- Email: [email protected]
Assessment of early ventricular systole by first pass radionuclide angiography: Useful method for detection of left ventricular dysfunction at rest in patients with coronary artery disease
METHODS
Assessment of Early Ventricular Systole by First Pass Radionuclide Angiography: Useful Method for Detection of Left Ventricular Dysfunction at Rest in Patients With Coronary Artery Disease
ROBERT SLUTSKY, MD DONALD GORDON, MD, FACC JOEL
KARLINER,
ALEXANDER STEVE JOHN
WALASKI, VERBA,
MATTHIAS KIRK
MD, FACC
BATTLER, PhD
PFISTERER,
PETERSON,
WILLIAM
MD
RT
MD,
ASHBURN,
MD FACC
MD
San Diego, California
To identify abnormal left ventricular function without exercise stress in patients with coron ry artery disease first-pass radionuclide angiograms were analyzed in 3 h normal subjects (Group I); 31 patients with coronary disease and normal contrast ventriculograms (Group II); and 17 patients with coronary disease and depressed left ventricular function (Group III). Total ejection fraction (EF) was computed with standard angiographic methods and from each time-activity curve. During the first third of systole, ejection fraction was determined manually by averaging three to five beats and the value compared with that obtained with contrast ventriculography: Contrast Group
Total EF
113 EF
I
-
-
II III
0.64 f 0.06 0.42 i 0.09’
0.23 f 0.04 0.16 f 0.04
Radionucltde Total EF 113 EF 0.62 f 0.04 0.59 f 0.05 0.34 f 0.09’
0.29 f 0.04 0.16 f 0.04* 0.13 f 0.04*
P < 0.01 versus II; + P < 0.01 versus I and II. * P < 0.01 versus I. l
Both total radionuclide ejection fraction (r = 0.95) and first-third ejection fraction (r = 0.91) correlated well with angiography. Intraobserver and interobserver variation was small, averaging 0.02 f 0.02 (range 0 to 0.05). The radionuclide first-third ejection fraction was 0.25 or greater in normal subjects and less than 0.25 in 29 of 31 patients (94 percent) in Group II and in all patients in Group III. It is concluded that the first-third ejection fraction obtained with first pass angiography identifies subtle abnormalities of left ventricular function at rest in more than 90 percent of patients with coronary disease that may not be recognized by total ejection fraction alone.
From the Divisions of Cardiology and Nuclear Medicine, University of California, San Diego, California. This study was supported by the Specialized Center of Research on lschemic Heart Disease, National Institutes of Health Research Grant HL-17662 awarded by the National Heart, Lung, and Blodd Institute, Bethesda Maryland. Manuscript received December 12, 1976; revised manuscript received April 4. 1979, accepted April 11, 1979. Address for reprints: Robert Slutsky, MD, University of California San Diego Medical Center, 225 Dickinson Street, San Diego, California 92103.
Our experience and that of other investigators indicate that one third or more of patients with chronic ischemic heart disease have a normal left ventricular ejection fraction at rest. Therefore, stress exercise studies have been developed to produce changes characteristic of ischemia shown by alterations in the S-T segment,l myocardial scintigraphy” or changes in ejection fraction.s-” However, Leighton et a1.6 and Johnson et a1.7 using contrast ventriculography were among the first investigators to suggest that alterations in early ejection phase indexes and wall motion might be a more sensitive indicator of ventricular dysfunction (thus suggesting the possibility of coronary artery disease) than the conventional ejection fraction. Accordingly, in this study we compared the sensitivity for detection of coronary artery disease of the ejection fraction determined at rest with first pass radionuclide angiography and that of the fraction ejected during the first one third of systole. September 1979
The American Journal of CARDIOLOGY
Volume 44
459
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY
ET AL.
Methods
(range 23 to 66). All had normal chest X-ray films and electrocardiograms; 2. Group II was composed of 31 patients with a history of
With the method described by Schelbert et al.,a 80 subjects underwent first pass radionuclide angiography. These 80 subjects were separated into three groups: 1. Group I consisted of 32 norm@ subjects. Twenty-two were normal volunteers who were not taking medication and had no evidence of cardiopulmonary disease by history or physical examination. Ten subjects in this group underwent contrast angiography for complaints of chest pain and were found to be free of cardiac disease. No patient had received nitroglycerin or propranolol within 96 hours of the study. There were 11 women and 21 men with a mean age of 48 years
chest pain whr, underwent diagnostic cardiac catheterization.
Patient Groups
Each of these patients had at least one significant area of greater than 70 percent coronary arterial stenosis and normal ejection phase indexes of left ventricular performance, including ejection fraction, mean velocity of circumferential fiber shortening and mean normalized systolic ejection rate assessed with conventional contrast angiography. Nine patients had been taking propranolol, but this agent was discontinued 96 hours before the study. None were taking digitalis and none had valve disease. This group was composed of 9 women and 22 men with a mean age of 52 years (range 33
COUNTS 130+ 120s 11Oi
f
loot 90+
* *
*
*
*
%*
**
so+** 70+
**
60+ 50+
*
**
*
*
* **
*
*
* * *
l
4ot
*
**
*
*
3ot
l
2ot
** * *
*
*
*
l
*
l
*
*f ‘r *
*
l *
**
* l
5‘
* *
*
;** * *** ** * ** * ** * ** *, ** * r’ $*
*
*
l
* *
* *
* ** * * ***
* *
1
**
*
l*
* * *
b*
*
**
+2
l
R
l
NORMAL SUBJECT EF = .71
*
** t l*
lo+
t
o+
+
+
+
+
+
TIME COUNTS 110+
PATIENT WITH CORONARY ARTERY DISEASE EF = .i’l
so+* 70+ * 60+
* * l l
50: 40+ 30+
*
*
** ** **
** ** f
10+ 0+
+
+
+
TIME
469
September
1979
The Amerkan
Journal
of CARDIOLDGY
Volume
44
FIGURE 1. Typical time-activity curves from two subjects after a single bolus dose of radioactive tracer through the left ventricle. Both subjects had a normal ejection fraction (EF).
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY ET AL.
to 64). No patient had historical or electrocardiographic evidence of a previous infarct. 3. Group III consisted of 17 patients with coronary artery disease as previously defined. All 17 had depressed ejection phase indexes (ejection fraction less than 0.56; mean velocity of circumferential fiber shortening less than 1.00 circumferencelsec; mean normalized systolic ejection rate less than 2 vols/sec). Thirteen of these patients were taking digoxin (mean serum level 1.34 f 0.19 ng/ml [* standard deviation]). No patient was receiving antihypertensive therapy or propranolol and all had normal blood pressure at the time of study. This group consisted of 15 men and 2 women with a mean age of 53 years (range 35 to 75). All subjects gave written informed consent for study according to a protocol approved by The Committee on Human Investigations Using Human Subjects and Volunteers, University of California, San Diego. Radionucllde Angiography Each subject underwent first pass radionuclide angiography performed using 14 to 20 miC of technetium-99m (ggmTc) bound to human serum albumin. The studies were made 1 to 24 hours before cardiac catheterization in all patients in Groups II and III and in the 10 patients in Group I who were catheterized. The data were collected using a mobile single crystal scintillation camera equipped with a general purpose parallel hole collimator and stored on digital videotape. The data were then transferred to a General Electric Med II Computer System and processed using a previously described method.8 In brief, a light pen assigned region-of-interest surrounding the left ventricle is used to monitor counts at 0.04 second intervals during the transit of the radioactive bolus dose through the chamber. Background correction is made and a plot is produced showing the changes in count rate during approximately 10 to 15 heartbeats. Ejection fraction: The high frequency curve is characterized by cyclic alterations in count rate between end-diastole (high points) and end-systole (subsequent low points) from which ejection fraction may be calculated by the formula Ejection fraction = (End-diastolic counts - End-systolic counts)/End-diastolic counts. After each left ventricular high frequency time-activity curve was printed (Fig. l), selected portions of the downward slopes of the curve were identified for beat by beat analysis. Three to six beats were then analyzed. End-diastole was chosen as the data point with the highest number of counts just before ejection. End-systole was taken as the data point with the smallest number of counts. Beats were excluded when the end-diastolic or end-systolic data were not clearly apparent, when the atria1 wave was not visualized or when one definite single peak was not clearly identified. This usually led to the elimination of approximately one out of every five beats. The distance between these two points was defined as the left ventricular ejection time, and was further divided into thirds (Fig. 2).
variability is due to the statistical error around the end-systolic and end-diastolic data points, which range from 30 to 150 counts/frame. Calculation of the first-third ejection fraction utilizes a manual method. The raw data curves undergo a 5 point curve smoothing, and then 7 to 10 beats are plotted. Only beats on the downslope of the left ventricular time-activity curve are utilized, thus eliminating the recirculation data. The statistical variation with this manual calculation is higher than that utilizing a root mean square method. The error in this method can be approximated by assuming the randomness of counts and employing the “law of propogation of error.“‘OThe standard error (SE) for one heartbeat is: SE
I(Csc+ B) + Cs2(CDG 1
_ n CDL
,-..l I;D‘
+ B)]1/2 1
where Cso = gross systolic counts (at first third of systole); Coo = gross diastolic counts; Cs = background corrected counts at the first third of systole; and CD = back-ground corrected end-diastolic counts. Because three to six beats were used in the calculation the resultant standard error is: SE(I) SE (n) = x where SE (n) is the error using N number of heartbeats. The CD counts ranged from 80 to 140 and the Cso from 40 to 80, with the background assumed to be 10 counts. For normal subjects, in whom the first-third ejection fraction is 0.25 or more, the mean statistical variation is just under 0.05 ejection fraction units using three to six beats. For patients with a first-third ejection fraction of 0.25 or less the mean statistical variation is approximately 0.058 ejection fraction units. Contrast ventriculography: Each biplane contrast ventriculogram was used to calculate total ejection fraction and ejection fraction during the first third of systole. The enddiastolic frame was chosen as either: (1) the frame before visualized expulsion of blood into the aorta; (2) the frame before inward motion of the ventricular walls; or (3) the frame corresponding to 0.04 second after the peak of the QRS complex on a simultaneous pulse counter. End-systole was chosen as the smallest silhouette determined with planimetry. The total number of frames was counted and the volume at completion of the first third of systole was then determined.
Normal
Abnormal
1
1FL
!‘i !
f i i i/
The ejection fraction during the first one third of systole was calculated using a manual method and the following formula: (End-diastolic counts - Counts at I/3 of systole)-
/End-diastolic counts. In each patient heart rate and systolic blood pressure were measured and compared. The total radionuclide ejection fraction was also calculated using the computer program mentioned previous1y.s Statistical evaluation of the ejection fraction with this method has previously been discussed.g A root-mean square curve smoothing technique is utilized to minimize the ejection fraction variability. Five to 10 beats are utilized and the variation around an ejection fraction of 0.50 is f 0.02 ejection fraction units. The
=
1
I
I/3 ET- r-L +
l/3 ET*_
ET TOTAL EF q .71 FIRST i/3 EF = .31
TIME
ET
TOTAL EF = .71 FIRST l/3 EF = .16
FIGURE 2. A selected beat from a time activity curve in two subjects. Note the delayed ejection in the patient with coronary artery disease (right). ED = enddiastole; EF = ejection fraction; ES = end-systole; ET = ejection time: l/3 ET = first third of ejection time.
September lB79
The American Journal of CARDKXCGY
Volume 44
461
EARLY VENTRICULAR SYSTDLE AND CORONARY DISEASE-SLUTSKY
Premature or postpremature beats were omitted and at least two beats were averaged. All patients had normal sinus rhythm at the time of both the radionuclide and the contrast ventriculographic studies. Statistical analysis and reproducibility: Statistical comparisons among the three patient groups were carried out using a repeated measures analysis of variance method.” Reproducibility was studied by comparison of the results of recalculation of the same data in 15 cases by the same observer 3 months apart. The interobserver variation was assessed by having a second observer perform these 15 studies without knowledge of the results of the first observer. Ten patients had repeat studies performed 7 days apart to assess reproducibility over time. Additionally we analyzed the beat to beat variation in each of the 385 beats evaluated.
Results Hemodynamic data: Subjects in Group I (normal) and Group II (coronary disease) all had a normal total ejection fraction (0.52 or greater) on first pass radionuclide angiography, and their mean values were not significantly different from each other (0.62 f 0.04 [standard deviation] versus 0.59 f 0.05) (Table I and Fig. 3). Values in Group III (coronary disease with abnormal ventriculograms) were significantly lower than in the other groups (0.34 f 0.09, P < 0.01) (Table I).
TABLE I Ejection Fraction and First One Third Ejection Fraction in Each Patient Group Case no.
Mean SD
EF
Group I ll3EF
0.61 0.74 0.68 0.60 0.65 0.54 0.65 0.53 0.71 0.52 0.65 0.59 0.66 0.71 0.55 0.58 0.66 0.78 0.54 0.55 0.63 0.60 0.66 0.71 0.58 0.54 0.63 0.54 0.79 0.81 0.79 0.50 0.62 0.04
0.28 0.35 0.26 0.29 0.25 0.36 0.25 0.31 0.25 0.25 0.26 0.29 0.33 0.36 0.33 0.25 0.31 0.29 0.25 0.29 0.34 0.31 0.30 0.28 0.27 0.26 0.26 0.29 0.31 0.30 0.29 0.32 0.29 0.04
EF
Group II ll3EF
0.47 0.62 0.52 0.57 0.54 0.55 0.57 0.58 0.54 0.53 0.55 0.71 0.58 0.76 0.65 0.53 0.60 0.58 0.62 0.64 0.71 0.53 0.57 0.58 0.60 0.68 0.52 0.64 0.53 0.56 0.67 0.59 0.05
0.19 0.17 0.23 0.16 0.14 0.18 0.17 0.22 0.19 0.19 0.17 0.17 0.22 0.27 0.19 0.14 0.24 0.14 0.17 0.21 0.29 0.17 0.14 0.19 0.14 0.15 0.10 0.12 0.10 0.11 0.13 0.18 0.04
Group III EF l/3 EF 0.30 0.43 0.44 0.47 0.46 0.24 0.40 0.37 0.37 0.21 0.23 0.26 0.19 0.22 0.17 0.30 0.38
0.15 0.19 0.16 0.19 0.17 0.09 0.19 0.15 0.06 0.04 0.12 0.09 0.07 0.10 0.07 0.12 0.15
September 1979
The Amerban
All patients in Group I who underwent angiography had a left ventricular end-diastolic pressure of 12 mm Hg or less (mean 7 f 3 mm Hg). Twenty-six of 32 patients in Group II had a left ventricular end-diastolic pressure of less than 12 mm Hg (mean 9.1 f 2.4 mm Hg) (difference from values in Group I not significant). Fourteen of 17 patients in Group III had a left ventricular end-diastolic pressure more than li mm Hg (mean 18.2 f 6.1) (P< 0.05 versus Groups I and II). Patients in Groups II and III had a lower mean radionuclide first-third ejection fraction (0.18 f 0.04 and 0.13 f 0.04, respectively) than the value in Group I (0.29 f 0.04) (both P < 0.01). The mean value in Group II was significantly different (P< 0.05) from that in Group III (Table I). The lower limit of normal for first-third ejection fraction, as established by the values in Group I, was chosen as 0.25. This value was exceeded in only two patients in Group II and in no patient in Group III. All data were compared using analysis of variance.” Additionally, all normal persons who underwent contrast ventriculography had a first-third ejection fraction by this method of 0.28 or greater, and 45 of 48 patients in Groups II and III had an early ejection fraction of 0.28 or less, thus confirming the radionuclide results. Correlation with contrast ventriculography: In the 58 patients with angiographic data the correlation between the two methods for total ejection fraction was 0.95; for first-third ejection fraction it was 0.91 (Fig. 4 and 5). As reported in previous studies,8 the angiographic ejection fraction tended to be slightly higher than the radionuclide ejection fraction. The average difference between the two techniques was 0.05 for the total ejection fraction and 0.06 for the first-third ejection fraction (l/3 EF). Utility of the method: To determine the usefulness of assessing left ventricular function during early systole
GROUP
I
n = 32
0.34 0.09
NOw4t SUBJECTS
0.13 0.04
EF = ejection fraction; Group I = normal subjects; Group II = patients with coronary artery disease and a normal ejection fraction; Group Ill = patients with coronary artery disease and a depressed ejection fraction; SD = standard deviation.
462
ET AL
Journal of CARDIOLOGY
EF = .BZ f .04
GROUP
I
n = 31 I
Wllll NORMIt “JI”,ON Fk4CVON EF = .59 f .OE
GROUP
dl
n =‘17 ,
E.WliON
FRWlON
EF=.Yi.O#
FIGURE 3. First-third ejection fraction plotted for each patient in the three qoups. EF = total ejection fraction (normal 0.52). See Table I for individual results.
Volume 44
EARLY
with radionuclide angiography, the following standard calculations12 were also made: 1.
VENTRICULAR
=
= 32132or 100 percent. 2. Sensitivity
=
Number of patients with coronary disease and reduced l/3 EF (<0.25) Total number of patients with coronary disease = 46/48 or 96 percent.
In Group II, the specificity was 94 percent (29 of 31). One patient with a one third ejection fraction of 0.27 had severe three vessel coronary artery disease, but he also had a hypercontractile left ventricle and an ejection fraction of 0.85 on contrast ventriculography. The other patient with a normal first-third ejection fraction (0.29) had a single 70 percent stenotic lesion of the left anterior descending coronary artery. 3. Predictive accuracy = True positives (coronary disease and l/3 EF < 0.25) All positives = 46/46 = 106 percent. There were 2 false negatives (2/48 or 4 percent) and no false positives. Reproducibility of method: For first-third ejection fraction, the intraobserver variation was small, with a mean difference of 0.017 ejection fraction unit (range 0 to 0.04) when 15 randomly selected studies were evaluated 2 months later using the same previously recorded raw data. A second observer performed the first-third ejection fraction calculations without
CORONARY
DISEASE-SLUTSKY
ET AL.
Heart Rate and Blood Pressure in the Three Patlent Groups
Group
Total number of normal subjects
AND
TABLE II
Specificity Number of normal subjects with normal l/3 EF (3 0.25)
SYSTDLE
II III
Heart Rate (beatslmin) fS:: 71f
11
Systolic Blood Pressure (mm Hg) 118f 116 f 121 f
13 14 11
First Pass Ejection Fraction
First-Third Ejection Fraction
0.59 0.62 f 0.04 0.05 0.34 f 0.09
0.29 f 0.04 0.18 0.13 f 0.04
All values are expressed as mean f standard deviation.
knowledge of the results of the first observer. Interobserver variation was modest with a mean difference of 0.023 ejection fraction unit (range 0 to 0.05). Ten patients were studied twice, 7 days apart. The mean difference was 0.03 ejection fraction unit (range 0 to 0.05). Mean beat to beat variation was 0.03 f 0.025 ejection fraction unit (range 0 - 0.10) in 385 studies. The greatest errors were seen in the patients with a very low (less than 0.10) first-third ejection fraction. Vital signs: There was no significant difference (as assessed with analysis of variance) among Groups I, II and III with respect to heart rate or blood pressure (Table II).
Discussion Rushmer13 suggested in the mid 1960s that evaluation of the early phase of the ventricular impulse might prove fruitful in assessing subtle abnormalities of left ventricular function. Gault et aLi4 showed that peak depression in instantaneous velocity of circumferential fiber shortening (Vcf) occurred early in systole, at a time when peak wall stress was recorded. Some investigators15J6 used peak aortic flow and ventricular acceleration to assess left ventricular performance in dogs, whereas others6,7 suggested that assessment of the early phases of left ventricular ejection may yield information not available when only end-systolic data are compared
.50
COMPARSION OF CONTRASTANO AAOIDNUCLIOE EJECTIONFRACTIONS
.
'0
FIRST l/3 EJECTION FRACTION FIRST-PASS flAOlONUCLlDE ANGIOGRAPHY
FKSURE 4. Comparison of contrast and radionuclide first-third ejection fraction in 58 subjects. n = number of subjects: r = correlation coefficient.
FIRST-PASS RADtONUCLlDE EJECTION FRACllON
FIGURE 5. Comparison of contrast and radionuclide ejection fraction in 58 subjects. Abbreviations as in Figure 4.
September 1979
The American Journal of CARWOLGGY
Volume 44
463
EARLY VENTRICULAR SYSTOLF
AND CORONARY DISEASE-SLUTSKY
with end-diastolic measurements. By contrast, Hammermeister et al.17 were unable to differentiate normal from abnormal patients when the rate of change of left ventricular volume was analyzed. However, peak flows during differing phases of ejection were not analyzed. Depressed first-third ejection fraction as an index of impaired left ventricular function: In this study we used first-pass radionuclide angiography to test the hypothesis that patients with coronary artery disease exhibit impairment of early ventricular ejection. In this study, 46 of 48 patients (96 percent) with coronary artery disease had a depressed ejection fraction during the first-third of systole, although 31 had a normal total ejection fraction on both contrast and radionuclide angiography. We were unable to define an abnormality in left ventricular function at rest in only two patients; one patient was somewhat atypical in that he had a “hypercontractile” left ventricle and the second patient had only single vessel coronary disease. Our data also suggest that the sensitivity, specificity and predictive accuracy of this approach support its utility in assessing left ventricular function in patients with known or suspected coronary artery disease, although this group of patients was selected for the presence or absence of coronary artery disease. We have found that this abnormality of ventricular ejection also occurs in other diseases, such as aortic stenosis,18 and Johnson et al.7 also found dysfunction in patients with cardiomyopathy. Clearly sensitivity, specificity and predictive accuracy will depend on patient selection. The correlation between angiographic and radionuclide tech-
10
4
.9 SF z s z .= 9
0.65
.8 0 60 .lL1.18 6-
=
.5-
2 t
.4-
I= J, 2
. 77 . 70
.22 20
0 57
w
g
A.30
A.2o
A 16
25
l .55
niques was excellent and the inter- and intraobserver variations appear to be modest. Reproducibility over time was also reasonably good. Many patients with a depressed ejection fraction had a similar (although depressed) first-third ejection fraction when compared with patients with a normal resting ejection fraction. This may limit the utility of the method in comparing patient groups. Cause of depressed early ejection fraction: The exact basis for this observed abnormality is unclear, One possibility is that patients with coronary artery disease may have decreased aortic compliance, which could reduce the amount of blood ejected early in systole.1g~20 Additionally, it has been shown that areas of abnormal contraction have increased amounts of fibrosis,21 which may lead to contractile abnormalities as a result of decreased developed wall force. Focal conduction abnormalities or chronically depressed blood flow might contribute to these contraction abnormalities, although this would seem less likely. Potential limitations of method: It appears that early ejection is a useful period to study, and based on our preliminary observations can often distinguish normal patients from patients with cardiac disease without utilizing an exercise stress.3l4 The method is simple to perform, nontraumatic and reproducible. The major pitfalls in utilizing this method would include studying patients with such severely depressed left ventricular function that end-diastole cannot be confidently determined or patients with a regurgitant mitral valve in whom a low impedance leak might occur. Some problems may be due to the higher statistical error around lower ejection fractions, and difficulty with defining the end-diastolic point. Additional theoretical limitations include statistical problems due to difficulties with the accuracy of the counts at the enddiastolic and end-systolic data points, and the occasional patients with curves that are too noisy for legitimate analysis. Clearly, with larger patient populations and a greater splay in the normal population, the sensitivity, specificity and predictive accuracy will change (and probably decrease). Part of this may be resolvable with computer analysis, higher count rate acquisition and shorter frame time. The theoretical
l EJECTION
FR4CllOM
VALUES
A FIRSl.~Hl~EJECTIONFR4C~ON VALUES
0.50.52 .3"45
.2 -
.l -
.l
.2
.3
.4
.5
.6
.7
.6
I
1
.S
1.0
1 SPECIFICITY(FALSE POSITIVEFRACTION) FIGURE 6. The sensitivity of the ejection fraction and first-third ejectii fraction compared with the percent of false posltive studies at several threshold levels of normal. A false positive fraction is defined as: Number of false positives/Number of actually negative cases or lspecificity. Note that by lowering the cutoff point for a normal first-thlrd ejection fraction to 0.20 the sensitivity level falls to 85 percent, although the false positive fraction remains unchanged.
464
ET AL.
September 1979
7he Arnerkan Journal d CARMOLOGY
limitations of this first pass method
are based on the statistical variability of the results and the patient populations selected. Figure 6 depicts a curve that compares the sensitivity of a test with the fraction of false positive tests that results when a variety of thresholds are compared. In this graph one can see that the sensitivity of the first-third ejection for coronary disease in this population is well over 80 percent at a strict threshold for normal of 0.20 or greater; at this threshold few false positives occur. With lower normal ranges, sensitivity decreases and with more moderate or lax thresholds, the sensitivity increases but the proportion of false positive studies increases.22 The conventional ejection fraction has a high sensitivity for coronary disease only when a high false positive fraction is accepted. Thus, in general, future analysis of early systolic ejection by this method has significant potential
Volume 44
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY
variability in sensitivity depending on the threshold accepted as normal. If one excludes the patients with a depressed conventional ejection fraction, the sensitivity at a threshold of 0.20 decreases to approximately 75 percent. Similar analysis with gated equilibrium radionuclide angiography using composite time-activity curves in our laboratory has not been fruitful. Correlation of first-third equilibrium and contrast angiographic ejection has been poor (r = 0.65 or less) ‘resulting, as would be expected, in much less patient separation (23). Thus, noninvasive analysis is restricted to first pass methods in our laboratory.
ET AL.
Clinical implication: First-third ejection fraction obtained with first pass radionuclide angiography is an important new observation that permits identification of subtle abnormalities of left ventricular function at rest in patients with coronary artery disease that may not be recognized by analysis of total ejection fraction alone. At present caution must be observed; it will be important for other studies to be performed to confirm these observations and validate their reproducibility before the measurement can become an integral part of the clinical evaluation of the patient with a chest pain syndrome.
References 1. Bruce RA: Method of exercise testing: step testing, treadmill and isometrics. In Exercise in Cardiovascular Disease and Health (Amsterdam E, Wilmore J, DeMaria A, ed). New York, Yorke Medical Books, 1977, p 149-158 2. Berman D, Amsterdam EA, Mason DT: Detection of myocardial ischemia by rest and exercise thallium-201 scintigraphy. In Ref 1, p 24-31 3. Borer JB, Badhrach !X, Green MV, Kent KM, Epstein SE, Johnston GS: Real time-radionuclide cineangiography In the noninvasive evaluation of global and regional left ventricular function at rest and during exercise in patients with coronary artery disease. N Engl J Med 296:839-8451977 4. Pfieterer M, S&&r G, WCC/ D, Swanson S, Gordon D, Slutsky R, Peterson K, Ashburn W: Profiles of left ventricular ejection by equilibrium radionuclide angiography during exercise and in the recovery period between normals and patients with CAD (abstr). J Nucl Med 19:710, 1978 5. Bfutaky RA, Kartluer J, Rkcl D, Schuler G, Pflsterer M, Peterson K, Ashbum W: Response of left ventricular volume to exercise in normals and!patients with angina pectoris as calculated by radionucltde angiography (abstr). J Nucl Med 19:7 11, 1978 6. LeIghton R, Polla& ME, Wekh T6: Abnormal left ventricular wall motion at mid-ejection in patients with coronary heart disease. Circulation 52:238-248, 1975 7. Johnson L, Etgs K, Bchmldt D, Weiss M, Cannon P: Volume ejected in early systole. Circulation 52:378-386, 1975 8. ScheMerI H, Verba J, Johnson AD, Block G, Alazrakl N, Rose F, Ashburn W: Non-traumatic determination of left ventriculation ejection fraction of radionuclide angiography. Circulation 51: 902-909, 1975 9. Ashburn W, Verba J, Schelbert H: Left ventricular ejection fraction-A review of several radionuclide angiographic approaches using the scintillation camera. Prog Cardiovasc Dis 20:267-284, 1978 10. Bevington PR: Data Reduction and Erorr Analysis for Physical Sciences. New York, McGraw-Hill, 1989, p 60-78
Il.
12. 13. 14. 15. 16.
17. 18.
19.
20.
2 1.
22. 23.
September 1979
Winer BJ: Multifactorial experiments having repeat measures on the same element. In, Statistical Principles in Experimental Design. New York, McGraw-Hill, 1971, p 570-595 Froelicher VF: The detection of asymptomatic coronary artery disease. Annu Rev f&d 28:1-12, 1977 Rushmer R: Initial ventricular impulse. Circulation 29:269-280, 1964 GauH J, Ross J Jr, Braunwatd E: Contractile state of the left ventricle in man. Circ Res 22:457-469, 1966 Nobte MIM: The contribution of blood momentum to left ventricular ejection in the dog. Circ Res 23:663-871, 1968 Nutter W, Noble JR, Hurst VW: Peak aortic flow and acceleration as indices of ventricular performance in the dog. J Lab Clin Med 77:307-313,197l Harnmermefsier KE, Brooks BC, Warbasse JR: The rate of change of left ventricular volume in man. Circulation 49:729-738, 1974 Slutsky R, Karliner J, Battier A, Peterson K, Johnson A, Ross J Jr: The first-third ejection fraction in patients with aortic stenosis with and without coronary disease (abstr). Clin Res 27:204, 1979 Urschel C, Cove8 J, Bonnenblkk E, Ross J, Braunwatd E: Effects of decreased aortic contpliance on performance of the left ventricle. Am J Physiol 214:298-303, 1988 Wilcken D, Charlier A, Hoffman J, Guz A: Effects of alterations in aortic impedance on the performance of the ventricles: Circ Res 14:283-289, 1984 better R, Behar V, Wagner 0, Start J, Starmer F, Lee K, Hackel D: Evaluation of asynergy as an indicator of myocardial fibrosis. Circulation 57:715-725, 1978 Metr C: Basic principles of ROC analysis. Semin Nucl f&d 8: 283-292, 1978 Gordon D, Ashburn A, Stutsky R: Radionuclide angiography: first-pass method of assessing ejection fraction. In. Nuclear Cardiology (Mason D. Berman D, ed). New York, Yorke Medial Books, 1979
The Amertcan Journal of CARDtDLDGY
Volume 44
465
Assessment of Early Ventricular Systole by First Pass Radionuclide Angiography: Useful Method for Detection of Left Ventricular Dysfunction at Rest in Patients With Coronary Artery Disease
ROBERT SLUTSKY, MD DONALD GORDON, MD, FACC JOEL
KARLINER,
ALEXANDER STEVE JOHN
WALASKI, VERBA,
MATTHIAS KIRK
MD, FACC
BATTLER, PhD
PFISTERER,
PETERSON,
WILLIAM
MD
RT
MD,
ASHBURN,
MD FACC
MD
San Diego, California
To identify abnormal left ventricular function without exercise stress in patients with coron ry artery disease first-pass radionuclide angiograms were analyzed in 3 h normal subjects (Group I); 31 patients with coronary disease and normal contrast ventriculograms (Group II); and 17 patients with coronary disease and depressed left ventricular function (Group III). Total ejection fraction (EF) was computed with standard angiographic methods and from each time-activity curve. During the first third of systole, ejection fraction was determined manually by averaging three to five beats and the value compared with that obtained with contrast ventriculography: Contrast Group
Total EF
113 EF
I
-
-
II III
0.64 f 0.06 0.42 i 0.09’
0.23 f 0.04 0.16 f 0.04
Radionucltde Total EF 113 EF 0.62 f 0.04 0.59 f 0.05 0.34 f 0.09’
0.29 f 0.04 0.16 f 0.04* 0.13 f 0.04*
P < 0.01 versus II; + P < 0.01 versus I and II. * P < 0.01 versus I. l
Both total radionuclide ejection fraction (r = 0.95) and first-third ejection fraction (r = 0.91) correlated well with angiography. Intraobserver and interobserver variation was small, averaging 0.02 f 0.02 (range 0 to 0.05). The radionuclide first-third ejection fraction was 0.25 or greater in normal subjects and less than 0.25 in 29 of 31 patients (94 percent) in Group II and in all patients in Group III. It is concluded that the first-third ejection fraction obtained with first pass angiography identifies subtle abnormalities of left ventricular function at rest in more than 90 percent of patients with coronary disease that may not be recognized by total ejection fraction alone.
From the Divisions of Cardiology and Nuclear Medicine, University of California, San Diego, California. This study was supported by the Specialized Center of Research on lschemic Heart Disease, National Institutes of Health Research Grant HL-17662 awarded by the National Heart, Lung, and Blodd Institute, Bethesda Maryland. Manuscript received December 12, 1976; revised manuscript received April 4. 1979, accepted April 11, 1979. Address for reprints: Robert Slutsky, MD, University of California San Diego Medical Center, 225 Dickinson Street, San Diego, California 92103.
Our experience and that of other investigators indicate that one third or more of patients with chronic ischemic heart disease have a normal left ventricular ejection fraction at rest. Therefore, stress exercise studies have been developed to produce changes characteristic of ischemia shown by alterations in the S-T segment,l myocardial scintigraphy” or changes in ejection fraction.s-” However, Leighton et a1.6 and Johnson et a1.7 using contrast ventriculography were among the first investigators to suggest that alterations in early ejection phase indexes and wall motion might be a more sensitive indicator of ventricular dysfunction (thus suggesting the possibility of coronary artery disease) than the conventional ejection fraction. Accordingly, in this study we compared the sensitivity for detection of coronary artery disease of the ejection fraction determined at rest with first pass radionuclide angiography and that of the fraction ejected during the first one third of systole. September 1979
The American Journal of CARDIOLOGY
Volume 44
459
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY
ET AL.
Methods
(range 23 to 66). All had normal chest X-ray films and electrocardiograms; 2. Group II was composed of 31 patients with a history of
With the method described by Schelbert et al.,a 80 subjects underwent first pass radionuclide angiography. These 80 subjects were separated into three groups: 1. Group I consisted of 32 norm@ subjects. Twenty-two were normal volunteers who were not taking medication and had no evidence of cardiopulmonary disease by history or physical examination. Ten subjects in this group underwent contrast angiography for complaints of chest pain and were found to be free of cardiac disease. No patient had received nitroglycerin or propranolol within 96 hours of the study. There were 11 women and 21 men with a mean age of 48 years
chest pain whr, underwent diagnostic cardiac catheterization.
Patient Groups
Each of these patients had at least one significant area of greater than 70 percent coronary arterial stenosis and normal ejection phase indexes of left ventricular performance, including ejection fraction, mean velocity of circumferential fiber shortening and mean normalized systolic ejection rate assessed with conventional contrast angiography. Nine patients had been taking propranolol, but this agent was discontinued 96 hours before the study. None were taking digitalis and none had valve disease. This group was composed of 9 women and 22 men with a mean age of 52 years (range 33
COUNTS 130+ 120s 11Oi
f
loot 90+
* *
*
*
*
%*
**
so+** 70+
**
60+ 50+
*
**
*
*
* **
*
*
* * *
l
4ot
*
**
*
*
3ot
l
2ot
** * *
*
*
*
l
*
l
*
*f ‘r *
*
l *
**
* l
5‘
* *
*
;** * *** ** * ** * ** * ** *, ** * r’ $*
*
*
l
* *
* *
* ** * * ***
* *
1
**
*
l*
* * *
b*
*
**
+2
l
R
l
NORMAL SUBJECT EF = .71
*
** t l*
lo+
t
o+
+
+
+
+
+
TIME COUNTS 110+
PATIENT WITH CORONARY ARTERY DISEASE EF = .i’l
so+* 70+ * 60+
* * l l
50: 40+ 30+
*
*
** ** **
** ** f
10+ 0+
+
+
+
TIME
469
September
1979
The Amerkan
Journal
of CARDIOLDGY
Volume
44
FIGURE 1. Typical time-activity curves from two subjects after a single bolus dose of radioactive tracer through the left ventricle. Both subjects had a normal ejection fraction (EF).
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY ET AL.
to 64). No patient had historical or electrocardiographic evidence of a previous infarct. 3. Group III consisted of 17 patients with coronary artery disease as previously defined. All 17 had depressed ejection phase indexes (ejection fraction less than 0.56; mean velocity of circumferential fiber shortening less than 1.00 circumferencelsec; mean normalized systolic ejection rate less than 2 vols/sec). Thirteen of these patients were taking digoxin (mean serum level 1.34 f 0.19 ng/ml [* standard deviation]). No patient was receiving antihypertensive therapy or propranolol and all had normal blood pressure at the time of study. This group consisted of 15 men and 2 women with a mean age of 53 years (range 35 to 75). All subjects gave written informed consent for study according to a protocol approved by The Committee on Human Investigations Using Human Subjects and Volunteers, University of California, San Diego. Radionucllde Angiography Each subject underwent first pass radionuclide angiography performed using 14 to 20 miC of technetium-99m (ggmTc) bound to human serum albumin. The studies were made 1 to 24 hours before cardiac catheterization in all patients in Groups II and III and in the 10 patients in Group I who were catheterized. The data were collected using a mobile single crystal scintillation camera equipped with a general purpose parallel hole collimator and stored on digital videotape. The data were then transferred to a General Electric Med II Computer System and processed using a previously described method.8 In brief, a light pen assigned region-of-interest surrounding the left ventricle is used to monitor counts at 0.04 second intervals during the transit of the radioactive bolus dose through the chamber. Background correction is made and a plot is produced showing the changes in count rate during approximately 10 to 15 heartbeats. Ejection fraction: The high frequency curve is characterized by cyclic alterations in count rate between end-diastole (high points) and end-systole (subsequent low points) from which ejection fraction may be calculated by the formula Ejection fraction = (End-diastolic counts - End-systolic counts)/End-diastolic counts. After each left ventricular high frequency time-activity curve was printed (Fig. l), selected portions of the downward slopes of the curve were identified for beat by beat analysis. Three to six beats were then analyzed. End-diastole was chosen as the data point with the highest number of counts just before ejection. End-systole was taken as the data point with the smallest number of counts. Beats were excluded when the end-diastolic or end-systolic data were not clearly apparent, when the atria1 wave was not visualized or when one definite single peak was not clearly identified. This usually led to the elimination of approximately one out of every five beats. The distance between these two points was defined as the left ventricular ejection time, and was further divided into thirds (Fig. 2).
variability is due to the statistical error around the end-systolic and end-diastolic data points, which range from 30 to 150 counts/frame. Calculation of the first-third ejection fraction utilizes a manual method. The raw data curves undergo a 5 point curve smoothing, and then 7 to 10 beats are plotted. Only beats on the downslope of the left ventricular time-activity curve are utilized, thus eliminating the recirculation data. The statistical variation with this manual calculation is higher than that utilizing a root mean square method. The error in this method can be approximated by assuming the randomness of counts and employing the “law of propogation of error.“‘OThe standard error (SE) for one heartbeat is: SE
I(Csc+ B) + Cs2(CDG 1
_ n CDL
,-..l I;D‘
+ B)]1/2 1
where Cso = gross systolic counts (at first third of systole); Coo = gross diastolic counts; Cs = background corrected counts at the first third of systole; and CD = back-ground corrected end-diastolic counts. Because three to six beats were used in the calculation the resultant standard error is: SE(I) SE (n) = x where SE (n) is the error using N number of heartbeats. The CD counts ranged from 80 to 140 and the Cso from 40 to 80, with the background assumed to be 10 counts. For normal subjects, in whom the first-third ejection fraction is 0.25 or more, the mean statistical variation is just under 0.05 ejection fraction units using three to six beats. For patients with a first-third ejection fraction of 0.25 or less the mean statistical variation is approximately 0.058 ejection fraction units. Contrast ventriculography: Each biplane contrast ventriculogram was used to calculate total ejection fraction and ejection fraction during the first third of systole. The enddiastolic frame was chosen as either: (1) the frame before visualized expulsion of blood into the aorta; (2) the frame before inward motion of the ventricular walls; or (3) the frame corresponding to 0.04 second after the peak of the QRS complex on a simultaneous pulse counter. End-systole was chosen as the smallest silhouette determined with planimetry. The total number of frames was counted and the volume at completion of the first third of systole was then determined.
Normal
Abnormal
1
1FL
!‘i !
f i i i/
The ejection fraction during the first one third of systole was calculated using a manual method and the following formula: (End-diastolic counts - Counts at I/3 of systole)-
/End-diastolic counts. In each patient heart rate and systolic blood pressure were measured and compared. The total radionuclide ejection fraction was also calculated using the computer program mentioned previous1y.s Statistical evaluation of the ejection fraction with this method has previously been discussed.g A root-mean square curve smoothing technique is utilized to minimize the ejection fraction variability. Five to 10 beats are utilized and the variation around an ejection fraction of 0.50 is f 0.02 ejection fraction units. The
=
1
I
I/3 ET- r-L +
l/3 ET*_
ET TOTAL EF q .71 FIRST i/3 EF = .31
TIME
ET
TOTAL EF = .71 FIRST l/3 EF = .16
FIGURE 2. A selected beat from a time activity curve in two subjects. Note the delayed ejection in the patient with coronary artery disease (right). ED = enddiastole; EF = ejection fraction; ES = end-systole; ET = ejection time: l/3 ET = first third of ejection time.
September lB79
The American Journal of CARDKXCGY
Volume 44
461
EARLY VENTRICULAR SYSTDLE AND CORONARY DISEASE-SLUTSKY
Premature or postpremature beats were omitted and at least two beats were averaged. All patients had normal sinus rhythm at the time of both the radionuclide and the contrast ventriculographic studies. Statistical analysis and reproducibility: Statistical comparisons among the three patient groups were carried out using a repeated measures analysis of variance method.” Reproducibility was studied by comparison of the results of recalculation of the same data in 15 cases by the same observer 3 months apart. The interobserver variation was assessed by having a second observer perform these 15 studies without knowledge of the results of the first observer. Ten patients had repeat studies performed 7 days apart to assess reproducibility over time. Additionally we analyzed the beat to beat variation in each of the 385 beats evaluated.
Results Hemodynamic data: Subjects in Group I (normal) and Group II (coronary disease) all had a normal total ejection fraction (0.52 or greater) on first pass radionuclide angiography, and their mean values were not significantly different from each other (0.62 f 0.04 [standard deviation] versus 0.59 f 0.05) (Table I and Fig. 3). Values in Group III (coronary disease with abnormal ventriculograms) were significantly lower than in the other groups (0.34 f 0.09, P < 0.01) (Table I).
TABLE I Ejection Fraction and First One Third Ejection Fraction in Each Patient Group Case no.
Mean SD
EF
Group I ll3EF
0.61 0.74 0.68 0.60 0.65 0.54 0.65 0.53 0.71 0.52 0.65 0.59 0.66 0.71 0.55 0.58 0.66 0.78 0.54 0.55 0.63 0.60 0.66 0.71 0.58 0.54 0.63 0.54 0.79 0.81 0.79 0.50 0.62 0.04
0.28 0.35 0.26 0.29 0.25 0.36 0.25 0.31 0.25 0.25 0.26 0.29 0.33 0.36 0.33 0.25 0.31 0.29 0.25 0.29 0.34 0.31 0.30 0.28 0.27 0.26 0.26 0.29 0.31 0.30 0.29 0.32 0.29 0.04
EF
Group II ll3EF
0.47 0.62 0.52 0.57 0.54 0.55 0.57 0.58 0.54 0.53 0.55 0.71 0.58 0.76 0.65 0.53 0.60 0.58 0.62 0.64 0.71 0.53 0.57 0.58 0.60 0.68 0.52 0.64 0.53 0.56 0.67 0.59 0.05
0.19 0.17 0.23 0.16 0.14 0.18 0.17 0.22 0.19 0.19 0.17 0.17 0.22 0.27 0.19 0.14 0.24 0.14 0.17 0.21 0.29 0.17 0.14 0.19 0.14 0.15 0.10 0.12 0.10 0.11 0.13 0.18 0.04
Group III EF l/3 EF 0.30 0.43 0.44 0.47 0.46 0.24 0.40 0.37 0.37 0.21 0.23 0.26 0.19 0.22 0.17 0.30 0.38
0.15 0.19 0.16 0.19 0.17 0.09 0.19 0.15 0.06 0.04 0.12 0.09 0.07 0.10 0.07 0.12 0.15
September 1979
The Amerban
All patients in Group I who underwent angiography had a left ventricular end-diastolic pressure of 12 mm Hg or less (mean 7 f 3 mm Hg). Twenty-six of 32 patients in Group II had a left ventricular end-diastolic pressure of less than 12 mm Hg (mean 9.1 f 2.4 mm Hg) (difference from values in Group I not significant). Fourteen of 17 patients in Group III had a left ventricular end-diastolic pressure more than li mm Hg (mean 18.2 f 6.1) (P< 0.05 versus Groups I and II). Patients in Groups II and III had a lower mean radionuclide first-third ejection fraction (0.18 f 0.04 and 0.13 f 0.04, respectively) than the value in Group I (0.29 f 0.04) (both P < 0.01). The mean value in Group II was significantly different (P< 0.05) from that in Group III (Table I). The lower limit of normal for first-third ejection fraction, as established by the values in Group I, was chosen as 0.25. This value was exceeded in only two patients in Group II and in no patient in Group III. All data were compared using analysis of variance.” Additionally, all normal persons who underwent contrast ventriculography had a first-third ejection fraction by this method of 0.28 or greater, and 45 of 48 patients in Groups II and III had an early ejection fraction of 0.28 or less, thus confirming the radionuclide results. Correlation with contrast ventriculography: In the 58 patients with angiographic data the correlation between the two methods for total ejection fraction was 0.95; for first-third ejection fraction it was 0.91 (Fig. 4 and 5). As reported in previous studies,8 the angiographic ejection fraction tended to be slightly higher than the radionuclide ejection fraction. The average difference between the two techniques was 0.05 for the total ejection fraction and 0.06 for the first-third ejection fraction (l/3 EF). Utility of the method: To determine the usefulness of assessing left ventricular function during early systole
GROUP
I
n = 32
0.34 0.09
NOw4t SUBJECTS
0.13 0.04
EF = ejection fraction; Group I = normal subjects; Group II = patients with coronary artery disease and a normal ejection fraction; Group Ill = patients with coronary artery disease and a depressed ejection fraction; SD = standard deviation.
462
ET AL
Journal of CARDIOLOGY
EF = .BZ f .04
GROUP
I
n = 31 I
Wllll NORMIt “JI”,ON Fk4CVON EF = .59 f .OE
GROUP
dl
n =‘17 ,
E.WliON
FRWlON
EF=.Yi.O#
FIGURE 3. First-third ejection fraction plotted for each patient in the three qoups. EF = total ejection fraction (normal 0.52). See Table I for individual results.
Volume 44
EARLY
with radionuclide angiography, the following standard calculations12 were also made: 1.
VENTRICULAR
=
= 32132or 100 percent. 2. Sensitivity
=
Number of patients with coronary disease and reduced l/3 EF (<0.25) Total number of patients with coronary disease = 46/48 or 96 percent.
In Group II, the specificity was 94 percent (29 of 31). One patient with a one third ejection fraction of 0.27 had severe three vessel coronary artery disease, but he also had a hypercontractile left ventricle and an ejection fraction of 0.85 on contrast ventriculography. The other patient with a normal first-third ejection fraction (0.29) had a single 70 percent stenotic lesion of the left anterior descending coronary artery. 3. Predictive accuracy = True positives (coronary disease and l/3 EF < 0.25) All positives = 46/46 = 106 percent. There were 2 false negatives (2/48 or 4 percent) and no false positives. Reproducibility of method: For first-third ejection fraction, the intraobserver variation was small, with a mean difference of 0.017 ejection fraction unit (range 0 to 0.04) when 15 randomly selected studies were evaluated 2 months later using the same previously recorded raw data. A second observer performed the first-third ejection fraction calculations without
CORONARY
DISEASE-SLUTSKY
ET AL.
Heart Rate and Blood Pressure in the Three Patlent Groups
Group
Total number of normal subjects
AND
TABLE II
Specificity Number of normal subjects with normal l/3 EF (3 0.25)
SYSTDLE
II III
Heart Rate (beatslmin) fS:: 71f
11
Systolic Blood Pressure (mm Hg) 118f 116 f 121 f
13 14 11
First Pass Ejection Fraction
First-Third Ejection Fraction
0.59 0.62 f 0.04 0.05 0.34 f 0.09
0.29 f 0.04 0.18 0.13 f 0.04
All values are expressed as mean f standard deviation.
knowledge of the results of the first observer. Interobserver variation was modest with a mean difference of 0.023 ejection fraction unit (range 0 to 0.05). Ten patients were studied twice, 7 days apart. The mean difference was 0.03 ejection fraction unit (range 0 to 0.05). Mean beat to beat variation was 0.03 f 0.025 ejection fraction unit (range 0 - 0.10) in 385 studies. The greatest errors were seen in the patients with a very low (less than 0.10) first-third ejection fraction. Vital signs: There was no significant difference (as assessed with analysis of variance) among Groups I, II and III with respect to heart rate or blood pressure (Table II).
Discussion Rushmer13 suggested in the mid 1960s that evaluation of the early phase of the ventricular impulse might prove fruitful in assessing subtle abnormalities of left ventricular function. Gault et aLi4 showed that peak depression in instantaneous velocity of circumferential fiber shortening (Vcf) occurred early in systole, at a time when peak wall stress was recorded. Some investigators15J6 used peak aortic flow and ventricular acceleration to assess left ventricular performance in dogs, whereas others6,7 suggested that assessment of the early phases of left ventricular ejection may yield information not available when only end-systolic data are compared
.50
COMPARSION OF CONTRASTANO AAOIDNUCLIOE EJECTIONFRACTIONS
.
'0
FIRST l/3 EJECTION FRACTION FIRST-PASS flAOlONUCLlDE ANGIOGRAPHY
FKSURE 4. Comparison of contrast and radionuclide first-third ejection fraction in 58 subjects. n = number of subjects: r = correlation coefficient.
FIRST-PASS RADtONUCLlDE EJECTION FRACllON
FIGURE 5. Comparison of contrast and radionuclide ejection fraction in 58 subjects. Abbreviations as in Figure 4.
September 1979
The American Journal of CARWOLGGY
Volume 44
463
EARLY VENTRICULAR SYSTOLF
AND CORONARY DISEASE-SLUTSKY
with end-diastolic measurements. By contrast, Hammermeister et al.17 were unable to differentiate normal from abnormal patients when the rate of change of left ventricular volume was analyzed. However, peak flows during differing phases of ejection were not analyzed. Depressed first-third ejection fraction as an index of impaired left ventricular function: In this study we used first-pass radionuclide angiography to test the hypothesis that patients with coronary artery disease exhibit impairment of early ventricular ejection. In this study, 46 of 48 patients (96 percent) with coronary artery disease had a depressed ejection fraction during the first-third of systole, although 31 had a normal total ejection fraction on both contrast and radionuclide angiography. We were unable to define an abnormality in left ventricular function at rest in only two patients; one patient was somewhat atypical in that he had a “hypercontractile” left ventricle and the second patient had only single vessel coronary disease. Our data also suggest that the sensitivity, specificity and predictive accuracy of this approach support its utility in assessing left ventricular function in patients with known or suspected coronary artery disease, although this group of patients was selected for the presence or absence of coronary artery disease. We have found that this abnormality of ventricular ejection also occurs in other diseases, such as aortic stenosis,18 and Johnson et al.7 also found dysfunction in patients with cardiomyopathy. Clearly sensitivity, specificity and predictive accuracy will depend on patient selection. The correlation between angiographic and radionuclide tech-
10
4
.9 SF z s z .= 9
0.65
.8 0 60 .lL1.18 6-
=
.5-
2 t
.4-
I= J, 2
. 77 . 70
.22 20
0 57
w
g
A.30
A.2o
A 16
25
l .55
niques was excellent and the inter- and intraobserver variations appear to be modest. Reproducibility over time was also reasonably good. Many patients with a depressed ejection fraction had a similar (although depressed) first-third ejection fraction when compared with patients with a normal resting ejection fraction. This may limit the utility of the method in comparing patient groups. Cause of depressed early ejection fraction: The exact basis for this observed abnormality is unclear, One possibility is that patients with coronary artery disease may have decreased aortic compliance, which could reduce the amount of blood ejected early in systole.1g~20 Additionally, it has been shown that areas of abnormal contraction have increased amounts of fibrosis,21 which may lead to contractile abnormalities as a result of decreased developed wall force. Focal conduction abnormalities or chronically depressed blood flow might contribute to these contraction abnormalities, although this would seem less likely. Potential limitations of method: It appears that early ejection is a useful period to study, and based on our preliminary observations can often distinguish normal patients from patients with cardiac disease without utilizing an exercise stress.3l4 The method is simple to perform, nontraumatic and reproducible. The major pitfalls in utilizing this method would include studying patients with such severely depressed left ventricular function that end-diastole cannot be confidently determined or patients with a regurgitant mitral valve in whom a low impedance leak might occur. Some problems may be due to the higher statistical error around lower ejection fractions, and difficulty with defining the end-diastolic point. Additional theoretical limitations include statistical problems due to difficulties with the accuracy of the counts at the enddiastolic and end-systolic data points, and the occasional patients with curves that are too noisy for legitimate analysis. Clearly, with larger patient populations and a greater splay in the normal population, the sensitivity, specificity and predictive accuracy will change (and probably decrease). Part of this may be resolvable with computer analysis, higher count rate acquisition and shorter frame time. The theoretical
l EJECTION
FR4CllOM
VALUES
A FIRSl.~Hl~EJECTIONFR4C~ON VALUES
0.50.52 .3"45
.2 -
.l -
.l
.2
.3
.4
.5
.6
.7
.6
I
1
.S
1.0
1 SPECIFICITY(FALSE POSITIVEFRACTION) FIGURE 6. The sensitivity of the ejection fraction and first-third ejectii fraction compared with the percent of false posltive studies at several threshold levels of normal. A false positive fraction is defined as: Number of false positives/Number of actually negative cases or lspecificity. Note that by lowering the cutoff point for a normal first-thlrd ejection fraction to 0.20 the sensitivity level falls to 85 percent, although the false positive fraction remains unchanged.
464
ET AL.
September 1979
7he Arnerkan Journal d CARMOLOGY
limitations of this first pass method
are based on the statistical variability of the results and the patient populations selected. Figure 6 depicts a curve that compares the sensitivity of a test with the fraction of false positive tests that results when a variety of thresholds are compared. In this graph one can see that the sensitivity of the first-third ejection for coronary disease in this population is well over 80 percent at a strict threshold for normal of 0.20 or greater; at this threshold few false positives occur. With lower normal ranges, sensitivity decreases and with more moderate or lax thresholds, the sensitivity increases but the proportion of false positive studies increases.22 The conventional ejection fraction has a high sensitivity for coronary disease only when a high false positive fraction is accepted. Thus, in general, future analysis of early systolic ejection by this method has significant potential
Volume 44
EARLY VENTRICULAR SYSTOLE AND CORONARY DISEASE-SLUTSKY
variability in sensitivity depending on the threshold accepted as normal. If one excludes the patients with a depressed conventional ejection fraction, the sensitivity at a threshold of 0.20 decreases to approximately 75 percent. Similar analysis with gated equilibrium radionuclide angiography using composite time-activity curves in our laboratory has not been fruitful. Correlation of first-third equilibrium and contrast angiographic ejection has been poor (r = 0.65 or less) ‘resulting, as would be expected, in much less patient separation (23). Thus, noninvasive analysis is restricted to first pass methods in our laboratory.
ET AL.
Clinical implication: First-third ejection fraction obtained with first pass radionuclide angiography is an important new observation that permits identification of subtle abnormalities of left ventricular function at rest in patients with coronary artery disease that may not be recognized by analysis of total ejection fraction alone. At present caution must be observed; it will be important for other studies to be performed to confirm these observations and validate their reproducibility before the measurement can become an integral part of the clinical evaluation of the patient with a chest pain syndrome.
References 1. Bruce RA: Method of exercise testing: step testing, treadmill and isometrics. In Exercise in Cardiovascular Disease and Health (Amsterdam E, Wilmore J, DeMaria A, ed). New York, Yorke Medical Books, 1977, p 149-158 2. Berman D, Amsterdam EA, Mason DT: Detection of myocardial ischemia by rest and exercise thallium-201 scintigraphy. In Ref 1, p 24-31 3. Borer JB, Badhrach !X, Green MV, Kent KM, Epstein SE, Johnston GS: Real time-radionuclide cineangiography In the noninvasive evaluation of global and regional left ventricular function at rest and during exercise in patients with coronary artery disease. N Engl J Med 296:839-8451977 4. Pfieterer M, S&&r G, WCC/ D, Swanson S, Gordon D, Slutsky R, Peterson K, Ashburn W: Profiles of left ventricular ejection by equilibrium radionuclide angiography during exercise and in the recovery period between normals and patients with CAD (abstr). J Nucl Med 19:710, 1978 5. Bfutaky RA, Kartluer J, Rkcl D, Schuler G, Pflsterer M, Peterson K, Ashbum W: Response of left ventricular volume to exercise in normals and!patients with angina pectoris as calculated by radionucltde angiography (abstr). J Nucl Med 19:7 11, 1978 6. LeIghton R, Polla& ME, Wekh T6: Abnormal left ventricular wall motion at mid-ejection in patients with coronary heart disease. Circulation 52:238-248, 1975 7. Johnson L, Etgs K, Bchmldt D, Weiss M, Cannon P: Volume ejected in early systole. Circulation 52:378-386, 1975 8. ScheMerI H, Verba J, Johnson AD, Block G, Alazrakl N, Rose F, Ashburn W: Non-traumatic determination of left ventriculation ejection fraction of radionuclide angiography. Circulation 51: 902-909, 1975 9. Ashburn W, Verba J, Schelbert H: Left ventricular ejection fraction-A review of several radionuclide angiographic approaches using the scintillation camera. Prog Cardiovasc Dis 20:267-284, 1978 10. Bevington PR: Data Reduction and Erorr Analysis for Physical Sciences. New York, McGraw-Hill, 1989, p 60-78
Il.
12. 13. 14. 15. 16.
17. 18.
19.
20.
2 1.
22. 23.
September 1979
Winer BJ: Multifactorial experiments having repeat measures on the same element. In, Statistical Principles in Experimental Design. New York, McGraw-Hill, 1971, p 570-595 Froelicher VF: The detection of asymptomatic coronary artery disease. Annu Rev f&d 28:1-12, 1977 Rushmer R: Initial ventricular impulse. Circulation 29:269-280, 1964 GauH J, Ross J Jr, Braunwatd E: Contractile state of the left ventricle in man. Circ Res 22:457-469, 1966 Nobte MIM: The contribution of blood momentum to left ventricular ejection in the dog. Circ Res 23:663-871, 1968 Nutter W, Noble JR, Hurst VW: Peak aortic flow and acceleration as indices of ventricular performance in the dog. J Lab Clin Med 77:307-313,197l Harnmermefsier KE, Brooks BC, Warbasse JR: The rate of change of left ventricular volume in man. Circulation 49:729-738, 1974 Slutsky R, Karliner J, Battier A, Peterson K, Johnson A, Ross J Jr: The first-third ejection fraction in patients with aortic stenosis with and without coronary disease (abstr). Clin Res 27:204, 1979 Urschel C, Cove8 J, Bonnenblkk E, Ross J, Braunwatd E: Effects of decreased aortic contpliance on performance of the left ventricle. Am J Physiol 214:298-303, 1988 Wilcken D, Charlier A, Hoffman J, Guz A: Effects of alterations in aortic impedance on the performance of the ventricles: Circ Res 14:283-289, 1984 better R, Behar V, Wagner 0, Start J, Starmer F, Lee K, Hackel D: Evaluation of asynergy as an indicator of myocardial fibrosis. Circulation 57:715-725, 1978 Metr C: Basic principles of ROC analysis. Semin Nucl f&d 8: 283-292, 1978 Gordon D, Ashburn A, Stutsky R: Radionuclide angiography: first-pass method of assessing ejection fraction. In. Nuclear Cardiology (Mason D. Berman D, ed). New York, Yorke Medial Books, 1979
The Amertcan Journal of CARDtDLDGY
Volume 44
465