Comparison of exercise perfusion and ventricular function imaging: An analysis of factors affecting the diagnostic accuracy of each technique

272

lACC Vol. 3, No.2 february 1984:272-83

Comparison of Exercise Perfusion and Ventricular Function Imaging: An Analysis of Factors Affecting the Diagnostic Accuracy of Each Technique MARY D. OSBAKKEN , MD, PhD,* ROBERT D. OKADA, MD, FACC, CHARLES A, BOUCHER, MD, FACC, H. WILLIAM STRAUSS, MD, FACC, GERALD M, POHOST, MD, FACC Boston, Massachusetts and Hershey, Pennsylvania

Exercise thallium-201 perfusion scans and gated equilibrium blood pool scans were performed in 120 catheterized patients with a chest pain syndrome. Eighty-six patients had coronary artery disease and 34 patients did not. The effects of gender, propranolol, exercise level, exercise ischemia, history of typical angina, history of previous myocardial infarction, electrocardiographic Q waves, number of diseased vesselsand extent of coronary artery obstruction on diagnostic accuracy were evaluated. The overall sensitivity and specificity of thallium scans were 76 and 68%, respectively, and those of gated blood pool scans 80 and 62% (p = not significant). Propranolol decreased the specificity of thallium scans (propranolol = 42%; no propranolol = 87%, P < 0.05). Thallium scans and anginal history were less sensitive for detecting coronary disease in women (men: thallium 79%; angina 77%; women: 54 and 46%, respectively; p < 0.05). Exercise level did not significantly affect the diagnostic accuracy of either scan. Thallium and gated scans were both highly sensitive (95%) in detecting disease in 20 patients with a prior myocardial infarction, angina and a positive electrocardiogram. The sensitivity of the thallium scan signifi-

=

=

At rest, both myocardial perfusion and function may be adequate even in patients with significant coronary artery obstruction. However, when the oxygen demand of myoFrom the Cardiac Unit and Division of Nuclear Medicine, Massachusetts General Hospital, Boston, Massachusetts and the Departments of Medicine and Radiology, * The Milton S. Hershey Medical Center, Pennsylvania State University, Hershey, Pennsylvania. Drs. Pohost, Okada and Osbakken are Established Investigators of the American Heart Association. This study was supported in part by U.S. Public Health Service Grants HL21751 and HL26215 and Cardiovascular Nuclear Training Grant I-T32 HL-07416 from the National Institutes of Health, Bethesda, Maryland. Manuscript received February 22, 1983; revised manuscript received August 16, 1983, accepted August 16. 1983. Address for reprints: Mary D. Osbakken. MD, Department of Medicine and Radiology, The Milton S. Hershey Medical Center, The Pennsylvania State University, P.O. Box 850, Hershey, Pennsylvania 17033. © 1984 by the American College of Cardiology

cantly decreased as the number of diseased vessels decreased. Both thallium and gated scans were lessfrequently positive in patients with atypical angina or no Q waves, but were not significantly influenced by electrocardiographic ischemia. The sensitivity and specificity of both scans were low in 57 patients with the combination of atypical angina, no history of infarction and equivocal stress electrocardiogram (thallium = 61 and 63%, re61 and 67%). spectively; gated When stress thallium scan evaluation included the electrocardiogram and thallium scan interpretation, the diagnostic accuracy was 81%. When all the information from gated scans (wall motion, ejection fraction, pulmonary blood volume) was combined for final gated scan evaluation, the diagnostic accuracy was 83 %. When electrocardiographic data were added to all three gated scan variables, diagnostic accuracy was 77%. In conclusion, thallium perfusion and gated blood pool scans have reasonable diagnostic accuracy for coronary artery disease in a group of patients with a moderately high prevalence of disease. However, combined variables from each test are needed to provide reliable diagnostic accuracy.

=

cardial cells is increased with exercise, patients with coronary artery disease cannot adequately increase oxygen supply (perfusion), and as a result myocardial function can be depressed. Two scintigraphic techniques that are used to detect coronary artery disease, the thallium perfusion scan (1-16) and the gated blood pool angiogram (17-23), rely on this close relation between perfusion and function in order to detect coronary artery disease (24-30). The purpose of the present investigation was to determine whether one of these studies might be a preferred approach for detecting coronary artery disease. Various exercise thallium scan (exercise electrocardiogram and thallium distribution characteristics) and gated scan (regional wall motion, ejection fraction and pulmonary blood volume) variables were eval0735-1097/84/$3.00

lACC Vol. 3, No.2 February t984:272-83

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOL SCANS

uated alone and in combination to determine their diagnostic accuracy. In addition, we compared scintigraphic diagnosis with other means of diagnosis (anginal history, rest and exercise electrocardiograms). We also evaluated the effects of a variety of clinical variables (sex, propranolol, exercise level, number and extent of diseased vessels) on the sensitivity, specificity and diagnostic accuracy of each type of scan.

Methods Patients. The 120 patients included in this study were selected randomly from a group of patients who underwent cardiac catheterization for a chest pain syndrome. These patients were studied within I week of cardiac catheterization. All patients gave informed consent. There were 86 patients with coronary artery disease defined as 50% or greater diameter narrowing in one or more major coronary vessels, and 34 patients without significant coronary artery disease (25 with < 20% and 9 with 20 to 49% diameter narrowing in any vessel). None of the patients had valvular heart disease, left main coronary artery disease, unstable angina or congestive heart failure. The patients without significant coronary artery disease had a left ventricular enddiastolic pressure less than 12 mm Hg. Thirty-three of 86 patients with coronary artery disease subsequently underwent coronary artery bypass graft surgery. Because we intended to determine the effect of propranolol on diagnostic accuracy, no attempt was made to alter this medication. Long-acting nitrates were withheld 4 hours before the scintigraphic studies. Cardiac catheterization. Each patient had routine coronary angiography and contrast ventriculography. Right- and left-sided pressures were measured. The presence and number of diseased vessels were determined by two cardiac radiologists by consensus and without knowledge of the scintigraphic findings.

Scintigraphic Image Acquisition Thallium perfusion imaging. Each patient was studied after a 12 hour fast within 1 week of cardiac catheterization. An intravenous line was inserted into an antecubital vein and the patient was placed in a supine position on a bicycle exercise table (Engineering Dynamic Corporation, Lowell, Massachusetts). This exercise mode was chosen to be identical to that used for gated blood pool imaging. Precordial electrocardiographic leads were placed for continuous monitoring. Exercise commenced at a work load of 150 kilopond meters (kpm)/min and was increased in 150 kpm/min increments every 3 minutes until peak tolerance (85% maximal predicted heart rate) or until the development of angina, fatigue, shortness of breath, ventricular irritability or 3 mm or more ST depression on the electrocardiogram. One minute before the anticipated cessation of exercise, 1.5 mCi of

273

thallium-20 I was injected intravenously. Within 5 minutes after completion of exercise, scanning was begun with an Ohio Nuclear series 400 gamma-camera (Technicare, Cleveland, Ohio) using a medium sensitivity parallel hole collimator. The pulse height analyzer was set at 87 keY with a 20% window. Images were stored on a Mugacart computer system (Medical Data Systems, Ann Arbor, Michigan) for subsequent analysis. Anterior and 45° left anterior oblique images were collected for 6 minutes each. Redistribution images were collected 2 to 3 hours after exercise in both the anterior and 45° left anterior oblique views, also for 6 minutes each. Gated blood pool imaging. Multigated blood pool scans were performed on the same day, 4 hours after completion of the thallium exercise protocol. Twenty minutes after injection of stannous pyrophosphate (Pyrolyte, New England Nuclear, North Billerica, Massachusetts), 20 mCi of technetium-99m pertechnetate were administered for in vivo labeling of the red blood cells. After 10 minutes of equilibration, supine rest imaging was performed in the anterior and left anterior oblique projection that optimized separation of the right and left ventricle, with a gamma-camera equipped with a medium sensitivity all purpose parallel hole collimator. The pulse height analyzer was set at 140 keY with a 20% window. Data acquisition was performed on a Mugacart cardiac computer system. A Brattle physiologic trigger was employed to synchronize digital data recording with the electrocardiogram. Each cardiac cycle was divided into 14 frames; approximately 150,000 to 200,000 counts were obtained in each frame (and were collected for a preset number of counts, which at rest required 5 to 10 minutes per view depending on heart rate). After the rest study, graded supine exercise was performed using a similar protocol for similar durations, work loads and end point (that is, to 85% maximal predicted heart rate or to symptom-limited maximum) as for the thallium perfusion exercise study. All patients attained similar heart rates and blood pressures for both exercise regimens (Table I). These heart rate and blood pressure responses occurred with similar work loads and at similar durations of exercise. Left anterior oblique images were recorded during each stage of exercise. Each image was collected for the last 2 minutes (150,000 counts/frame) at each exercise level. Thallium perfusion image analysis. The thallium images were reviewed by three independent observers unaware of the clinical, catheterization and gated scan data. In each view, the left ventricle was divided into three segments for qualitative analysis (anterior, apical and inferior for the anterior view; septal, apical-inferior and posterior for the 45° left anterior oblique view). An average score (from the three observers) for each segment was obtained for both the initial and delayed scans in a manner previously reported (31). = no disBriefly, the scoring system was as follows: cernible activity, + 1 = myocardial activity present but

°

274

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLI UM AND GATED BLOOD POOL SCANS

JACC Vol. 3. No. 2 February 1984:272- 83

Table 1. Hemodynamic Responses to Exercise in Patients With and Without Propranolol Therapy 86 Patients With Coronary Artery Disease Prop Age (yr) BP (mm Hg) Rest Ex HR (beals/min) Resl Ex % Pred HR Work load (kprn-s) Duration (min) EF (%) Rest Ex

No Prop

52 ± 0.14 122177 ± 4.6/2.5 155/87 ± 4. 6/36t 66 108 65. 1 504 8.3

52.5 ± 3.3 122173 ± 4.1/3.1 167/82 = 1.5/0.5*t

±1.7 3.4 t ± 1.3 ± 58 ± 1.0

::!:

62. 1 ± 1.3 59 .8 ± 1.7

75 123 77.5 477 7. 1

± 3.6* ± 5.2*t ::!: 1.3* ± 49 ::!: 0 .5

57.4 ::!: 3.6 57.4 ::!: 3.6

34 Patients Without Coronary Artery Disease Prop

No Prop

50 ± 3.2

48 ± 2.2

128/80 ± 0. 74/0.97 167/88 ± 5.8/4.0t 71 135 78.6 450 6 .9

± 3.6 ± 4.7t ::!: 3.3 ± 30 ± 0 .3

63.6 ± 1.9 70.8 ± 2.l t

131179 ± 3.8/ 1.9 164/89 ± 4.8 /3.4 t 73 136 81 538 7.8

± ± ± ± ±

3.2 I.2t 0 .7 57 0.8

64.9 ± 2.4 70.5 ± 2.5

Similar heart rate and blood pressure responses were obtained for thallium and gated blood pool exercise scans. *p < 0 .05 propranolol versus no propranolol; t p < 0.05 rest versus exercise. BP = biood pressure; EF = ejection fraction; Ex = exercise; HR = heart rate; % Pred HR = % predicted maximal heart rate; Prop = propranolol.

reduced in thickness or intensity and + 2 = normal myocardial thickness and intensity. If the observer felt the defect was intermediate between two grades, half scores were permitted. Criteria fo r abnormality were generatedfrom a previous analysis oj 50 patients (3 J) and were as f ollows . In men,

a score of 1.5 or less at maximal exercise was considered abnonnal for ihe anterior, inferior, septal and posterior walls. In women. a score of 1.0 or less in the anterior and septal walls and 1.5 or less for the inferior and posterolateralwalls was necessary for the initial scan to be read as abnormal . The different criteria for women were chosen because of the possibility of attenuation of activity due to breast tissue. For the apical and apicoinferior wall segments, an initial score of 0.75 or less was considered abnormal for men and women. The lower criterion for these walls was chosen because the apical walls may normally be thinner than the other walls, If the initial defect filled in (that is, score increased by 2: 0.5 point), a transient defect existed. If the defect did not fill in (that is, score increased < 0.5 point), a persistent defect existed. Gated scan ventricular function analysis. The left ventricle was similarly divided into six segments in two views for rest wall motion analysis, and into three segments in the left anterior oblique view for exercise wall motion analysis. A fi ve point scoring system was used for each wall: - I = dyskinesia, 0 = akinesia, + I = moderate hypokinesia, + 2 = mild hypokinesia and + 3 = normal. A score for each segment at rest and. during exercise was generated by averaging the score of three independent observers in a manner previously reported (32). Criteria for abnormality were as fo llows. I) Rest wall motion was considered abnormal if the average score was 1.5 or less for the anterior, inferior, septal or posterolateral segments and 1.0 or less for the apical and apicoinferior

segments. 2) Exercise wall motion was considered abnormal if there was at least a 1.0 point decrease from the rest value. Ejection fraction. Rest and exercise ejection fractions were determined from the left anterior oblique view using a semiautomated edgedetection and counts technique (MUGE program, Medical Data Systems. Ann Arbor. Michigan) by two independent observers. The results were averaged as previously reported (32). An abnormal rest ejection fraction was considered to be less than 50%. In patients with a rest ejection fraction below 70%, an increase of less than 5% was considered an abnormal exercise response. For a rest ejection fraction greater than 70%, an abnormal exercise response was a decrease of 5% or greater (33). The range of rest ejection fractions was 25 to 80%. However, because our patient group was chosen because of its relative clinical stability (that is, no congestive heart failure or unstable angina), most of our patients had a rest ejection fraction above 40%. Only fi ve patients had a rest ejection fraction less than 40%. Mean rest and exercise ejection fractions are presented in Table I. Pulmonary blood volume data. Changes in pulmonary blood volume from rest to exercise were determined by comparing the exercise and rest activity in a region of interest placed over the lung (normalized for collection time). A previous study in our laboratory (34) determined that an exercise/rest ratio of 1.06 or greater was abnormal and indicative of exercise-induced ventricular dysfunction as seen in coronary artery disease. Electrocardiographic interpretations. Exercise electrocardiograms were considered diagnostic of coronary artery disease (ischemia) if there was I mm or more horizontal or downsloping ST segment depression 0.08 second after the J point. Final scan data interpretation. Thallium scans were considered abnormal if there were transient or persistent

lACC Vo!. 3, No.2 February 1984:272-83

OSBAKKEN ET AL. DIAGNOSTIC ACCl.'RACY OF THALLIUM AND GATED BLOOD POOL SCANS

defects, Gated scans were considered abnormal if there were rest or exercise wall motion or ejection fraction abnormalities as defined in a previous section. The effect of pulmonary blood volume was evaluated separately and was not included in the final gated scan score. Statistical analysis. Sensitivity, specificity and diagnostic accuracy were determined as follows: Sensitivity Specificity Diagnostic accuracy

True positives True positives and False negatives' True negatives True negatives and False positives' True positives + True negatives Total patients

Student t tests and Fisher exact tests were used for comparisons of various patient data. Patient data were analyzed for the total study group and then for various subgroups based on clinical symptoms, medication, sex, exercise level. electrocardiogram (rest and exercise) and number and extent of diseased vessels. Rest and exercise scores were combined to produce a final score for each scan. Hemodynamic data are expressed as mean values ± standard deviation.

Results Characteristics of study patients. The mean age (± standard deviation) was similar for those with coronary artery disease (52 ± 1.8 years) and those without coronary artery disease (49 ± 2,6 years) (Table I). There were 73 men and 13 women with coronary disease and 16 men and 18 women without coronary disease. Twenty-six patients with coronary artery disease had single vessel disease, 32 had double vessel disease and 28 had triple vessel disease. Nine of the patients without coronary artery disease had subcriticallesions (> 20%, :s 49% narrowing in any vessel) and 25 had no disease (:s 20% narrowing in any vessel). Among the patients with coronary disease, 66 were taking propranolol and 20 were not; of the patients without coronary disease. 19 were taking propranolol and 15 were not. Forty-four patients had a history of previous myocardial infarction, Overall sensitivity, specificity and diagnostic accuracy (Table 2). Except for gated wall motion. which was less sensitive and more specific for coronary artery disease (probability [p] < 0.05), there were no significant differences between the sensitivity, specificity or diagnostic accuracy of any scan or clinical variable. Anginal history, electrocardiogram, thallium scan and gated scan results were combined in groups of two. three and four (anyone positive variable constituting a positive study for the combination) to determine sensitivity, specificity and diagnostic accuracy in our total patient group. The combination of two or more variables generally in-

275

Table 2. Total Group Sensitivity, Specificity and Diagnostic Accuracy of Anginal History, Electrocardiogram, Thallium Scan and Gated Scan Alone and Combined

Typical anginal history (A) Exercise electrocardiogram

Sensitivity (n = 86)

Specificity

72'1<

71%

72%

72'1<

710/<

720/<

(n =

34)

Diagnostic Accuracy (n = 120)

(E)

Thallium (f) Gated wall motion Ejection fraction Pulmonary blood volume Gated wall motion + ejection fraction (G) Gated wall motion + ejection fraction + pulmonary blood volume (G')

76'/'

73'7r 54"*

43'/'

73'1<

74'*

69')f

73";' 77'lr

80'/,

61"1<

75"*

9Y)f

59'X

830/<

A -t E A + T

91'1< 87'7r

53'/' 53')(

A+G E + T

88'/'

44Ck 65'7< 41Ck 44Ck 500/, 440/< 290/< 350/< 260/<

80% 78'7, 76'7< 720/< 75Ck 770/< 81'7< 750/, 760/< 740/,

21%

75"1<

+ G

87"1< 84'1<

T+G

87'/'

E G

+ E

87"1<

A A

+E+T +E+G

96 c)f 93'/'

E A

+ T + G

9Y;'

+

93')f 97'/'

T

+

G

A+E+T+G

81%

creased the sensttivity. decreased the specificity, and did not affect the diagnostic accuracy. Effect of clinical variables on scan results. History of angina, previous infarction and electrocardiographic Q waves. Individual scan results in patients with typical angina

(62 patients) were not significantly different from those in patients with atypical angina (24 patients) (Fig. I). However, both scans were more frequently positive in the typical angina patients (p < 0.05). More frequent rest or exercise wall motion abnormalities. or both, were found in patients with a history of previous myocardial infarction (44 patients) (p < 0,05) (Fig. 2). Patients with Q waves (31 patients) on the electrocardiogram had significantly more frequent wall motion abnormalities than did patients without Q waves (Fig. 2). Both scans were more frequently positive in patients with Q waves (p < 0.05), In the 20 patients with a comhination of anKina, previous infarction and Q waves, both the thallium and gated scans,

individually or combined, were highly sensitive in predicting the existence of disease (Fig. 3). Sensitivity of thallium

OSBAKKEN ET AL. DIAGNOSTIC ACCU RACY OF THALLIUM AND GATED BLOOD POOL SCANS

276

100 ~

80 70 60 %50

SENSITIVITY (n = 86) Typical Angina Atypical Angina (n = 61) 100 (n =25) ,... ~ .....

.....

.-

40 30

.....

..........

80 70 60 %50

..... ..... .....

40 30

I-

20 10

.....

.....

.....*

-

~

20 10 T

G

EFPBY G/ GI EFEfI

T

T. Tor G Q

G

EFPBV Gf G1 EF Ef I

T + T or G G

PaY

PaY

SPECIFICITY (n = 34) (n= 10)

100

.....

90

%

80 70 60 50

lACC Vol. 3. No.2 February 1984:272-83

..... .....

-

90

1--

-

%

80 70 60 50

40 30

40 30

20 10

20 10 T

G

(n=24)

100

EFPBV GI GI EFEFI

..........

.....1I-

.....

T

T+ Tor G G

-

G

EF PRY GJ GJ EF EFI

.BY

Figure 1. Sensitivity, specificity and diagnostic accuracy of thallium and gated equilibrium blood pool scans in patients with and without coronary artery disease and without a history of typical angina. EF = ejection fraction (rest or exercise abnormality, or both); G = gated wall motion (rest or exercise abnormality, or both); PBV = pulmonary bloodvolume; T = thallium scan (transient or persistent defect, or both).

T . Tor G G

.BY

DIAGNOSTIC ACCURACY (n = 120) (n =71)

100 90

%

80 70 60 50

.....

.....

- ..... -

90

.....

80 70

-

40 30

(n =49)

100

%

-

.....

60 50

I-

.....

- ..... r-

40 30

20 10

20 10 T

G

EF POV GI

GI

EF EFI 'BY

T + T or

G

G

T

G

EF PaY GJ GI EF EF'

T . TOf GG

.BY

and gated wall motion scans was significantly lower in the 23 patients with coronary disease without this combination (p < 0.05). Specificity in the 34 patients without typical angina, history of infarction and Q waves is presented in Table 2 and represents our normal group. Gender (Table 3) . Anginal history, thallium scans and gated wall motion were significantly less sensitive for coronary artery disease in women than in men (p < 0.05). Propranolol (Table 4). A significant proportion of our patients were taking propranolol at the time of scintigraphic study; 77% of the coronary disease group and 56% of the noncoronary group. In the group with coronarydisease, those taking propranolol had significantly lower peak exercise blood pressure, lower rest and exercise heart rate responses and a lower exercise level than those patients not taking propranolol (Table I) . Propranolol did not significantly af-

feet rest or exercise hemodynamic variables in the group without coronary disease. Propranolol administration did not significantly affect the sensitivity of any diagnostic scan variable in the total group (Table 4) nor in any patient subgroups. The specificity of thallium scans was significantly lower for patients taking propranolol (p < 0.05). There were 28 apparent transient defects in 19 normal patients taking propranolol: 6 in the anterolateral, I in the inferior, 2 in the posterior, 7 in the septal and 12 in the apicoinferior segment. Apicoinferior defects were only included as defects if they were seen in both the anterior and the left anterior oblique 45° views. Only two defects were seen in patients not taking propranolol; one was in the septum and one in the apex. This pattern of defects was not significantly different from those found in patients with coronary disease.

lACC Vol. 3. No.2 February 1984:272-83

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOL SCANS

277

SENSITIVITY

100

%

Myocardial Infarct (n =44) ,.-

90 80 70 60 50 40 30 20

,.-

No Myocardial Infarct (n = 42)

-

100

rr-

%

10

Figure 2. Sensitivity of thallium and gated equilibrium blood pool scans in patients with and without a history of myocardial infarction and with and without Q waves. *p < 0.05 (myocardial infarct vs. no infarct and Q waves vs. no Q wave). Abbreviations in Figure I.

r-

90 80

r-

r-

70 60 50 40 30 20 10

- r-

r-

r-*

TGEFP~WW

TGEF~VWW

EFE~

EFE~

PBV

PBY

*P
100

%

90 80 70 60 50 40 30 20

r-rr-

rr-

r%

10 G

EF PBY GI

GI

EF EFt

T+ Tor G

r-

90 80

r- .....

T

No Q Waves (n = 55)

100

70 60 50 40 30 20 10

rr-

*

r-r-r"'"

*

r-

* tT

G

.BV

G

EFPBYGf GI EFEFI .BV

T+Tor G G

P
There was no significant difference in diagnostic accuracy of any of the evaluated variables in patients who were or were not taking propranolol. Effects of exercise. Exercise heart rate, blood pressure, percent maximal predicted heart rate and ejection fraction data are presented in Table I. Exercise level did not significantly affect thallium or gated scan sensitivity, specificity or diagnostic accuracy (Table 5). There was no significant difference in thallium scan and gated scan sensitivity or specificity based on exercise angina or ischemia (I mm ST segment depression 0.08 second out from the J point), or both (Fig. 4). Number and extent of diseased vessels (Table 6). The sensitivity of each scan variable increased as the number of diseased vessels increased. However, the difference was significant only for thallium scans in patients with triple vessel disease compared with single and double vessel disease (p < 0.05). There was no significant difference in any scan variable based on luminal narrowing (Fig. 5) when the patients were separated into groups with 50% narrowing, 50 to 75% narrowing and 75% or more narrowing, based on the largest obstruction in any vessel.

Discussion Sensitivity,

specificity

and diagnostic accuracy.

Evaluation of our total group (120 patients) demonstrated that anginal history, exercise electrocardiogram, thallium scan (transient or persistent defects) and gated blood pool scans (rest or exercise wall motion and ejection fraction) had similar sensitivity, specificity and diagnostic accuracy. Combinations of two or more diagnostic variables (clinical plus scintigraphic) increased sensitivity, decreased specificity and did not change overall diagnostic accuracy for coronary artery disease in the total group. Combined thallium scan scores (electrocardiogram plus thallium distribution interpretation) and gated scan scores (wall motion, ejection fraction plus pulmonary blood volume) were more accurate in diagnosing disease from nondisease than either variable alone. Thus, none of the clinical or scan variables, singly or in combination, is the ultimate diagnostic tool for evaluation of coronary artery disease in a general population of patients with moderately high disease prevalence.

Effect of clinical variables on scan performance. History. electrocardiogram and number ofdiseased vessels.

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOL SCANS

278

lACC Vol. 3. No. 2

February 1984:272- 83

SENSITIVITY CAD - Angina , - Myocardial Infarct, - Q Waves (n = 23)

CAD

+ Angina, + Myocardial Infarct, + Q Waves (n =20) 100

...-

90

...-

I-

80 70

100

...--

80 70

.....

60

60

%

...-

90

%

50 40 30

...-*

T

G

r--

I*-

50 40 30

20 10

20 10

Figure 3. Sensitivity of thallium and gated equi librium blood pool scans in patients with coronary artery disease (CAD) with angina (A) plus myocardial infarction (MI) plus Q waves (Q) vs. those without angina, myocardial infarction or Q waves. Abbreviations in Figure 1.

r--

I*T

EFPBVGI GI EF EFI PBV

EFPBVGI GI EF EFI PBV

G

·P<0 .05 (+ A, MI, Q vs - A, MI, Q)

In the present study. thallium and gated blood pool scans were slightly more accurate in the diagnosis of coronary artery disease in patients with typical angina than in those without. We found that patients with a history of previous myocardial infarction more frequently had wall motion abnormalities than did those without (p < 0.05), and that patients with Q waves more frequently had ejection fraction and combined thallium and gated scan abnormalities than did those without Q waves (p < 0.05). This can be interpreted as an obvious consequence of permanently damaged myocardium in patients with a prior infarction. The presence or absence of ischemia on the exercise electrocardiogram did not affect diagnostic accuracy of any of the scan variables. In addition, patients with triple vessel disease had a greater number of positive scan variables than did patients with single and double vessel disease. Thus, in patient groups with a high prevalence of postmyocardial infarction and triple vessel disease, both thallium and gated scans would be expected to be quite sensitive for coronary disease. However, it is in this group that cardiac scintigraphy is least necessary to provide a diagnosis.

Effects of gender. It is well known from clinical experience that women have a higher incidence of atypical angina than do men. This may account for the decreased diagnostic accuracy of angina in women. Women have signifi cantly more breast tissue than do men. This tissue may attenuate thallium activity and make qualitative interpretation of scan data difficult, thus causing the test to have less diagnostic accuracy. Therefore, more stringent criteria for abnormality were developed to evaluate thallium scans in women. Eff ects of propranolol . One mechanism that may be responsible for variable sensitivity and specifi city of thallium and gated scans may be due to a direct effect of propranolol of the vasculature, that is, propranolol may decrease blood fl ow to nonischemic myocardium (35-37). This may be related to direct beta-receptor blockade of the coronary arteries (36) or to reflex phenomena based on a decreased oxygen demand secondary to a propranolol-related decrease in heart rate, left ventricular pressure-time index and velocity and force of contraction (35,36). It is possible that one or a combination of these mechanisms decreased myo-

Table 3. Gender Effects on Sensitivity, Specificity and Diagnostic Accuracy Sensitivity Men (n

Angina ECG T

GWM EF PBV

GWM + GWM +

EF EF

+

PBV

= 73) 77%74% 79 %48%73% 72% 8 1% 88%

Specificity Women (n

=

13)

46% 6 10/0 54% 15% 77% 54% 77% 77%

(n

Men = 171 7 1% 82% 65% 7 1% 65% 100% 53% 47%

Diagnostic Accuracy Women = I7l

(n

71% 82% 59% 94% 76% 100% 71% 76%

Men (n

= 90) 76% 76% 77% 52% 7 1% 76% 76% 80%

Women (n = 30 ) 60o/c 73% 57% 60 % 77% 67 % 73% 77%

-p < 0.0 5 (men versus women). ECG = Q wave or exercise ischemia; EF = ejection fraction (rest or exercise abnormality, or both); GWM = gated wall motion (rest orexercise abnormality, or both); PBV = pulmonary blood volume; T = thallium scan (transient or persistent defect , or both ).

lACC

279

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF TH AL LI UM AND GATED BLOOD POOL SCANS

vot. 3. No.2

February IlJ84:272-83

Table 4. Effects of Propranolol (Prop) on Se nsitivi ty. Specifici ty and Diagnostic Acc uracy

=

66)

(n =

T

GWM EF PBY

GWM or EF GWM + EF + PBY

(n

20)

609, 60'11 7SCk 45% 80
76
Angina

ECG

Prop

No Prop

Prop (n

Diagnostic Accuracy

Specificity

Sensitivity =

No Prop

19)

(n =

6YI< 74
80Ck 9YI< 8Yk 8Yk 6Yk 10Wk

6lVk 79Cfr

5S
No Prop

Prop

15)

(n =

85)

(n

69% 74% lIO% 63% 74'7( 66% 69% 62%

7YYr 75'"k 68'"k 51 '11 7Yk 6Wk 7l1'11 8l1'"k

5N

= 35)

*p < O.OS propranolol versus no propranolol. Angina = history of typical angina. Other abbreviations as before.

cardial perfusion in our propranolol-treated normal patients, resulting in decreased thallium uptake and. thus. false positive thallium scans. This decreased fl ow could have been more pronounced in the subendocardium than in the epicardium, which could be perceived as wall thinning. We also found a lower specificity for gated blood pool scan for patients on propranolol. Because of its beta-receptor blockade effects, propranolol may decrease ventricular function by decreasing force and velocity of contraction. Thus, it is possible that patients treated with propranolol may appear to have decreased function, as evidenced by low ejection fraction measured by gated blood pool scans. This could cause decreased ventricular function in patients with coronary disease and in normal patients, thus increasing sensitivity and decreasing specificity of gated blood pool scans. Alternatively. propranolol could prevent ischemia during exercise and allow the heart to function more normally (38- 40). This may decrease the sensitivity of both thallium perfusion and gated blood pool function scans, thus decreasing the overall diagnostic accuracy of each scan. Comparison of our data with those of other investigators. Comparison of imaging techniques. Differences in scanning techniques may infl uence the quality of thallium and gated images and thus affect the fi nal interpretation. We performed nuclear imaging with a single crystal camera (25-30) within I week of contrast angiography, similar to several other groups (24,27 ,30). so there was minimal time

for anatomic change between the contrast and scintigraphic studies. Two view thallium scans were acquired in our study (26.27 .30), whereas three views were used by several other groups (24,25,28.29) , which may potentially decrease our sensitivity. However. other groups (26.27 ,30) who used two views obtained similar sensitivity to those using three views (24,25). Therefore, the number of views does not seem to be a factor in overall scan reliability. Wall motion at rest was evaluated in two views (anterior and 45° left anterior oblique views in the gated equilibrium mode), while exercise wall motion was evaluated in one view (45° left anterior oblique view) (27-30) . Some groups (24 .25) used one view and fi rst pass techniques for wall motion analysis. This may also alter the sensitivity of gated blood pool scans. However, the use of two views would tend to increase the sensitivity of locating wall motion abnormalities. Both of our exercise protocols were performed with supine bicycle exercise on the same day. Although hemodynamic responses to supine exercise are somewhat different than those to upright exercise, the value of the rest-exercise response in diagnosing coronary artery disease has been shown to be similar in both positions (4 1.42). A variety of different exercise protocols have been used by other groups. These include treadmill (26.28), upright and supine bicycle (25.27.4 1,42) and handgrip (24) exercise. Some groups (28.30) performed both exercise protocols on the same day, while others (24-27) performed them on separate days.

Table 5. Effec t of Exerci se Level (% predi cted max imal heart rate) on Sensiti vity . Specificity and Diagnosti c Accuracy Sensitivity

> 85%, Pred HR (n = 10)

ECG T

GWM EF PBY

GWM or EF

60% 70'7< 50o/c

60% 70% 90%

Abbreviations as before.

Diagnostic Accuracy

Specificity

< lIS'7< Pred HR

> 8S (1< Prcd HR

=

In = II I

(n

76)

< lI5'if Pred HR In

=

23)

73'1<

87'if

64';'

6 I r;\

92'1< 5Yl<

78Cfr 78ck 100%

100'1< 55<1f

6591

> lI5Ck Prcd HR (n =

21)

< lI5 c/c Pred HR (n =

99)

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOL SCANS

280

JACC Vol. 3. No.2

February 1984:272-83

SENSITIVITY (n = 86)

+ Angina (n = 29) + Ischemia

0/0

100 90 rI80 70 -~ 60 50 I40 30 20 10 T

G

-

- .....

EF PBY GJ GJ EFEFI

100 90 80 70 60 50 40 30 20 10

T+ Tor G G

+ Angina (n =30) - Ischemia

.....

-

r-

r-

_r-

..........

T

G

EFPBV GJ GJ EFEFI

- Angina (n =13)

100 90 80 70 f060 50 40 I30 20 10

T+ Tor G G

T

G

-

r-

-

rr-

.....

EF PBV GJ GI EFEF/

.BY

.BY

- Angina (n = 14) -Ischemia

+ Ischemia

100 90 80 I 70 60 50 40 30 20 10

T+ Tor G G

T

.....

rr-

-

G

.....

r-

,....

EFPBVG/ GJ EFEFJ

.BY

1+10r G G

.BY

SPECIFICITY (n = 34)

100 90 80 70 60 0/0 50 40 30 20 10

-

(n=5)

(n=5)

r-

.....

T

G

- - .....

r-

EF PBY GJ GJ EF EFI

T+ G

100 90 80 70 60 50 40 30 20 10

-

T

-

G

-

Performance of exercise on the same day or different days, may cause different problems. In same day exercise protocols, the first exercise procedure may be sufficiently exhausting, so that the second exercise level is less extreme. However, this did not appear to be the case in our patient group, which exercised for similar time periods to similar work loads for each study. Extent of exercise may influence the scan results. However, if exercise was performed to similar predefined levels (symptom-limited maximum), the type of exercise used is probably not important. In protocols where the two exercise scans were performed on different days, the physiologic state of the patient may be different, which could result in nonhomogeneous perfusion-function results. In addition, although our scan scoring procedures were similar to those of others, our criteria for normalcy were more stringent, which could be reflected by our decreased specificity. Patients. Patient selection plays a role in the accuracy of a test in diagnosing the presence or absence of disease. A patient group that is heavily weighted toward having

r-

EF PlY GI

.BY

Figure 4. Sensitivity and specificity of thallium and gated equilibrium blood pool scans in patients with and without coronary artery disease with and without exercise electrocardiographic ischemia (I mm ST segment depression 0.08 second from the J point) or angina. or both. Abbreviations as in Figure I.

(n =24)

-,....

GJ

T+

EF EFI .BY

G

.....

100 90 80 70 60 50 40 30 20 10

-

T

G

-

EFPBVGJ OJ EF EFI

-

T+ G

.BY

disease or toward normalcy will have a greater tendency to have positive or negative tests than a more randomly selected group. Our patient group was much larger, but with similar age, sex, and clinical stability distributions when compared with clinical groups of other authors (20-35). Although a significant number of our patients (44 of 86) had a history of previous myocardial infarction (similar to other groups; 25,27-29), there was a fairly homogeneous distribution of single, double and triple vessel disease. Several previous Table 6. Effect of Number of Diseased Coronary Vessels on Sensitivity Coronary Artery Disease

Angina ECG T GWM EF

PBV GWM or EF GWM + EF

I Vessel (n = 26)

2 Vessel (n = 32)

62% 58% 69% 23% 65% 68%

78% 72% 69% 47% 75% 69% 81% 90%

690/c

+

PBV

77%

3 Vessel = 28)

(n

75% 86%* 89%t 57% 79% 71% 89% 92%

*p < 0.05 (I vessel versus 3 vessel disease); tp < 0.05 (I and 2 vessel versus 3 vessel disease. Abbreviations as before.

SENSITIVITY

100 80

%

90

90 80 10 60 50 40 30 20 10

r-

80

r-~

~50% 100

100

r-

r-

70

60 50 40 30 20 10

Group III Obstruction (n =86)

>50% >75%

-

90

(n = 86)

Group II Obstruction (n=9)

Group I Obstruction (n = 77)

~ 75%

Figure 5. Sensitivityand specificityof thallium and gated equilibrium blood pool scans in patients with and without coronary artery disease grouped accordingto severity of vessel obstruction. Abbreviations as in Figure I.

281

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOLSCANS

lACC Vol. 3, No.2 February 1984:272-83

r-

70 60

%

f-

A

50 40 30 20 10

%

;.. ~

F ECGTGEPBYGJGJ

... ECG T

EF PlY GI

0

EF EFI

PBV

PBV

r-

A ECG T

'"

EF EFI

r-

r-

0

EF PlY GI 0'

EFEFI POV

·P<0.05 (Group II vs Group I) (n = 34)

SPECIFICITY

<50%

<50% Obstruction (n = 34) r-

100

70

60

r-

,....

ft-~

0/0 50 40 30 20 10

90

r-

80

80

70

10

60

%

<20% Obstruction (n = 25) 100

r-

90

r-

Obstruction (n=9)

100

90 80

~20%

rr-

~r-

%

10 A ECG T

studies (24,28) were more heavily weighted toward triple vessel disease and one study (29) included only patients with previous myocardial infarction. These clinical weighting factors could account for the greater positivity (sensitivity) of thallium and gated scans in these patient groups. Other groups (24,25) used more stringent criteria for significant coronary artery disease (~ 75% obstruction). Close scrutiny of our patients without coronary artery disease showed that we may have had a greater incidence of subcritical disease than some other groups. All of these factors, singly or in combination, could account for lower diagnostic accuracy. Thus, patient population differences could explain differences in scan sensitivity, specificity and diagnostic accuracy. Clinical implications. Our data indicate that although both thallium and gated blood pool scans provide relevant information concerning ventricular perfusion and function, neither scan alone or combined can unequivocally diagnose coronary artery disease in a general group of patients with a moderately high prevalence of disease. The following points should be made. Best results from thallium and gated scans are obtained when combined variables for each scan are used; that is, thallium distribution plus electrocardiographic interpretation for thallium scans; wall motion, ejection fraction and pulmonary blood volume for gated scans. Thallium scans must be interpreted with caution in women

G

EFP8VG/ GJ EF EFI PBV

A

co

t

0

f-

60

~

50 40 30 20

r

r-

EF PBV GI 01 EF EFI FBV

50 40 30 20 10 "ECGT

G

EF PBVGI GI EF EFI FBV

because of their low diagnostic accuracy. Propranolol should be discontinued sufficiently in advance of thallium and gated scans so that its effects can be cleared from the system, because of the lowered diagnostic accuracy of both types of scan in patients taking propranolol. Thallium or gated scans, or both, can be useful in providing information concerning myocardial perfusion and function in patients with documented coronary artery disease. In patients with equivocal evidence of coronary disease, neither scan has good sensitivity or specificity in diagnosing coronary artery disease and cannot be used to provide an accurate diagnosis. We conclude that thallium and gated blood pool scans have many limitations when used for diagnosis of coronary artery disease as discussed in this report. Thus, they can be used more advantageously when these factors are considered. We extend special thanks to Judy Perry for her thoughtful and patient help in preparation of this manuscript.

References I. Turner DA, Battle WE. Deshmukh H, et al. The predictive value of myocardial perfusion scintigraphy after stress in patients without previous myocardial infarction. J Nucl Med 1975;19:249-59.

282

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLIUM AND GATED BLOOD POOL SCANS

lACC Vol. 3. No.2

February 1984:272-83

2. Botvinick EH, Taradosh MR. Shames OM, Parmley WW. Thallium201 myocardial perfusion scintigraphy for the clinical clarification of normal, abnormal, and equivocal electrocardiographic stress test. Am 1 Cardiol 1978;41:43-51.

20. Brady TJ, Thrall JH, Lo K, Pitt B. Factors affecting the sensitivity and specificity of exercise radionuclide ventriculography in coronary artery disease detection: the importance of adequate exercise. 1 Nucl Med [980;21:339-43.

3. Massie BM, Botvinick EH, Brundage BH. Correlation of thallium20 I scintigrams with coronary anatomy: factors affecting region by region sensitivity. Am 1 Cardiol 1979;44:616-22.

21. Bodenheimer MM, Banka YS, Fooshee CM, Gillespie lA. Helfant RH. Detection of coronary heart disease using radionuclide determined regional ejection fraction at rest and during hand grip exercise: correlation with coronary arteriography. Circulation 1978:58:640-8.

4. Ritchie Jl., Zaret BL. Strauss HW, et al. Myocardial imaging with thallium-20 I: a multicenter study in patients with angina pectoris or acute myocardial infarction. Am 1 Cardiol 1978;42:345-50. 5. Dunn RF, Bailey [K. Wren R, Kelly DT. Exercise induced ST-segment elevation: correlation of thallium-20 I myocardial perfusion scanning and coronary arteriography. Circulation 1980;61 :989-5. 6. Bodenheimer MM, Banka YS. Fooshee CM. Helfant RH. Extent and severity of coronary heart disease as determined by thalljum-20 I myocardial perfusion scanning: comparison with stress electrocardiography. Arch Intern Med 1979;139:630-4. 7. McCarthy OM, Blood OK, Sciacca RR, Cannon Pl. Single dose myocardial perfusion imaging with thallium-20I: application in patients with nondiagnostic electrocardiographic stress tests. Am 1 Cardiol 1979;43:899-906. 8. Bodenheimer MM. Banka YS. Helfant RH. Nuclear cardiology. [I. The role of myocardial perfusion imaging using thallium-201 in diagnosis of coronary heart disease. Am 1 Cardiol 1980;45:674-84. 9. Rigo P, Bailey [K, Griffith LSC, et al. Yalue and limitations of segmental analysis of stress thallium myocardial imaging for localization of coronary artery disease. Circulation 1980:61 :973-81. 10. Bodenheimer MM. Banka YS, Fooshee CM. Herman GA. Helfant R. Relationship between regional myocardial perfusion and the presence, severity, and reversibility of asynergy in patients with coronary heart disease. Circulation 1978;58:789-94. II. Massie BM, Botvinick EH, Brundage BH. Greenberg B, Shames 0, Gelberg H. Relationship of regional myocardial perfusion to segmental wall motion: a physiologic basis for understanding the presence and reversibility of asynergy. Circulation 1978:58:1154-63.

12. Ritchie Jl., Trobaugh GB, Hamilton GW, et al. Myocardial imaging with thallium-201 at rest and during exercise: comparison with coronary arteriography and resting and stress electrocardiography. Circulation 1977:56:66-71. 13. Dash H. Massie BM, Botvinick EH, Brundage BH. The non-invasive identification of left main and three-vessel coronary artery disease by myocardial stress perfusion scintigraphy and treadmill exercise electrocardiography. Circulation 1979;60:276-84. 14. Bailey [K, Griffith SC, Rouleau 1, Strauss W. Pitt B. Thallium-201 myocardial perfusion imaging at rest and during exercise: comparative sensitivity to electrocardiography in coronary artery disease. Circulation 1977;55:79-87. 15. Mclin lA. Piret LG, Yanbukele RIM. et al. Diagnostic value of exercise electrocardiography and thallium myocardial scintigraphy in patients without previous myocardial infarction. A Bayesian approach. Circulation 1981:63:1019-24.

22. Jones RH, McEwan P, Newman GE. et al. Accuracy of diagnosis of coronary artery disease by radionuclide measurement of left ventricular function during rest and exercise. Circulation 1981 ;64:586-601. 23. Diamond GA, Forrester IS. Improved interpretation of a continuous variable in diagnostic testing: probabilistic analysis of scintigraphic rest and exercise left ventricular ejection fraction for coronary disease detection. Am Heart 1 1981:102:189-96. 24. Bodenheimer MM, Banka YS, Fooshee CM, Helfant RH. Comparative sensitivity of the exercise electrocardiogram: thallium imaging and stress radionuclide angiography to detect the presence and severity of coronary heart disease. Circulation 1979;60:1270-8. 25. lengo lA, Freeman R, Brizendine M, Mena I. Detection of coronary artery disease: comparison of exercise stress radionuclide angiography and thallium stress perfusion scanning. Am 1 Cardiol 1980:45:535-41. 26. Caldwell IH, Hamilton GW, Sorenson SG. Ritchie n., Williams DL, Kennedy lW. The detection of coronary artery disease with radionuclide techniques: a comparison of rest-exercise thallium imaging and ejection fraction response. Circulation 1980;61 :610-9. 27. lohnstone DE. Sands MJ, Berger HJ, et al. Comparison of exercise radionuclide angiography and thallium-201 myocardial perfusion imaging in coronary artery disease. Am 1 Cardiol 1980:45:1113-9. 28. Elkayam \,J, Weinstein M, Berman 0, et al. Stress thallium-201 myocardial scintigraphy and exercise technetium ventriculography in detection and location of chronic coronary artery disease: comparison of sensitivity and specificity of these non-invasive tests alone and in combination. Am Heart 11981:101:657-66. 29. Tiefenbrunn Al, Biello DR, Gellman EM, Sobel BE, Siegel BA, Roberts R. Gated cardiac blood pool imaging and thallium-201 myocardial scintigraphy for detection of remote myocardial infarction. Am 1 Cardiol 1981;47: 1-6. 30. Kirshenbaum HD. Okada RD, Boucher CA. Kushner FG. Strauss HW, Pohost GM. Relationship of Thallium-201 myocardial perfusion pattern to regional and global left ventricular function with exercise. Am Heart 1 1981;101:734-9. 31. Okada RD, Boucher CA, Kirshenbaum HK, et al. Improved diagnostic accuracy of thallium-201 stress test using multiple observers and criteria derived from intcrobscrver analysis of variance. Am 1 Cardiol 1980;46:619- 25. 32. Okada RD, Kirshenbaum HD. Kishner FG, et al. Observer variance in the qualitative evaluation of left ventricular wall motion and the quantitation of left ventricular ejection fraction using rest and exercise multigated blood pool imaging. Circulation 1980;61:128-36.

16. Ong YS. Quaife MA. Dzindzio BS, Emery IF, Kotlyarov EY, Forker AD. Clinical decision making with treadmill testing and thallium-201. Am 1 Med 1980:69:31-8.

33. Osbakken MD, Boucher CA, Okada RD, Bingham lB, Strauss HW, Pohost GM. Spectrum of global left ventricular responses to supine exercise: limitation in the use of ejection fraction in identifying patients with coronary artery disease. Am 1 Cardiol 1983:51:28-35.

17. Bodenheimer MM, Banka YS, Fooshee CM. Gillespie lA, Helfant RH. Detection of coronary heart disease using radionuclide determined regional ejection fraction at rest and during hand grip exercise: correlation with coronary arteriography. Circulation 1978;58:640-8.

34. Okada RD, Pohost GM, Kirshenbaum HD. et al. Radionuclide determined change in pulmonary blood volume with exercise: improved sensitivity of rnultigated blood pool scanning in detecting coronary artery disease. N Engl 1 Med 1979;301 :569-76.

18. Borer 15, Kent RM, Bacharach SL, et al. Sensitivity. specificity. and

35. Kloner RA. Reimer KA. lennings RB. Distribution of coronary collateral flow in acute myocardial ischemic injury: effect of propranolol. Cardiovasc Res 1976:10:81-90.

predictive accuracy of radionuclide cineangiography during exercise in patients with CAD: comparison with exercise electrocardiography. Circulation 1979;60:572-80. 19. Bodenheimer MM, Banka YS, Helfant RH. Nuclear cardiology. I. Radionuclide angiographic assessment of left ventricular contraction: uses, limitations. and future directions. Am 1 Cardiol 1980;45:661-73.

36. Pitt B, Craven P. Effect of propranolol on regional myocardial bloop now in acute ischemia. Cardiovasc Res 1970;4:176-9. 37. Becker LC, Ferreira R. Thomas M. Effect of propranolol and iso-

lACC Vol. 3, No.2 February 1984:272- 83

OSBAKKEN ET AL. DIAGNOSTIC ACCURACY OF THALLI UM AND GATED BLOOD POOL SCANS

prenaline on regional left ventricular blood flow in experimental myocardial ischemia. Cardiovasc Res 1975:9:178- 86. 38. Marshall RC, Berger HJ, Reduto LA, Cohen LS, Gottschalk A, Zaret

BL. Assessment of cardiac performance with quantitative radionuclide angiography: effects of oral propranolol on global and regional left ventricular function in coronary artery disease. Circulation

283

Serial exercise radionuclide angiography: validation of count-derived changes in cardiac output and quantitation of maximal exercise ventricular volume change after nitroglycerin and piqpranolol in normal I man. Circulation 1980:61:600-9 . 4 1. Manyari DE, Kostuk WJ. Left and right ventricular function at rest

39. Battler A. Ross J Jr, Slutsky R. Pfisterer M. Ashburn W. Froelicher

and during bicycle exercise in the supine and sitting positions in normal subjects and patients with coronary artery disease: assessment by radionuclide ventriculography. Am J Cardiel 1983:51:36-42,

V, Improvement of exercise-induced left ventricular dysfunction with oral propranolol in patients with coronary heart disease. Am J Cardiol

42 . Polimer LR, Kehmer GJ. Lewis ~E , Parkey RW. Blomgrist CG, Willerson JT. Left ventricular performance in normal subjects: a com-

1978:58:808- I 4.

1979;44:318- 24. 40 . Sorenson SG, Ritchie JL, Caldwell JH, Hamilton GW, Kennedy JW.

parison of the responses to exercise in the upright and supine positions. Circulation 1980:62:528- 34 .