were injected at rest with 24 mCi of Tc-99m tetrofosmin, and imaging was performed in the same manner as for the stress study. To minimize gall bladder activity, all patients were instructed to consume a light fatty meal after tracer injection and before imaging. Image processing and analysis. From the raw transaxial tomograms, orthogonal images were generated by means of oblique angle reconstruction, producing short-axis, horizontal long-axis, and vertical long-axis views. Representative sections composed of 3 short-axis sections (apical, midventricular, and basal) and the midventricular vertical long-axis combined to assess apical segments were selected for analysis, as described by Garcia et al.36 Myocardial segments were also grouped according to anterior, inferior, and lateral regions.29 Defect extension and severity were separately analyzed by 2 blinded observers. A 4-point scoring system for segmental uptake of Tc99m tetrofosmin (in which 0 = normal, 1 = mildly reduced, 2 = moderately reduced, and 3 = severely decreased uptake) was used. Patterns of perfusion defects were separately determined in each segment for rest and stress studies. The reversibility of radiotracer uptake was defined according to the change in segmental score between the stress and rest studies: nonreversible (3-3, 2-2, 2-3), partially reversible (3-2, 2-1), and completely reversible (3-0, 2-0, 3-1). A reversible defect was considered to be present when a completely reversible or partially reversible uptake defect was present. A nonreversible uptake defect was scored when uptake was worse in the rest image, as compared with the stress image (1-2, 0-3, 1-3, 0-2). Patterns of completely reversible perfusion defects (1-0) and nonreversible perfusion
Galassi et al Accuracy of exercise testing for myocardial ischemia
577
defects (1-1, 0-1) were not considered to be significant for analysis and were grouped as normal.37 A total ischemic perfusion score that takes into account both the extent and severity of reversible perfusion defects was determined by adding the difference in the perfusion score from stress to rest for those segments in which the defects proved to be partially or completely reversible. Of the total 13.056 segments scored, agreement on the total ischemic perfusion score was reached in 12.142 (93%); whereas, in the remaining segments, a split decision was resolved by means of consensus.
Statistical Analysis A stepwise discriminant analysis was performed to identify variables that were a means of predicting patients classified according to the ischemic perfusion score. These 23 variables were selected as possible predictors for discrimination: age (years), sex (male and female), risk factors (family history of coronary artery disease, smoking, hypertension, diabetes mellitus, hyperlipidemia; each at 2 levels, yes or no), test referral (suspected and known coronary artery disease), previous myocardial infarction (at 3 levels, Q wave, non-Q wave, absent), resting HR (beats/min), resting rate-pressure product (beats/min × mm Hg), peak exercise HR (beats/min), peak exercise rate-pressure product (beats/min × mm Hg), peak exercise time (minutes), threshold criteria HR (beats/min), threshold criteria rate-pressure product (beats/min × mm Hg), threshold criteria exercise time (minutes), exercise angina (yes or no), percent of maximal heart rate achieved during exercise (%), location of myocardial ischemia (anterior, inferior, lateral), maximal amount of ST segment depression (mm), ST/HR index (µV/beats/min), and slope of ST segment depression (at 3 levels, up-sloping, not up-sloping, negative). The BMDP statistical package was used with equal prior probabilities assigned to groups. A logistic regression analysis, with GLIM, was performed as a means of analyzing differences among results given by means of the jackknife classification. The latter is a validation method by which each case is in turn classified by discrimination functions derived by performing discriminant analysis on the remaining n-1 cases. Thus sensitivity and specificity rates can be calculated with higher validity. Because of the classification of extent and severity of ischemia in more than 2 groups, sensitivity was expressed as the percent of cases correctly classified into each group; whereas specificity was expressed as the correct classification of cases classified into either of the remaining groups. BMDP is a registered trademark of BMDP Statistical Software. Glim is a registered trademark of the Royal Statistical Society. The chi-square test and analysis of variance were used as a means of testing and assessing univariate significant differences for variables among patients with different ischemic perfusion scores. To correct increases of type-I error caused by multiple comparisons, a more conservative P value of .01 replaced the traditional .05 level of significance. Quantitative data are expressed as the mean ± SD. For categorical data, relative frequencies are given.
578
Galassi et al Accuracy of exercise testing for myocardial ischemia
Journal of Nuclear Cardiology November/December 2000
Table 1. Clinical variables in patients with an ischemic perfusion score of 0 (group 1), with an ischemic perfusion score of 1 or greater but less than 11 (group 2), and an ischemic perfusion score of 11 or greater (group 3)
Reversible perfusion defects Group 1 Patients 284 (35%) Age (y) 55 ± 10 Sex (men/women) 172/112 (61%/39%) Test referral (suspected 114/170 (40%/60%) CAD/known CAD) Family history of CAD 125/159 (44%/56%) (yes/no) Smoker (yes/no) 124/160 (44%/56%) Hypertension (yes/no) 144/140 (51%/49%) Diabetes mellitus (yes/no) 43/241 (15%/85%) Hyperlipidemia (yes/no) 134/150 (47%/53%) MI: Q-wave/ 50/23/211 (18%/8%/74%) non–Q-wave/absent
Group 2
Group 3
P value
457 (51%) 57 ± 10 388/69 (85%/15%) 51/406 (11%/89%)
75 (9%) 59 ± 10 68/7 (91%/9%) 6/69 (8%/92%)
<.0001 <.0001 <.0001
229/228 (50%/50%)
32/43 (43%/57%)
NS
254/203 (56%/44%) 195/262 (43%/57%) 90/367 (20%/80%) 210/247 (46%/54%) 229/62/166 (50%/14%/36%)
37/38 (49%/51%) NS 38/37 (51%/49%) NS 23/52 (31%/69%) NS 36/39 (48%/52%) NS 31/16/28 (41%/21%/37%) <.0001
Data are presented as mean value ± SD or number (%) of patients. MI, Myocardial infarction; CAD, coronary artery disease; NS, not significant.
RESULTS There were 614 maximal tests (75%) and 202 submaximal tests (25%). Exercise tests were stopped because of ST-segment depression 3 mm or greater in only 8 cases (1%). An exercise ischemic threshold was achieved in 326 cases (40%). Prediction of Impaired Myocardial Perfusion According to discriminant analysis, differences among patients were maximized when they were allocated in 3 groups according to ischemic perfusion score: patients without reversible perfusion defects (group 1), patients with an ischemic perfusion score of 1 or higher but less than 11 (group 2), and patients with an ischemic perfusion score of 11 or higher (group 3). This classification was confirmed on the basis of previous reports, which suggested that the severity of impaired myocardial perfusion by means of scintigraphy is a means of stratifying patients with different prognoses.38 An exercise ischemic threshold was achieved in 27 of 284 cases of patients in group 1 (10%), in 231 of 457 cases of patients in group 2 (51%), and in 68 of 75 cases of patients in group 3 (91%; P < .0001). Tables 1 and 2 summarize all electrocardiographic parameters measured in the 3 groups of patients and P values for differences at univariate analysis. Variables that were altogether significant (P < .001) by means of the U statistic for discrimination were hier-
archically moved by means of this stepwise method: maximal amount of ST-segment depression, presence and type of myocardial infarction, rate-pressure product threshold criteria, slope of ST-segment depression, sex, angina, ST/HR index, peak exercise HR. Exercise variables were also combined into a new global exercise index according to this formula: [Rate-pressure product threshold criteria / 100 + Maximal amount of ST segment depression / Peak exercise heart rate + (1 / ST / HR index)]1/2 in which if ST/HR index equals 0, then 1/ST/HR index should equal 0.01, to discriminate patients in the 3 groups according to the ischemic perfusion score (Figure 1A and B). Discriminant analysis was then repeated and, by using the jackknife classification to validate the classification functions for the 3 groups of patients, provided a sensitivity rate of 75% (214/284), 50% (229/457), and 77% (58/75), respectively, and a specificity rate of 80% (425/532), 78% (281/359), and 82% (611/741), respectively. According to this method, among the misclassified patients of group 1, 65 (93%) were allocated in group 2, and the remaining 5 patients (7%) were allocated in group 3. Similarly, among the misclassified patients of group 3, 13 (76%) were allocated in group 2, and the remaining 4 patients (24%) were allocated in group 1. Conversely, among misclassified patients of group 2, 103 (45%) were allocated in group 1, and 125 (55%) were allocated in group 3.
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
Galassi et al Accuracy of exercise testing for myocardial ischemia
A
B Figure 1. A, Diagnostic clinical and exercise algorithm in male patients. B, Diagnostic clinical and exercise algorithm in female patients. MI, Myocardial infarction; group 1, patients without impaired myocardial perfusion; group 2, patients with mild to moderate impaired myocardial perfusion; group 3, patients with severe impaired myocardial perfusion; global exercise index, [rate-pressure product threshold criteria/100 + maximal amount of ST-segment depression/peak exercise HR + (1/ST/HR index)*]1/2, (*when ST/HR index = 0, then 1/ST/HR index = 0.01).
579
580
Galassi et al Accuracy of exercise testing for myocardial ischemia
Journal of Nuclear Cardiology November/December 2000
Table 2. Exercise variables in patients with an ischemic perfusion score of 0 (group 1), with an ischemic perfusion score of 1 or greater but less than 11 (group 2) and an ischemic perfusion score of 11 or greater (group 3)
Reversible perfusion defects
Maximal ST ↓(mm) ST/HR index (µV/beats/min) Rest HR (beats/min) Rate-pressure product (beats/min × mm Hg) Peak exercise HR (beats/min) Rate-pressure product (beats/min × mm Hg) Time (min) Threshold criteria HR (beats/min) Rate-pressure product (beats/min × mm Hg) Time (min) Maximal HR (%) Exercise angina (no/yes) Slope of ST↓ (↑/non-↑/negative) Location of ischemia (ant/lat/inf/ant+inf/ ant+lat/inf+lat)
Group 1
Group 2
0.2 ± 0.5 0.4 ± 0.9
1.0 ± 0.9 1.8 ± 2.1
Group 3 1.9 ± 0.9 4.7 ± 4.7
P value <.0001 <.0001
82 ± 14 11.362 ± 2.588
78 ± 14 10.777 ± 2.344
78 ± 13 10.806 ± 2.096
<.003 <.003
150 ± 15 27.022 ± 3.473
141 ± 17 25.325 ± 3.789
131 ± 19 22.644 ± 4.089
<.0001 <.0001
16 ± 3
15 ± 3
14 ± 5
<.0001
149 ± 15 26.725 ± 3.483
135 ± 19 23.506 ± 4.339
115 ± 17 18.453 ± 3.738
<.0001 <.0001
16 ± 3 95 ± 9 265/19 (93%/7%) 69/23/192 (24%/8%/68%) —
14 ± 4 88 ± 10 306/151 (67%/33%) 157/169/131 (34%/37%/29%) 156/97/175/19/7/3 (34%/21%/38%/4%/2%/1%)
10 ± 4 82 ± 10 21/54 (28%/72%) 12/59/4 (16%/79%/5%) 34/13/10/11/7/0 (45%/17%/13%/15%/10%/0%)
<.0001 <.0001 <.0001 <.0001 <.0001
Data are presented as mean value ± SD or number (%) of patients. ST ↓ , ST segment depression; ↑, up-sloping; non-↑, non-up-sloping; ant, anterior; lat, lateral; inf, inferior; NS, not significant.
An allocation tree was finally built up by using significant clinical and exercise variables to discriminate among patients in the 3 groups. According to each combination of clinical variables (sex and presence and type of myocardial infarction) and exercise variables (maximal amount of ST-segment depression, rate-pressure product threshold criteria, slope of ST-segment depression, angina, ST/HR index, peak exercise HR), cutoff limits of the global exercise index for allocation into each of the 3 groups were given (Figure 1A and B). We also performed the same analysis including clinical variables only. The number of correctly classified patients was 216 of 284 (76%) in group 1, 229 of 457 (50%) in group 2, and 16 of 75 (21%) in group 3.
DISCUSSION This study demonstrates that clinical and electrocardiographic exercise variables may satisfactorily discrim-
inate patients with no reversible perfusion defects or with severe impaired myocardial perfusion at scintigraphy, but is far less satisfactory in those patients with mild to moderate perfusion defects. We assessed whether easily measured exercise testing variables might be able to predict the severity of myocardial ischemia. Although coronary angiography has traditionally been used as the reference standard for the evaluation of coronary artery disease, myocardial perfusion is an integrated response of an anatomic-hemodynamic system in which the coronary stenosis is only one component.25,26 Coronary angiography may not accurately indicate the physiologic significance of a coronary artery stenosis and may generally suffer from the inherent methodological verification bias. Only patients with noninvasive positive test results are more likely to have the results verified by means of coronary angiography, whereas patients with negative exercise test results are rarely referred for subsequent invasive studies.39 This
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
increases sensitivity, because false-negative results are unlikely to be discovered, and decreases specificity. To overcome such limitations, we compared exercise testing variables with myocardial SPECT. Although some technical disadvantages, such as variable tracer myocardial extraction fraction at different flow rates and lack of attenuation correction factors, may prevent this technique from being a quantitative tool for measuring myocardial perfusion, myocardial imaging with Tc-99m–labeled compound with SPECT provides a sensitive means for detecting exercise-induced regional hypoperfusion and can be easily used as a means of evaluating all patients with suspected or known coronary artery disease.27-30 In the present study, among all clinical and exercise variables tested, multivariate discriminant analysis identified 8 independent predictors for the discrimination of patients without impaired myocardial perfusion and with different severity of impaired myocardial perfusion. By using a diagnostic algorithm that included clinical and exercise variables, we were able to discriminate patients with no reversible perfusion defects and patients with severe impaired myocardial perfusion with a sensitivity rate of 75% and 77%, respectively. However, a less satisfactory sensitivity rate of 50% was obtained for the identification of patients with mild to moderate impairment of myocardial perfusion. Among all electrocardiographic variables, maximal amount of ST-segment depression, rate-pressure product threshold criteria, ST/HR index, and peak exercise HR were significant. These results are concordant with those obtained by other authors5,6 and confirm the usefulness of ST-segment analysis for the discrimination between patients with severe disease and healthy patients. Furthermore, the rate-pressure product that best correlates with myocardial oxygen consumption represents an index of myocardial metabolic demand.1820 It has been used by different authors as an objective and accurate estimate of the severity of myocardial ischemia. Waters et al showed that considerable variability of exercise tolerance occurs even when ratepressure product at the ischemic threshold remains constant.22 Similarly, Kaski et al21 have shown that a higher rate-pressure product at the ischemic threshold after sublingual nitrates is a means of identifying those patients in whom the prevailing mechanism of action is an increase in coronary flow reserve. Although myocardial wall tension and contractility should also have been taken into account, because they might contribute to myocardial oxygen consumption, they cannot be directly measured during exercise testing and are not even indirectly measured by means of rate-pressure product. Recently, computer-based processing of ST-segment depression such as ST/HR index and ST/HR slope has
Galassi et al Accuracy of exercise testing for myocardial ischemia
581
been compared with standard ST-segment analysis to improve method sensitivity. Various authors have shown an improved sensitivity by using the ST/HR index,11-14 whereas other authors have not.15-17,40,41 The degree of increase in accuracy of ST/HR index may depend on the population being studied42 and on referral bias for coronary angiography.43 Indeed, our study showed that the ST/HR index may be of value when applied to an unselected population. As compared with our method, the added value of myocardial scintigraphy is as large as 24.6% of patients with false-positive results, in whom a perfusion defect would have been diagnosed. Conversely, the added value of myocardial scintigraphy is as great as 50% for those patients with mild to moderate perfusion defects, in whom a perfusion defect would not be diagnosed (22.5%) or in whom a severe perfusion defect would be diagnosed (27.3%). For those patients with a severe perfusion defect, myocardial scintigraphy proved to have the lowest added value of 22.7%, because only 5.3% of patients would have been found to not have any perfusion defects, whereas the remaining 17.3% would have been found to have mild to moderate perfusion defects. Conversely, the added value of our method, as compared with clinical variables alone, was highest for those patients with severe perfusion defects, in whom the added value was as great as 56%. Thus these findings suggest that, as a first investigation, clinical variables and exercise testing may be a reasonable method for discriminating patients with no reversible perfusion defects and with severe impaired myocardial perfusion, but it is of little value in identifying patients with mild to moderate myocardial perfusion abnormalities. Limitations of the Study Among all exercise variables measured, we used the rate-pressure product threshold criterion. This electrocardiographic index results from the rate-pressure product measured at the ischemic threshold in the case of positive exercise test results and from rate-pressure product value measured at peak exercise in the case of negative test results. Although it represents a measurement that differs among patients, it allows for the discrimination of the severity of myocardial hypoperfusion in all exercise tests performed. Our study did not take into account the ST/HR slope, a method that is much more cumbersome to measure and is only marginally more accurate than the ST/HR index.11-13 Furthermore, given the reported differences in the extraction between Tc-tetrofosmin, Tc-sestamibi, and thallium-201 imaging, findings obtained with tetrofosmin might be marginally
582
Galassi et al Accuracy of exercise testing for myocardial ischemia
different than those obtained with other tracers, especially when using pharmacologic coronary vasodilators. Conclusions The exercise test continues to play an important role in the assessment of patients with ischemic heart disease and, in many cases, continues to represent the first investigation for those patients who can exercise maximally and in whom exercise is interpretable. However, subset selection of patients will continue to be a problem if the intention is to identify a particular approach that will have universal applicability. We thank Iain Halliday, MA, for constructive comments and editorial assistance.
References 1. Ellestad MH. Stress testing: Principles and practice. New York: Davis; 1996. p. 111-38. 2. Bruce RA, Hossack KF, Rouen TA, Hofer V. Enhanced risk assessment for primary coronary heart disease events by maximal exercise testing: 10 years’ experience of Seattle Heart Watch. J Am Coll Cardiol 1983;2:565-73. 3. Detrano R, Gianrossi R, Froelicher V. The diagnostic accuracy of the exercise electrocardiogram: A meta-analysis of 22 years of research. Prog Cardiovasc Dis 1989;3:173-206. 4. Martin CM, McConahay DR. Maximal treadmill exercise electrocardiography: Correlations with coronary arteriography and cardiac hemodynamics. Circulation 1972;26:956-62. 5. Weiner DA, McCabe CH, Ryan TJ. Identification of patients with left main and three vessel coronary disease with clinical and exercise test variables. Am J Cardiol 1980;46:21-7. 6. Goldschlanger N, Selzer A, Cohn K. Treadmill stress tests as indicators of presence and presence and severity of coronary artery disease. Ann Int Med 1976;85:277-86. 7. Podrid PJ, Graboys TB, Lown B. Prognosis of medically treated patients with coronary artery disease with profound ST-segment depression during exercise testing. N Engl J Med 1981;305:1111-6. 8. Borer JS, Brensike JF, Redwood DR, et al. Limitations of electrocardiographic response to exercise in predicting coronary artery disease. N Eng J Med 1975;293:367-71. 9. Weiner DA, Ryan TJ, McCabe CH, et al. Exercise stress testing: Correlations among history of angina, ST segment response and prevalence of coronary artery disease in the Coronary Artery Surgery Study (CASS). N Engl J Med 1979;301:230-5. 10. Bogaty P, Guimond J, Robitaille NM, et al. A reappraisal of exercise electrocardiographic indexes of the severity of ischemic heart disease: Angiographic and scintigraphic correlates. J Am Coll Cardiol 1997;29:1497-504. 11. Kligfield P, Ameisen O, Okin PM. Heart rate adjustment of ST segment depression for improved detection of coronary artery disease. Circulation 1989;79:245-55. 12. Okin PM, Kligfield P. Heart rate adjustment of segment depression and performance of the exercise electrocardiogram: A critical evaluation. J Am Coll Cardiol 1995;25:1726-35. 13. Okin PM, Lauer MS, Kligfield P. Chronotropic response to exercise. Improved performance of ST-segment depression criteria after adjustment for heart rate reserve. Circulation 1996;94:3226-31.
Journal of Nuclear Cardiology November/December 2000
14. Elamin MS, Boyle R, Kardash MM, et al. Accurate detection of coronary heart disease by new exercise test. Br Heart J 1982;48: 311-20. 15. Lachterman B, Lehmann KG, Detrano R, Neutel J, Froelicher VF. Comparison of ST segment/heart rate index to standard ST criteria for analysis of exercise electrocardiogram. Circulation 1990;82:44-50. 16. Bobbio M, Detrano R, Schmid J-J, et al. Exercise-induced ST depression and ST/heart rate index to predict triple-vessel or left main coronary disease: A multicenter analysis. J Am Coll Cardiol 1992; 19:11-8. 17. Morise AP, Duval RD. Accuracy of ST/heart rate index in the diagnosis of coronary artery disease. Am J Cardiol 1992;69:603-6. 18. Gobel FL, Nordstrom LA, Nelson RR, Jorgensen CR, Wang Y. The ratepressure product as an index of myocardial oxygen consumption during exercise in patients with angina pectoris. Circulation 1978;57: 549-56. 19. Kitamura K, Joregensen CR, Gobel FL, Taylor HL, Wang Y. Hemodynamic correlates of myocardial oxygen consumption during upright exercise. J Appl Physiol 1972;32:516-20. 20. Nelson RR, Gobel FL, Jorgensen CR, Wang Y, Wang Y, Taylor HL. Hemodynamic predictors of myocardial oxygen consumption during static and dynamic exercise. Circulation 1974;50:1179-84. 21. Kaski JC, Plaza LR, Meran DO, Araujo L, Chierchia S, Maseri A. Improved coronary supply: Prevailing mechanism of action of nitrates in chronic stable angina. Am Heart J 1985;110:238-45. 22. Waters DD, McCans JL, Crean PA. Serial exercise testing in patients effort angina: Variable tolerance, fixed threshold. J Am Coll Cardiol 1985;6:1011-5. 23. Panza JA, Arshed AQ, Diodati JG, Callahan TS, Bonow RO, Epstein SE. Long-term variation in myocardial ischemia during daily life in patients with stable coronary artery disease: Its relation to changes in the ischemic threshold. J Am Coll Cardiol 1992;19:500-6. 24. Garber CE, Carleton RA, Camaione DN, Heller GV. The threshold for myocardial ischemia varies in patients with coronary artery disease depending on the exercise protocol. J Am Coll Cardiol 1991;17: 1256-62. 25. Wilson RF, Marcus ML, Christensen BV, Talman C, White CW. Accuracy of exercise electrocardiography in detecting physiologically significant coronary arterial lesions. Circulation 1991;83:412-21. 26. Folland ED, Vogel RA, Hartigan P, et al. Relation between coronary artery stenosis assessed by visual, caliper, and computer methods and exercise capacity in patients with single-vessel coronary artery disease. Circulation 1994;89:2005-14. 27. Gould KL. Identifying and measuring severity of coronary artery stenosis: Quantitative coronary arteriography and positron emission tomography. Circulation 1988;78:237-45. 28. Di Carli M, Czernin J, Hoh CK, et al. Relation among stenosis severity, myocardial blood flow, and flow reserve in patients with coronary artery disease. Circulation 1995;91:1944-51. 29. Kahn JK, McGhie I, Akers MS, et al. Quantitative rotational tomography with 201Tl and 99mTc 2-methoxy-isobutyl-isonitrile; a direct comparison in normal individuals and patients with coronary artery disease. Circulation 1989;79:1282-93. 30. Zaret BL, Rigo P, Wackers FJT, et al. Myocardial perfusion imaging with 99mTc Tetrofosmin: Comparison to 201Tl imaging and coronary angiography in a phase III multicenter trial. Circulation 1995;91:313-9. 31. Tamaki N, Yonekura Y, Mukai T, et al. Stress thallium-201 transaxial emission computed tomography: Quantitative versus qualitative analysis for evaluation of coronary artery disease. J Am Coll Cardiol 1984;4:1213-20. 32. DePasquale EE, Nody AC, DePuey EG, et al. Quantitative rotational thallium-201 tomography for identifying and localizing coronary artery disease. Circulation 1988;77:316-21.
Journal of Nuclear Cardiology Volume 7, Number 6;575-83
Galassi et al Accuracy of exercise testing for myocardial ischemia
33. Gibbons RJ, Balady GJ, Beasley JW et al. ACC/AHA Guidelines for exercise testing: A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee on exercise testing). J Am Coll Cardiol 1997;30(1):260-315. 34. Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake and nomografic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 1973;85:546-62. 35. Higley B, Smith FW, Smith T, et al. Technetium-99m-1, 2bis [bis(2ethoxyethyl) phosphino] ethane: Human biodistribution, dosimetry and safety of a new myocardial perfusion imaging agent. J Nucl Med 1993;34:30-8. 36. Garcia EV, Van Train K, Maddahi J, et al. Quantification of rotational thallium-201 myocardial tomography. J Nucl Med 1985;26:17-26. 37. Chouraqui P, Maddahi J, Ostrzega E, et al. Quantitative exercise thallium-201 rotational tomography for evaluation of patients with prior myocardial infarction. Am J Cardiol 1990;66:151-7. 38. Hachamovitch R, Berman DS, Shaw LJ, et al. Incremental prognostic value of myocardial perfusion single photon emission computed
O
39. 40.
41.
42.
43.
583
tomography for the prediction of cardiac death: Differential stratification for risk of cardiac death and myocardial infarction. Circulation 1998;97:535-43. Detrano R, Froelicher VF. Exercise testing: Uses and limitation considering recent studies. Prog Cardiovasc Dis 1988;31:173-204. Okin PM, Kligfield P, Ameisen O, Goldberg HL, Borer JS. Improved accuracy of exercise electrocardiogram: Identification of three-vessel coronary disease in stable angina pectoris by analysis of peak raterelated changes in ST segments. Am J Cardiol 1985;55:271-6. Bobbio M, Detrano R. A lesson from the controversy about heart rate adjustment of ST segment depression. Circulation 1991;84: 1410-3. Okin PM, Kligfield P. Population selection and performance of the exercise ECG for the identification of coronary artery disease. Am Heart J 1994;127:296-304. Morise A, Duval R. Diagnostic accuracy of heart rate-adjusted ST segment compared with standard ST-segment criteria. Am J Cardiol 1995;75:118-21.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS:
NEW ADDRESS:
Affix the address label from a recent issue of the journal here.
Clearly print your new address here. Name Address City / State / ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 1-800-654-2452 Outside the U.S., call 407-345-4000