Exercise Echocardiography as a Screening lest for Coronary Artery Disease and Correlation with Coronary Arteriography Linda J. Crouse, MD, James J. Harbrecht, MD, James L. Vacek, MD, Thomas L. Rosamond, MD, and Paul H. Kramer, MD
We evaluated exercise echocardiography as a screening test for coronary artery disease in 228 patients, all of whom underwent subsequent coronary angiography. After an echocardiogram at rest was obtained, each patient performed maximal, symptom-limited, upright treadmill exercise, immediately after which repeat imaging was performed. The exercise echocardiogram was abnormal if any segment failed to become hypercontractile with exercise, and these regional wall motion abnormalities were used to predict the extent and distribution of coronary disease. At subsequent angiography, coronary stenosis was defined as significant if luminal diameter was reduced 150%. Compared with electrocardiography, exercise echocardiography was more sensitive (97 vs 51%) and specific (64 vs 62%), and had higher positive (90 vs 62%) and negative (87 vs 26%) predictive accuracies. Exercise echocardiography was also highly predictive of the extent (no, 1-, 2- or 3-vessel disease) and distribution (which vessel) of coronary stenoses. It is concluded that exercise echocardiography is an excellent screening test for the presence, extent and distribution of coronary artery disease. (Am J Cardiol 1991;67:1213-1218)
From the Mid America Heart Institute, St. Luke’s Hospital of Kansas City, Kansas City, Missouri. Manuscript received December 3, 1990; revised manuscript received and accepted February 13,199 1. Address for reprints: Linda J. Crouse, MD, Suite 40-II,4320 Wornall Road, Kansas City, Missouri 64111.
oronary artery disease (CAD) is the leading cause of death in adults in the United States,’ and its detection is central to the application of a variety of primary and secondary interventions intended to reduce morbidity and mortality in afflicted patients. Stress electrocardiography has gained wide acceptance as a simple, inexpensive and widely available screening test for CAD, but has proved to be of limited reliability in many patient groups.2 The ability to detect coronary stenosis was strengthened by supplementing stress electrocardiography with imaging modalities, namely thallium scintigraphy3 and technetium radioventriculography,4 which can depict stress-related changes in regional myocardial perfusion or contraction. Echocardiography is an excellent method for evaluating regional wall motion and systolic thickening, but early efforts to couple this imaging method with exercise testing were hampered by technical obstacles resulting from exertional hyperpnea.5$6 The application of digital technology to echocardiographic imaging has largely overcome these impediments. This study evaluates exercise echocardiography as a screening test for CAD.
C
METHODS
The study population consisted of all patients with no previous history of heart disease who were referred during the study period for evaluation of possible CAD and underwent exercise echocardiography and subsequent cardiac catheterization. Two hundred twentyeight patients (153 men and 75 women, age range 33 to 84 years [mean 621) were enrolled. Before exercise, with patients in the left lateral decubitus position, images at rest were sequentially acquired in the parasternal short- and long-axis, and apical 2- and 4-chamber views. Images were recorded on M-inch videotape and were also digitized on-line and stored on floppy disk using a Microsonics Prevue system. This system acquires and digitizes 8 serial echocardiographic images at 50-ms intervals during systole of a single cardiac cycle, triggered by the electrocardiogram. The images can then be displayed in a continually recycling, tine-loop format. EXERCISE
ECHOCARDIOGRAPHY
1213
Patients then performed maximal, symptom-limited upright treadmill exercise, using the Bruce or modified Bruce protocol. Immediately after cessation of exercise, patients resumed the left lateral decubitus position, and imaging was repeated in the 4 views described, and recorded on videotape and floppy disk as rapidly as possible (completion within 60 seconds of exercise termination). All exercise echocardiograms were interpreted by a single observer who was unaware of the clinical history or the exercise electrocardiogram result. Regional wall motion was described for each segment in each view: anterior, anteroseptal, apical, lateral, posterior and inferior (Figure 1). To predict the extent of CAD (l-, 2- or 3-vessel disease), segments were then consolidated into 3 zones, namely anterior, lateral and inferoposterior. Wall motion was described as normal, hypokinetic, akinetic or hypercontractile. Segmental response was termed normal only if the segment became hypercontractile after exercise, regardless of resting function. Any segment failing to do so, even if it improved somewhat after exercise, was classified as abnormal. Description of regional wall motion was based on review of the digitized images, but this interpretation was supplemented by reviews of videotaped images. To predict the territorial distribution of CAD, abnormal regional wall motion responses in the apical, anteroseptal and anterior segments were attributed to disease of the left anterior descending artery. Lateral wall abnormalities were used to predict left circumflex stenosis. Because attribution of inferoposterior circulation depends on coronary dominance, exertional contractile failure in those segments was interpreted as evidence of stenosis of the right or left circumflex arteries, or both. No patient was excluded from this series because of technical difficulties (e.g., suboptimal image quality) or because of medications-that might interfere with exer-
cise tolerance, such as ,8 blockers. The exercise electrocardiogram was evaluated according to standard criteria.7 All patients underwent coronary angiography within 3 months after exercise echocardiography (generally within 3 weeks). Angiograms were interpreted by experienced angiographers who were unaware of, and did not participate in, the stress test interpretations. Coronary stenoses were assessedqualitatively and graded in severity according to estimated percent diameter reduction. A stenosis was considered hemodynamically significant if it produced 250% diameter reduction. With use of the coronary angiogram as the “gold standard,” the sensitivity, specificity and predictive accuracy (for our patient population) of exercise echocardiography were derived and compared with exercise electrocardiography. Statistical analysis was performed with the chi-square test. RESULTS At coronary angiography, 175 patients had “significant” stenosis, as previously defined, in at least 1 coronary artery. Fifty-three patients did not meet this criterion and were considered to be “normal.” Of the 175 patients with CAD, 170 had abnormal exercise echocardiograms (overall sensitivity 97%). Among the 53 patients comprising the angiographically normal group, exercise echocardiography was normal in 34 (overall specificity 64%). Fifteen of the 19 patients with normal coronary anatomy but abnormal exercise echocardiograms had resting regional wall motion abnormalities on echocardiography and ventriculography. For the entire group of 228 patients, the positive predictive accuracy was 90% and the negative predictive accuracy was 87%. An example of an exercise echocardiogram from a patient with normal coronary anatomy is shown in Figure 2.
Anteroseptal
nterior Posterolateral Parasternal Long Axis
Parasiernal Short Axis
lE# Anteriol
Anterior
1 1
Apical 4-Chamber
1214
THE AMERICAN
2-
JOURNAL
OF CARDIOLOGY
Posterolateral
Apical’ Chamber
VOLUME
El Inferior
67
JUNE
1, 1991
FIGURE 1. Regional nomenclature in the 4 echocardiographic views. Shading indicates consolidation oi segments into 3 distributions (anterior, posterolateral and inferior) for correlation with angiographic anatomy.
Beyond its ability to detect the presence of CAD, exercise echocardiography correlated closely with the distribution (specific vessel[s]) and extent (number of arteries) of disease. This correlation is summarized in Table I. Among the 175 patients with angiographically defined significant CAD, there were only 5 false-negative test results. CAD was confined to a diagonal branch in 3 of these patients. In 1, a 60% distal left circumflex stenosis was present. In the fifth patient, a high-grade stenosis was present in the apical segment of the left anterior descending coronary artery. Figure 3 is an example of an exercise echocardiogram in a patient with an isolated severe left anterior descending stenosis. Figure 4 is an example of an exercise echocardiogram in a patient with 3-vessel CAD. Exercise electrocardiography correlated more poorly with angiography. Of 175 patients with CAD, 90 had stress tests (sensitivity = 51%). In 53 patients without CAD, the routine treadmill test was normal in 33 and abnormal in 20 (specificity = 62%). This test’s positive predictive accuracy was 82%, whereas the negative predictive accuracy was only 28%. The sensitivity, specificity and predictive value of exercise echocardiography and electrocardiography are summarized and compared in Table II. The relative
FIGURE IURE 2. Exercise echocardiogram echocardior from a patient with noroatiit mal coronary anatomy showing improved contraction of all segments with exercise. All images are from end-systote. Rq+t images are on the /err; paired postexercise images are on the right A,B = parastemal long axis; C,D = parasternal short axis: E,F = apical 4 chamber; G,H = apical 2 chamber.
TABLE I Correlation of Extent of Coronary Distribution with Exercise Echocardiography
Disease
and
Exercise Echocardiogram Coronary
Disease
Distribution
by Angiography
l-vessel
disease (n = 66) 2-vessel disease (n = 65) 3-vessel disease (n = 41)
Normal
Abnormal
5
61”
0 0
65+ 41%
* Abnormal vessel correctly predicted in each case. + Both vessels predicted in 52 patients, 1 vessel in 13. *All vessels predicted in 32 patlents. 2 vessels in 9.
risk ratio indicates the relative likelihood of an incorrect prediction based on the electrocardiographic stress test versus that based on the echocardiographic stress test. DISCUSSION Identification of patients with CAD is the first step in the selective application of measures that are intended to prevent or reduce coronary morbidity and mortality. Although coronary stenosis may be manifest in patients at rest, most screening tests for coronary artery disease seek to detect the consequences of provoked
FIGURE 3. Exercise echocardiogram from a patient with isolated severe midstenosis of the left anterior descending artery, before (upper quad screen) and after (lower quad screen) bypass surgery. Only the apical views are shown. Rest images are on the lert; postexercise images are on the righf in the sameviews. A,B,E,F=apical4chamber; C,B,G,H=apicai2 chamber. The preoperative study demonstrates exertional aldnesia of the anteroapical segments (arrows). The postoperative study, by comparison, is normal. EXERCISE
ECHOCARDIOGRAPHY
1215
TABLE
II Exercise
Test Comparison 95%
Exercise
Confidence Interval
Test
Echo ECG
Sensitivity 97% 51% Specificity
Echo ECG
64% 62%
Echo ECG
(95, m-3 (4% 59)
(51,77) (4475)
Positive predictive accuracy 90%
82%
@SW (75.89)
p Value
Relative Ratio
We evaluated exercise echocardiography as a screening test for coronary artery disease in 228 patients, all of whom underwent subsequent coronary angiography. After an echocardiogram at rest was obtained, each patient performed maximal, symptom-limited, upright treadmill exercise, immediately after which repeat imaging was performed. The exercise echocardiogram was abnormal if any segment failed to become hypercontractile with exercise, and these regional wall motion abnormalities were used to predict the extent and distribution of coronary disease. At subsequent angiography, coronary stenosis was defined as significant if luminal diameter was reduced 150%. Compared with electrocardiography, exercise echocardiography was more sensitive (97 vs 51%) and specific (64 vs 62%), and had higher positive (90 vs 62%) and negative (87 vs 26%) predictive accuracies. Exercise echocardiography was also highly predictive of the extent (no, 1-, 2- or 3-vessel disease) and distribution (which vessel) of coronary stenoses. It is concluded that exercise echocardiography is an excellent screening test for the presence, extent and distribution of coronary artery disease. (Am J Cardiol 1991;67:1213-1218)
From the Mid America Heart Institute, St. Luke’s Hospital of Kansas City, Kansas City, Missouri. Manuscript received December 3, 1990; revised manuscript received and accepted February 13,199 1. Address for reprints: Linda J. Crouse, MD, Suite 40-II,4320 Wornall Road, Kansas City, Missouri 64111.
oronary artery disease (CAD) is the leading cause of death in adults in the United States,’ and its detection is central to the application of a variety of primary and secondary interventions intended to reduce morbidity and mortality in afflicted patients. Stress electrocardiography has gained wide acceptance as a simple, inexpensive and widely available screening test for CAD, but has proved to be of limited reliability in many patient groups.2 The ability to detect coronary stenosis was strengthened by supplementing stress electrocardiography with imaging modalities, namely thallium scintigraphy3 and technetium radioventriculography,4 which can depict stress-related changes in regional myocardial perfusion or contraction. Echocardiography is an excellent method for evaluating regional wall motion and systolic thickening, but early efforts to couple this imaging method with exercise testing were hampered by technical obstacles resulting from exertional hyperpnea.5$6 The application of digital technology to echocardiographic imaging has largely overcome these impediments. This study evaluates exercise echocardiography as a screening test for CAD.
C
METHODS
The study population consisted of all patients with no previous history of heart disease who were referred during the study period for evaluation of possible CAD and underwent exercise echocardiography and subsequent cardiac catheterization. Two hundred twentyeight patients (153 men and 75 women, age range 33 to 84 years [mean 621) were enrolled. Before exercise, with patients in the left lateral decubitus position, images at rest were sequentially acquired in the parasternal short- and long-axis, and apical 2- and 4-chamber views. Images were recorded on M-inch videotape and were also digitized on-line and stored on floppy disk using a Microsonics Prevue system. This system acquires and digitizes 8 serial echocardiographic images at 50-ms intervals during systole of a single cardiac cycle, triggered by the electrocardiogram. The images can then be displayed in a continually recycling, tine-loop format. EXERCISE
ECHOCARDIOGRAPHY
1213
Patients then performed maximal, symptom-limited upright treadmill exercise, using the Bruce or modified Bruce protocol. Immediately after cessation of exercise, patients resumed the left lateral decubitus position, and imaging was repeated in the 4 views described, and recorded on videotape and floppy disk as rapidly as possible (completion within 60 seconds of exercise termination). All exercise echocardiograms were interpreted by a single observer who was unaware of the clinical history or the exercise electrocardiogram result. Regional wall motion was described for each segment in each view: anterior, anteroseptal, apical, lateral, posterior and inferior (Figure 1). To predict the extent of CAD (l-, 2- or 3-vessel disease), segments were then consolidated into 3 zones, namely anterior, lateral and inferoposterior. Wall motion was described as normal, hypokinetic, akinetic or hypercontractile. Segmental response was termed normal only if the segment became hypercontractile after exercise, regardless of resting function. Any segment failing to do so, even if it improved somewhat after exercise, was classified as abnormal. Description of regional wall motion was based on review of the digitized images, but this interpretation was supplemented by reviews of videotaped images. To predict the territorial distribution of CAD, abnormal regional wall motion responses in the apical, anteroseptal and anterior segments were attributed to disease of the left anterior descending artery. Lateral wall abnormalities were used to predict left circumflex stenosis. Because attribution of inferoposterior circulation depends on coronary dominance, exertional contractile failure in those segments was interpreted as evidence of stenosis of the right or left circumflex arteries, or both. No patient was excluded from this series because of technical difficulties (e.g., suboptimal image quality) or because of medications-that might interfere with exer-
cise tolerance, such as ,8 blockers. The exercise electrocardiogram was evaluated according to standard criteria.7 All patients underwent coronary angiography within 3 months after exercise echocardiography (generally within 3 weeks). Angiograms were interpreted by experienced angiographers who were unaware of, and did not participate in, the stress test interpretations. Coronary stenoses were assessedqualitatively and graded in severity according to estimated percent diameter reduction. A stenosis was considered hemodynamically significant if it produced 250% diameter reduction. With use of the coronary angiogram as the “gold standard,” the sensitivity, specificity and predictive accuracy (for our patient population) of exercise echocardiography were derived and compared with exercise electrocardiography. Statistical analysis was performed with the chi-square test. RESULTS At coronary angiography, 175 patients had “significant” stenosis, as previously defined, in at least 1 coronary artery. Fifty-three patients did not meet this criterion and were considered to be “normal.” Of the 175 patients with CAD, 170 had abnormal exercise echocardiograms (overall sensitivity 97%). Among the 53 patients comprising the angiographically normal group, exercise echocardiography was normal in 34 (overall specificity 64%). Fifteen of the 19 patients with normal coronary anatomy but abnormal exercise echocardiograms had resting regional wall motion abnormalities on echocardiography and ventriculography. For the entire group of 228 patients, the positive predictive accuracy was 90% and the negative predictive accuracy was 87%. An example of an exercise echocardiogram from a patient with normal coronary anatomy is shown in Figure 2.
Anteroseptal
nterior Posterolateral Parasternal Long Axis
Parasiernal Short Axis
lE# Anteriol
Anterior
1 1
Apical 4-Chamber
1214
THE AMERICAN
2-
JOURNAL
OF CARDIOLOGY
Posterolateral
Apical’ Chamber
VOLUME
El Inferior
67
JUNE
1, 1991
FIGURE 1. Regional nomenclature in the 4 echocardiographic views. Shading indicates consolidation oi segments into 3 distributions (anterior, posterolateral and inferior) for correlation with angiographic anatomy.
Beyond its ability to detect the presence of CAD, exercise echocardiography correlated closely with the distribution (specific vessel[s]) and extent (number of arteries) of disease. This correlation is summarized in Table I. Among the 175 patients with angiographically defined significant CAD, there were only 5 false-negative test results. CAD was confined to a diagonal branch in 3 of these patients. In 1, a 60% distal left circumflex stenosis was present. In the fifth patient, a high-grade stenosis was present in the apical segment of the left anterior descending coronary artery. Figure 3 is an example of an exercise echocardiogram in a patient with an isolated severe left anterior descending stenosis. Figure 4 is an example of an exercise echocardiogram in a patient with 3-vessel CAD. Exercise electrocardiography correlated more poorly with angiography. Of 175 patients with CAD, 90 had stress tests (sensitivity = 51%). In 53 patients without CAD, the routine treadmill test was normal in 33 and abnormal in 20 (specificity = 62%). This test’s positive predictive accuracy was 82%, whereas the negative predictive accuracy was only 28%. The sensitivity, specificity and predictive value of exercise echocardiography and electrocardiography are summarized and compared in Table II. The relative
FIGURE IURE 2. Exercise echocardiogram echocardior from a patient with noroatiit mal coronary anatomy showing improved contraction of all segments with exercise. All images are from end-systote. Rq+t images are on the /err; paired postexercise images are on the right A,B = parastemal long axis; C,D = parasternal short axis: E,F = apical 4 chamber; G,H = apical 2 chamber.
TABLE I Correlation of Extent of Coronary Distribution with Exercise Echocardiography
Disease
and
Exercise Echocardiogram Coronary
Disease
Distribution
by Angiography
l-vessel
disease (n = 66) 2-vessel disease (n = 65) 3-vessel disease (n = 41)
Normal
Abnormal
5
61”
0 0
65+ 41%
* Abnormal vessel correctly predicted in each case. + Both vessels predicted in 52 patients, 1 vessel in 13. *All vessels predicted in 32 patlents. 2 vessels in 9.
risk ratio indicates the relative likelihood of an incorrect prediction based on the electrocardiographic stress test versus that based on the echocardiographic stress test. DISCUSSION Identification of patients with CAD is the first step in the selective application of measures that are intended to prevent or reduce coronary morbidity and mortality. Although coronary stenosis may be manifest in patients at rest, most screening tests for coronary artery disease seek to detect the consequences of provoked
FIGURE 3. Exercise echocardiogram from a patient with isolated severe midstenosis of the left anterior descending artery, before (upper quad screen) and after (lower quad screen) bypass surgery. Only the apical views are shown. Rest images are on the lert; postexercise images are on the righf in the sameviews. A,B,E,F=apical4chamber; C,B,G,H=apicai2 chamber. The preoperative study demonstrates exertional aldnesia of the anteroapical segments (arrows). The postoperative study, by comparison, is normal. EXERCISE
ECHOCARDIOGRAPHY
1215
TABLE
II Exercise
Test Comparison 95%
Exercise
Confidence Interval
Test
Echo ECG
Sensitivity 97% 51% Specificity
Echo ECG
64% 62%
Echo ECG
(95, m-3 (4% 59)
(51,77) (4475)
Positive predictive accuracy 90%
82%
@SW (75.89)
p Value
Relative Ratio
