Download Relationship QRS Amplitude Changes During Exercise to

Relationship of QRS Amplitude Changes During Exercise
to Left Ventricular Function and Volumes
and the Diagnosis of Coronary Artery Disease
ALEXANDER BATTLER, M.D., VICTOR FROELICHER, M.D., ROBERT SLUTSKY, M.D.,
AND WILLIAM ASHBURN, M.D.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
SUMMARY Preliminary studies have suggested that QRS-amplitude changes due to exercise-induced
alterations in ventricular volume and function can improve the diagnostic value of the exercise test. To evaluate
this, electrocardiographic data and equilibrium radionuclide angiographic images were recorded simultaneously in 18 normal subjects and 60 coronary artery disease patients at rest and during supine bicycle exercise. In 24 of the 60 coronary artery disease patients, left ventricular volumes were also calculated. The
measured QRS amplitudes were the R waves in V, X, Y and Z, the Q wave in Z and the sum of amplitudes of
R waves in X and Y and the Q wave in Z (2iR). The mean left ventricular ejection fraction increased significantly from rest to peak exercise in the normal subjects; however, the mean left ventricular ejection fraction
and mean volumes did not change significantly in the coronary patients. There was no significant difference in
the mean QRS-amplitude changes during exercise between the coronary artery disease patients and the normal
subjects in any of the measured leads. The sensitivity and specificity of exercise-induced QRS-amplitude
changes for coronary disease were lower than ST-segment changes. For ST-segment changes, the sensitivity
was 57% and specificity was 100%; the best sensitivity and specificity for QRS amplitude occurred in RZ, 48%
and 67%, respectively. When ejection fraction was related to 2R at rest and peak exercise for both normal subjects and coronary patients the correlations were fair (0.50, 0.51 respectively); however, the correlation
between the magnitude of 2R and ejection fraction change from rest to peak exercise was poor and did not improve with any other measured QRS amplitudes or by separating normal subjects from coronary patients with
and without previous myocardial infarction. There were also poor correlations between end-diastolic and endsystolic volumes to QRS amplitudes at rest, peak exercise and their magnitude of change from rest to peak exercise. Thus, R-wave amplitude changes during exercise testing have little diagnostic value and are not related
to exercise-induced changes in left ventricular function or volumes.
R-WAVE CHANGES in CM5 have been reported to
increase the diagnostic value of the exercise test. An
increase in R-wave amplitude during exercise has been
reported to indicate severe left ventricular dysfunction
and coronary obstruction, while a decrease in R-wave
amplitude is consistent with normal left ventricular
function.1' 2 These changes are thought to be caused by
an increase in end-diastolic volume during exercise
and the Brody effect.3 The frequent S-wave increase in
V4 in normal subjects during exercise testing was
postulated to be the result of increased myocardial
function.4 Resting left ventricular ejection fraction obtained by contrast angiography has demonstrated a
positive correlation with the sum of various Q- and Rwave amplitudes,5 but this has not been studied during
exercise.
The gated equilibrium radionuclide angiography
technique correlates well with angiographic studies of
ejection fraction and is useful for detecting ejection
fraction responses to exercise.'-" With this method, it
is also possible to calculate end-diastolic and end-
systolic volumes independent of geometric assumptions.'1 Thus, for the first time, QRS-amplitude
changes during exercise could be simultaneously correlated with exercise-induced changes in left ventricular function and volumes. The aim of this study
was to test the diagnostic value of exercise-induced
QRS-amplitude changes and to evaluate the
hypothesized mechanisms for these changes.
Methods
Patient Population
Sixty patients with coronary artery disease and 18
control subjects were studied. The control subjects
were normal volunteers without clinical or electrocardiographic evidence of cardiovascular or pulmonary
disease. The mean age was 41 years (range 30-57
years). The patients with coronary artery disease were
males, and their mean age was 55 years (range 35-76
years). Coronary artery disease was documented by
previous myocardial infarction (MI) at least 3 months
before the study in 35 patients, coronary angiography
in 10 patients, and angina pectoris with abnormal
ECG and thallium treadmill test in the other 15
patients. An abnormal treadmill test was defined as
more than 0.1 mV of flat or downsloping ST-segment
depression or ST elevation in any of the 15 electrocardiographic/vectorcardiographic (ECG/VCG) leads
other than aVR. Coronary artery disease was defined
angiographically by the presence of at least a 70%
decrease in the luminal diameter in one or more of the
major coronary arteries. An abnormal thallium test
From the Division of Cardiology, Department of Medicine and
the Division of Nuclear Medicine, Department of Radiology,
University of California, San Diego, California.
Supported by SCOR on Ischemic Heart Disease grant HL 17682,
NHLBI, NIH.
Dr. Battler's current address: Heart Institute, Chaim Sheba
Medical Center, Tel Hashomer, Israel.
Address for correspondence: V. Froelicher, M.D., Director of
Cardiac Rehabilitation, Department of Medicine, 225 Dickinson
Street, San Diego, California 92103.
Received March 12, 1979; accepted April 24, 1979.
Circulation 60, No. 5, 1979.
1004
QRS AMPLITUDES AND LV FUNCTION/Battler et al.
was defined as a perfusion defect at peak exercise filling during the recovery period or a perfusion defect
remaining after redistribution. Ten of the coronary
artery disease patients were treated with antianginal
doses of propranolol (80-320 mg) and 15 of the
patients were treated with either nitroglycerin or isosorbide dinitrate. All normal subjects and coronary
artery disease patients were studied after they gave
written informed consent.
Equipment and Imaging Techniques
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Multiple ECG gated equilibrium radionuclide angiography was performed after complete mixing
throughout the vascular space of 20-25 mCi 99mTc
tagged to human serum albumin, administered via a
peripheral arm vein as previously described.6-9 -Imaging was performed by positioning the detector of an
Anger-type single-crystal scintillation camera over the
chest in a 40-50° left anterior oblique and 5-10°
caudad tilt in order to isolate the left ventricle optimally from the left atrium. Data acquisition was accomplished using a commercially available nuclear
medicine computer system, Medical Data SystemPAD (MUGA), which divided the RR interval into 28
equal time periods (frames) of 20-40-msec duration,
depending on heart rate. It then assembled several
hundred heart cycles at corresponding times to
generate composite images throughout the cardiac cycle. In a 2-minute acquisition, usually 2500-7000
counts could be accumulated in the left ventricular
region of interest at end-diastole in one 40-msec frame
(counts corrected for background). To generate a
time-activity curve, a rectangular region of interest
was arbitrarily placed around the left ventricle at enddiastole. A computer algorithm (MUGE) automatically determined the edge of the left ventricle
using a combination of the second derivative and a
count-rate threshold of 5% per element as guidelines.
Each subsequent frame was processed at the same
threshold so that throughout all 28 frames a
"variable" region of interest was used to determine
the changing count rate within the left ventricle. A
computer-assigned background region of interest outside the lower quadrant of the left ventricle was used
to correct for noncardiac activity.
The ejection fraction was then calculated from the
time-activity curve according to the formula:
EDc
where ED, = left ventricular counts at end-diastole
(usually the first frame of the time activity curve corrected for background) and ES, = left ventricular
counts at end-systole (nadir of curve corrected for
background). The ejection fraction calculated by this
technique has been shown to correlate well with those
determined from biplane cineangiography, with a correlation coefficient in our laboratory of 0.91.9 The
inter- and intraobserver variations of ejection fraction
calculated by equilibrium radionuclide angiography in
our laborRitory is 0.04 ejection fraction units. Thus, in
1005
any single patient a significant change of ejection fraction was considered only when the change was at least
+0.05 ejection fraction units. End-systolic and enddiastolic volumes were calculated from 24 coronary
artery disease patients using a radionuclide method
free of geometric assumptions as previously described
by Slutsky et al.'1 Briefly, both at rest and peak exercise, the end-diastolic and end-systolic counts were
divided by the number of processed heart beats and
normalized for counts in milliliters of plasma drawn
from a peripheral vein while the patient was at rest
and again at peak exercise. Using this technique the
correlation of radionuclide angiographic volumes with
biplane contrast angiographic volumes was 0.98."
ECG/VCG Technique
Twelve ECG leads and three orthogonal Frank
leads were obtained using 14 electrodes. Consistent
electrode placement was accomplished using the
Dalhouise square.'2 The chest electrodes were small
silver electrodes that did not interfere with imaging.
The data were acquired using a specially developed,
microcomputer-assisted, 15-lead ECG/VCG recorder
(Marquette Data Logger) that records on paper and
continuously stores on floppy discs. In this way we
could obtain accurate peak exercise ECG/VCG
changes. ST-segment changes were considered in any
of the 15 ECG/VCG leads. To measure the QRS
amplitudes, the isoelectric line was always defined as
the PR interval and the highest and lowest amplitudes
were averaged from consecutive beats during 10
seconds at rest and during the last minute of exercise.
The measured QRS amplitudes were R wave in V, R
wave in X, R wave in Y, R wave in Z, Q wave in Z, as
well as 2R,which is equal to RX + RY + QZ.
Study Protocol
Both the normal subjects and patients were studied
first in the supine basal state and then during a graded
supine bicycle exercise with up to three levels of increasing work load, each lasting 3 minutes. The test
was discontinued because of any of the following:
angina, more than 0.3 mV ST depression or fatigue.
The supine exercise test was performed on an imaging
table supporting an electronically braked bicycle
ergometer (Uniwork 845T, Quinton Instruments) that
kept a constant work load over a pedal speed range of
40-70 rpm. This allowed the work done by each
patient at each level of exercise to be held constant
over the 3 minutes. The ejection fraction and the 15
scalar leads were recorded in the supine resting state,
and then the supine bicycle exercise was performed.
Equilibrium data were recorded for 5 minutes at rest
and continuously throughout the exercise period. The
data from the last 2 minutes of the peak 3-minute
work load level were analyzed to avoid the rapid
changes in heart rate and other hemodynamic
variables that usually occur in the first minute of a new
exercise level and affect the accuracy of the ejection
fraction calculations. Blood pressure was taken with
an arm cuff at rest and at peak exercise.
VOL 60, No
CIRCULATION
1006
5,
NOVEMBER 1979
but in the coronary artery disease patients it increased
by only 52 beats/min (p < 0.001). There was no
significant difference in the mean systolic blood
pressure at rest and peak exercise between the normals
and coronary artery disease patients. The mean ejection fraction at rest was significantly higher
(p < 0.001) in the normals (0.62 ± 0.06) than in all
the coronary artery disease patients (0.54 ± 0.11) or
the group with previous MI. (0.49 ± 0.11). However,
in the coronary artery disease patients without
previous MI., the resting ejection fraction was similar
to that of normals (0.62 ± 0.06). At peak exercise, the
normal subjects had a mean increase of 0.12 ejection
fraction units from the resting value. The coronary
artery disease patients did not significantly change the
mean ejection fraction from rest to peak exercise in
either those with or without a previous myocardial infarction.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Statistical Analysis
Statistical analysis used was the t test for paired and
nonpaired samples with a level of significance of
<0.01. The correlation coefficients were significant at
a level of <0.05. All the results are presented as the
mean ±SD. Sensitivity and specificity were defined as
described by Yerushalmy. 13 Briefly, sensitivity is
defined as the percentage of coronary disease patients
with an abnormal response TP/(TP + FN);
specificity is defined as the percent of subjects without
coronary disease with a normal test TN/(TN + FP).
To compare the combined sensitivity and specificity of
the various ECG/VCG variables, Youden's Jsummary index'4 was used, which has an algebraic
structure equivalent to the sum of sensitivity plus
specificity minus 1. The best summary index is closest
to 1, while negative summary indices indicate little
diagnostic value.
Results
In table 1 the heart rates, systolic blood pressures,
ejection fractions and QRS amplitudes (RV,, RX,
RY, RZ, QZ, XR) are presented at rest and peak exercise for the normals, all the coronary artery disease
patients and the coronary artery disease patients with
and without previous MI. The mean heart rate and
systolic blood pressure increased significantly from
rest to peak exercise in all the groups. The mean heart
rates at rest were not significantly different between
the groups; however, the mean exercise heart rate was
significantly higher in the normal subjects. In the normals the mean heart rate increased by 72 beats/min,
ECG/VCG Changes
The mean R-wave amplitudes in V5, X and ER
decreased significantly from rest to peak exercise both
in normals and in coronary artery disease patients.
There was no significant difference in the mean Rwave amplitudes in V5 or X either at rest or peak exercise between normal subjects and coronary patients.
The mean ZR amplitude was significantly lower
(p < 0.001) at rest and at peak exercise in all the coronary artery disease patients (rest 1.98 ± 0.64 mV,
peak exercise 1.68 ± 0.64 mV) and in patients with
previous MI (rest 1.79 ± 0.63 mV, peak exercise
1.51 ± 0.56 mV), relative to the normal subjects (rest
1.
Mean (- SD) Heart Rates, Systolic Blood Pressures, Ejection Fractions and QRS Amplitudes of the Normal Subjects and
Coronary Artery Disease Patients at Rest and Peak Exercise
HR
RX
RY
SBP
RZ
QZ
ER
RVs
(beats/
EF
(mm Hg)
min)
(mV)
(mV)
(mV)
(mV)
(mV)
(mV)
NL (n = 18)
rest
72 - 10 125 - 13 0.62 - 0.06 1.38 - 0.58 1.07 - 0.37 1.02 - 0.51 0.63 - 0.26 0.42 - 0.16 2.51 - 0.81
144 - 17 184 - 21 0.74 - 0.06 1.16 - 0.54 0.88 - 0.23 0.97 - 0.48 0.76 = 0.34 0.36 - 0.18 2.21 - 0.67
Pex
<0.001
NS
<0.001
<0.001
<0.005
<0.001
NS
NS
p
<0.001
TABLE
CAD (n = 60)
64 - 10 131
rest
Pex
116 = 21 173
p
<0.001
-
0.11 1.38 = 0.50 1.06
0.14 1.14 - 0.41 0.87
17 0.54
25 0.55
<0.001
NS
<0.001
=
0.34 0.64
0.31 0.60
<0.001
0.33 0.75
0.32 0.80
NS
-
NS
0.30 0.25 0.21 1.98 | 0.64
C.27 0.22 * 0.20 1.68 i 0.64
<0.001
<0.001
CAD with previous MI (n = 35)
rest
67 - 11 131 15 0.49 = 0.11 1.30 - 0.52 1.00 - 0.35 0.56 - 0.29 0.81 - 0.34 0.25 = 0.23 1.79 0.63
Pex
117 - 21 170 23 0.48 - 0.12 1.06 - 0.40 0.80 - 0.29 0.53 - 0.28 0.83 - 0.31 0.21 0.35 1.51 0.56
<0.001
p
<0.001
NS
<0.001
<0.001
NS
NS
<0.001
<0.001
=
CAD without previous MI (n = 25)
rest
63
8 131 = 17 0.62
- 0.06
1.50 0.46 1.15 0.30 0.77 - 0.35 0.66 0.21 0.34 - 0.18 2.25 0.58
Pex
116 22 179
28 0.64
0.11 1.28 - 0.40 0.98
0.32 0.70 d 0.36 0.76
0.21 0.26
0.19 1.94 - 0.66
<0.001
p
<0.001
NS
<0.001
<0.001
<0.001
NS
<0.01
<0.001
Values are mean = SD.
Abbreviations: NL = normal subjects; CAD = coronary artery disease patients; MI = myocardial infarction; HR = heart
rate; SBP = systolic blood pressure; EF = ejection fraction; Pex = peak exercise.
QRS AMPLITUDES AND LV FUNCTION/Battler et al.
ARX
1.5 r
ARY
1007
/ARZ
l
9
0.5 _
a
e
0
00
o
E
I1
o
0
-(DOW
-cccm co
-06
-Q=W-
on"
so
0=
-M
--cocamomoac-
-0.5
ow
0
-1
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
-1.5
_
CAD
Nl
NL
CAD
NI
CAD
NL
NI
CAD
CAD
FIGURE 1. QRS-amplitude changes (Am V)from rest to peak supine bicycle exercise for R waves in V5, X,
Y, Z and E2R (RX + R Y + QZ). Closed circles are normal subjects (NL) and open circles are coronary
artery disease patients (CA D). The mean change for each group is presented as an uninterrupted horizontal
line. There is no significant difference between normals and coronary artery disease patients in the
magnitude of QRS-amplitude changes from rest to peak exercise in any of the measured leads.
2.51 + 0.81 mV, peak exercise 2.21 0.67 mV) and
the patients without previous MI (rest 2.25 0.58
mV, peak exercise 1.94 + 0.66 mV). Mean R-wave
amplitude in Y did not change significantly from rest
to peak exercise in any of the groups of patients.
However, at rest and peak exercise, the mean R-wave
amplitude in Y was significantly lower in the coronary
artery disease patients compared with the normals.
Mean R-wave amplitude in Z did not change
significantly in normals, coronary artery disease
patients or coronary artery disease patients with
previous MI; however, it did increase significantly in
patients without a previous MI. The trend of the R
wave in Z amplitude was to increase during exercise.
The Q-wave amplitude in Z did not change significantly in the normals, however, it did decrease significantly in the coronary artery disease patients. The
mean resting and peak exercise Q waves in Z were significantly lower in amplitude in the coronary artery
disease patients than in the normals.
Figure 1 demonstrates the distribution of QRSamplitude changes from rest to peak exercise between
normals and coronary artery disease patients. An increase or no change in R wave was considered to be a
abnormal response to exercise; a decrease was considered normal. However, the Q wave in the Frank
scalar Z lead corresponds to the R wave in lead V2 of
the standard ECG, while the R wave of the Frank
scalar Z lead corresponds to the S wave in lead V2.
Thus for lead Z, any decrease or no change in the R
wave at peak exercise was considered an abnormal
response and any increase in RZ at peak exercise was
considered a normal response. The R wave in V5 in-
creased or did not change at peak exercise compared
with the preexercise amplitude in only 12 of the 60
coronary artery disease patients (for a sensitivity of
20%) and decreased in 14 of the 18 normals (for a
specificity of 78%). The R wave in X increased or did
not change in 10 of the 60 coronary artery disease
patients (for a sensitivity of 17%) and decreased in 14
of the 18 normals (for a specificity of 78%). The R
wave in Y increased or did not change at peak exercise
in 30 of the 60 coronary artery disease patients (for a
sensitivity of 50%) but decreased in only six of the 18
normal subjects (for a specificity of 34%). The R wave
in Z decreased or did not change in 29 of the 60 coronary artery disease patients (sensitivity 48%) and increased in 12 of the 18 normals (specificity 67%).
The R waves increased or did not change in nine of
the 60 coronary artery disease patients (sensitivity
15%) and decreased in 16 of 18 normals (specificity
89%). Thirty-four of the 60 coronary artery disease
patients (sensitivity 57%) had abnormal ST changes at
peak exercise, while none of the normals showed
similar changes (specificity 100%). Thus, the sensitivity and specificity of ST changes at peak exercise
were better than any of the QRS-amplitude changes
and according to the summary index, the QRS
changes reflecting the best sensitivity and specificity
occurred in RZ (table 2).
Radionuclide Ejection Fraction Changes
Figure 2 shows the changes in ejection fraction from
rest to peak exercise for each of the 18 normal subjects
and each of the 60 coronary artery disease patients.
Voi 60, No 5, NOVEMBER 1979
ClIRCULATION
1008
The Sensitivity, Specificity and Summary Index of the QRS Amplitude and ST-segment Criteria for
TABLE 2.
Diagnosing Coronary Disease.
R.X
RV.
Sensitivity
0.20 (12/60) 0.17 (10/60)
Specificity
0.78 (14/18) 0.78 (14/18)
Summary
index
ND
ND
=
Abbreviations: SI summary index; ND
were nondiagnostic.
RY
RZ
0.30 (30/60)
0.34 (6/18)
0.48 (29/60)
0.67 (12/18)
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
NL
18
n
0. 15 (9/60)
0.89 (16/18)
0.,7 (34/60)
1.00 (18/18)
ND
=
0.04
0.15
0.57
negative values for the SI indicating that the measurements
The coronary artery disease patients had a significantly lower mean resting ejection fraction than
the normals. At peak exercise the mean ejection fraction increased significantly in the normals and did not
change in the coronary artery disease patients. Every
normal subject increased the peak exercise ejection
fraction from a normal resting ejection fraction
(>0.50) by at least 0.05 ejection fraction units. Of the
coronary artery disease patients, 44 had either a
decrease or change in the peak exercise ejection fraction from rest.
Of the 16 coronary artery disease patients that increased peak exercise ejection fraction by at least 0.05
ejection fraction units, four had an abnormal resting
ejection fraction (<0.50),- three were taking antianginal doses of propranolol and three had isolated
right coronary artery stenosis. Of the remaining six,
three stopped exercise because of fatigue, and three
because of angina or significant ST changes.
90 F
ST
Volume Changes
In figure 3 the end-diastolic and end-systolic volume
changes from rest to peak exercise for 24 of the coronary artery disease patients are presented. Both the
mean end-diastolic and end-systolic volumes did not
change significantly from rest to peak exercise.
However, nine patients had increased end-diastolic
volume from rest to peak exercise, of whom seven
(78%) had accompanying angina or abnormal ST
changes at peak exercise, compared with only five of
the 15 patients (33%) that decreased or did not change
CAO n= 24
220 r-
EDV
ESV
200 H
CAD
n = 60
180 H
160 V
.80 K
I
.70 1
E
140
k
L*J
120-IF
.
C)
-
-j
0
.60
100
F
1
l-l
CC
cr
C-,)
80
.50
LU)
LU
60
I
p
.40
<0.001
T7
40
.30
F
20 L
.20 L
REST
REST
PEX
REST
PEX
FIGURE 2. Ejection fraction changes from rest to peak
supine bicycle exercise (PEX) in the 18 normal subjects (NL
closed circles) and the 60 coronary artery disease patients
(CAD
open circles). The average values are presented as
mean ± SD.
PEX
REST
PEX
FIGURE 3. End-diastolic volume (EDV) and end-systolic
volume (ESV) changes from rest to peak supine bicycle exercise (PEX) in 24 of the coronary artery disease patients. The
average values are presented as the mean ± SD. There is no
significant change in the mean EDV and ESVfrom rest to
peak exercise.
QRS AMPLITUDES AND LV FUNCTION/Battler et al.
NL + CAD n
=
78
NL
22
0.
n =
35
CAD WITHOUT PREVIOUS MI
n = 25
tL
06
0 4-
CAO WITH PREVIOUS Ml
CAD n = 60
n= 18
1009
rGL
7
01
Di-
uL
-
= REST
RE
B
62
RX
RY
.RZ
C] PEAK EXERCISE
mA
M
p 0 05
*
BUsBE5BXY
62
2.6R
RV5
RX
RY
RZ
2R
RVs
RX
RY
6RZ
YR
RVs
Y. R
RV5
RX
RY
RZ
'R
FIGURE 4. The correlation coefficients for the least-squares fit regressions of ejection fraction vs QRS
amplitudes (R waves in V5, X, Y, Z and ER) at restl and peak exercise and the magnitude of change from
rest to peak exercise (A) for both the normal subjects (NL) and coronary artery disease patients (CA D), for
the NL alone, for the CAD alone andfor the CAD patients with and without previous myocardial infarction
(MI).
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
end-diastolic volume at peak exercise. Of the nine
patients who had increased end-diastolic volume at
peak exercise only one also had an increase in one of
the QRS amplitudes (RV5, from 1.1 mV at rest to 1.2
mV at peak exercise); the other seven patients
decreased all QRS amplitudes measured at peak exercise.
Correlation of Ejection Fraction to QRS Amplitudes
Figure 4 shows the correlation coefficients between
ejection fractions and QRS amplitudes at rest and
peak exercise and their magnitude of change from rest
to peak exercise for normal subjects and coronary
artery disease patients together, for normal subjects,
for all the coronary artery disease patients, and for the
patients with and without previous MI. In all these
five groups there was a fair correlation between most
of the R-wave amplitudes and ejection fraction both at
rest and peak exercise. In most of the cases at peak exercise the correlation coefficients were -somewhat
lower than at rest. However, the correlation between
the magnitude of change of ejection fractions and
change of QRS amplitudes during exercise was never
significant. The best correlations were obtained in the
combined normals and coronary artery disease
patients at rest (r = 0.50) and peak exercise
(r = 0.51). Figure 5 is an example of the correlation
obtained from both the normals and the coronary
artery disease patients for the 2R at rest and peak exercise and the magnitude of changes from rest to peak
exercise. At rest, the correlation coefficient was 0.50
and at peak exercise it was 0.51. Both these correlations were significant (p < 0.001). However, the
correlation coefficient between the change of ejection
fraction and 2 R amplitude changes from rest to peak
exercise was not significant. Fourteen subjects had a
2R amplitude of less than 1.5 mV, and only two of
them were normals; eight of the 12 coronary artery
disease patients in this group also had resting ejection
fractions of less than 0.50 (fig. 5, rest). At peak exercise there were 24 coronary artery disease patients and
two normnal subjects who had 2;R amplitude of less
than 1.5 mV and 14 of them, all with coronary artery
disease, also had reduced peak ejection fractions of
less than 0.50 (fig. 5, peak exercise).
Volumes and QRS Amplitudes Correlations
In figure 6, the correlation coefficients between enddiastolic and end-systolic volumes and R-wave
amplitudes are presented at rest, peak exercise and for
the magnitude of change from rest to peak exercise.
All these correlations were poor, although they were
again somewhat better at rest and peak exercise than
for the magnitude of change from rest -to peak
exercise.
QRS Changes During Treadmill Testing
To test if similar results would be obtained, 25 of
the coronary artery disease patients had a treadmill
exercise test done within 24 hours of the supine bicycle
exercise test. R-wave amplitudes in V5 were measured
at rest and at maximal treadmill exercise. During the
treadmill exercise test the mean R-wave amplitude in
V5 decreased significantly (p < 0.001) from 1.35 +
0.63 mV at rest to 1.16 ± 0.53 mV at peak exercise
(fig. 7).
Discussion
Recently, Bonoris et al." 2 recommended the use of
R-wave amplitude changes during exercise to improve
the sensitivity and specificity of the exercise test. As
previously hypothesized,3' , 15-18 they suggested that
exercise-induced changes in left ventricular function
and volume cause the QRS-amplitude changes.
However, these theories have never been validated,
mainly because of technical problems involved in
simultaneously studying left ventricular function,
volume and QRS amplitude during exercise. Nuclear
cardiology and newer ECG techniques make it possible to measure dynamic changes in left ventricular
function, volumes and the ECG continuously during
exercise. The equilibrium radionuclide angiography
technique has been extensively validated by us9 and
others6-8, 10 and found to be very accurate. Thus, for
the first time, the relationship of exercise-induced
1010
VOL 60, No 5, NOVEMBER 1979
CI RCULATION
PEAK EXERCISE
PEAK EXERCISE - REST (,A)
r90
.30 r
80
70
;z
8
.60~
o
0
C-,
4
.20 -
.
0
..o8°. #
o
.50
0
0
o
.10
.0
F
00.
* 0
CD
40
O
C
C)
000
0
30~
O§°
z
20
y
a
50
0407 +0074x)2
p<0 00
.90
O -
*
8
o
0
0S,
0
0
LU-
C-L
LUJ
REST
0
-.10 F
0
000
0
-20 1
* NL
0 CAD -0. n
.80
70
0 9
0
z
0
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
c&
60
al
...
)0
cr-
.50
0
's
-<.O
0.
78
-0.09
0 ,,
0
*"
- 8
-
r=
-30
c
=
c
0
0 0
-3
-2
U
mV AIR
-1
C3
1
2
3
.30
r= 51
.20
A
y =
0.39+ 0.IIx)10 2
p<
0.01
L
1
0.5
2
1.5
32R
2.5
3
3.5
4
4.5
mV
FIGURE 5. The correlation between ejection fraction and mV E2R amplitude at peak exercise and at rest (left) and the
magnitude of change from rest to peak exercise (A) (right). Closed circles represent normal subjects (NL) and open circles
represent coronary artery disease patients (CAD).
0.5
0.4
0.3
ESV
EDV
-
0.20.1
CD,
00.
CC
-0.2
C-
-0.3r
L"
RV5
RX
RY
RZ
YR
RV5
RX
RY
RZ
IR
O REST
EM PEAK EXERCISE
*
p<0.05
FIGURE 6. The correlation coefficients between end-diastolic volumes (EDV) and end-systolic volume
(ESV), and QRS amplitudes (R V5, RX, R Y, RZ, 2R) at rest, peak exercise and the magnitude of change
from rest to peak exercise (A).
QRS AMPLITUDES AND LV FUNCTION/Battler et al.
RV5
n= 25
2.5
2.0
E
1.5
c
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
1.0
0.5
p <0.001
A
I
REST
Peak TM
Exercise
FIGURE 7. R-wave amplitude changes in V5 (mV) from
rest to peak treadmill (TM) exercise in 25 of the coronary
artery disease patients. The average values are presented as
the mean
±
SD.
changes in left ventricular volume and function to the
electrocardiogram could be assessed.
The magnitude and direction of exercise-induced
volume changes either in normals or coronary patients
is still controversial. Rushmer et al.19 used microcrystal ultrasound transducers in conscious dogs and
concluded that the heart dimensions often become
smaller during exertion. However, using the same
technique in swimming dogs, Erickson et al.20 showed
an increase in end-diastolic left ventricular diameter
during exercise. Using epicardial metal markers,
Braunwald et al.2' showed only small changes in both
end-diastolic and end-systolic dimensions during
supine exercise. Using the thermodilution technique in
normals and patients with a variety of heart diseases,
Gorlin et al.22 found variable changes in end-diastolic
volume during supine exercise. Using the echocardiographic technique, Stein et al.23 demonstrated a
1011
slight decrease in end-systolic dimension without a significant change in end-diastolic dimension during
supine exercise in normal subjects. Using contrast
angiography24 and nuclear angiography25 it was demonstrated that coronary artery disease patients who
developed angina during supine exercise usually increased both the end-diastolic and end-systolic
volumes, while patients who did not develop angina
usually did not change significantly the left ventricular
volumes. In another study, using erect bicycle exercise, both normals and coronary artery disease
patients increased the end-diastolic volume during exercise. 26 Thus, previous studies suggest that differences
in exercise position and severity of coronary disease
account for different volume responses to exercise. In
the present study the 24 coronary artery disease
patients in whom we could obtain volumes using the
radionuclear angiography technique did not change
significantly either the mean end-diastolic or endsystolic volumes from rest to peak supine exercise.
However, of nine patients that increased end-diastolic
volume during exercise, seven developed angina or
significant ST changes at peak exercise.
Ejection fraction changes during exercise have been
extensively investigated with radionuclide angiography. Most investigators agree that ejection fraction
increases during exercise in normal subjects, while
most coronary artery disease patients show no change
or a decrease.6 9, 25-27 In the present study the normal
subjects increased the ejection fraction significantly by
a mean of 0.12 ejection fraction units and each subject
increased the ejection fraction by at least 0.05; the
coronary artery disease patients did not significantly
change the mean ejection fraction from rest to peak
exercise. However, 16 of the 60 coronary artery disease patients increased the ejection fraction at peak
exercise by more than 0.05. Of these 16 patients, four
had an abnormal resting ejection fraction (<0.50),
which as an isolated finding is diagnostic of left ventricular dysfunction. Of the other 12 patients, three
were taking antianginal doses of propranolol and
three had isolated right coronary lesions, both causes
of a false-negative ejection fraction response to exercise as shown previously by our group.28' 29 Thus the
other six patients (10%) are true false-negative
responders. Of the six, three patients did not develop
angina or ST changes during exercise and stopped
because of fatigue. We can speculate that in these
patients the exercise challenge was not enough to
cause a mismatch between oxygen demand and
supply.
QRS-amplitude changes during exercise are as controversial as volume changes. Bonoris et al.2 measured
exercise-induced R-wave changes in CM5 in a highly
selected group of coronary artery disease patients and
in normal subjects. Most of these patients were falsepositive or false-negative responders by ST-segment
criteria. They defined an abnormal response as an increase or no change in R-wave amplitude and a normal response as a decrease in R-wave amplitude at
1012
CIRCULATION
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
peak exercise from rest. Using these criteria, the sensitivity and specificity of R-wave amplitude changes
during exercise were better than ST changes. They
also compared R-wave changes during exercise to the
severity of angiographic lesions and concluded that
exercise-induced increases in R-wave amplitude occurred with left ventricular dysfunction.' Katzeff and
Edwards4 predicted an electromechanical association
between S-wave amplitudes and cardiac mechanical
function. In normal subjects, they reported an increasing S-wave voltage in a horizontal bipolar lead
with increasing work load and concluded that the
amplitude of the S wave in certain leads acts as an indicator of left ventricular function. Kentala et al.'5
studied the variation of R wave from V5 during
progressive bicycle exercise testing as an index to
predict the results of physical training in patients with
coronary artery disease. They concluded that the Rwave changes were greatest in the cardiac patients
who responded to exercise training, and the smaller
the dynamic changes of the QRS amplitude during exercise testing the worse the training results. In another
paper Kentala and Luurila16 showed that R-wave
amplitude changes occurred with postural changes
and exercise in healthy subjects and patients surviving
acute MI.
In the present study, R-wave changes during exercise had the same directions in normal subjects as in
coronary artery disease patients. In both normals and
unselected coronary artery disease patients, the sensitivity and specificity of QRS-amplitude changes
were less than for ST changes. No single lead could
separate normals and coronary patients using the Rwave amplitude criteria, including RV5 and RX.
Among the nine patients who increased end-diastolic
volume at peak exercise, only one also increased the
R-wave amplitude in V5 (from 1.1 mV at rest to 1.2
mV at peak exercise). A possible explanation for these
different results could be that we used supine bicycle
exercise rather than treadmill or erect bicycle exercise.
To examine this, 25 of the coronary artery disease
patients were retested on a treadmill and RV,
amplitude was measured at rest and peak exercise.
There was again a significant decrease in RV,
amplitude from 1.35 ± 0.63 mV at rest to 1.16 ± .53
mV at peak exercise (p < 0.001) and only two patients
increased the amplitude of R wave in V5 from rest to
peak exercise. Another possible explanation is that
others may only allow their coronary patients with
severe disease to perform submaximal exercise. During submaximal exercise, most subjects have an increased R-wave amplitude, but decrease it at higher
work loads. As we described previously,30 exerciseinduced R-wave changes in the Z vectorcardiogram
lead had somewhat better sensitivity and specificity
than amplitudes measured in other leads. Using the
sum of R-wave amplitudes from X, Y and Z did not
improve the sensitivity or specificity of the test.
The relationship between QRS amplitudes and left
ventricular volumes and function at rest has been extensively investigated. The Brody effect,3 which ex-
VOL 60, No 5, NOVEMBER 1979
plains the effect of intracardiac blood on QRS potentials by augmentation of the initial portion of the QRS
complex and attenuation of the latter portion, was
derived from models3' 33 and animal experiments.34
However, human studies have not all agreed with this
theory of why an increase in left ventricular enddiastolic volume should increase R-wave amplitude.
Baxley et al.3' and Vine et al.36 found a positive correlation between QRS voltage and left ventricular
end-diastolic volume while Talbot et al.37 found an inverse correlation for volume and QRS voltage.
The relationship of QRS amplitudes and ejection
fraction was studied previously at rest. Gottwik et al.'
and Askenazi et al.38 compared QRS amplitudes obtained from the orthogonal Frank system with left
ventricular ejection fraction measured by contrast
angiography. They found that the sum of RX + RY +
QZ correlated significantly with ejection fraction, and
this correlation even improved with the ejection fraction of post extrasystolic beats. However, these studies
were not simultaneous ECG-function-volume studies
and were done only at rest. In the present study we
found a fair positive correlation at rest and peak exercise between R waves and ejection fractions and, as in
the previous works' 38 we also found that the best correlations were obtained using 2R. However, when we
correlated the magnitude of exercise-induced changes
in QRS amplitudes in the various leads to ejection
fraction changes, the correlations were not significant.
For left ventricular volumes, there were no significant
correlations either at rest or peak exercise or between
the magnitude of changes between rest and peak exercise. We could not find any improvement in the correlation coefficients when we separated the normals
from coronary patients, and coronary patients with or
without previous MI. Thus, although we agree that
there is a positive correlation between R-wave
amplitude and left ventricular ejection fraction either
at rest or peak exercise, this relationship does not explain the changes that occur in the QRS amplitudes
during exercise.
Other mechanisms than changes in left ventricular
volume and function must be responsible for exerciseinduced QRS-amplitude changes. Changes in wall
thickness during exercise could result in alterations in
size or number of individual muscle fiber dipoles.
Changes in the position of the heart in the thorax during exercise and in the amount of air in the lungs during exercise can affect the conductivity of the chest.
These are hypothesized mechanisms that require
further investigation.
In summary, during maximal exercise there is a
significant increase in ejection fraction in normals, no
change or a decrease in ejection fraction in most of the
coronary artery disease patients, no significant volume
changes in coronary patients who are not developing
angina pectoris, and a similar decrease in R-wave
amplitude during maximal exercise both in normals
and coronary artery disease patients.
Thus, our study has not confirmed the direction of
exercise-induced QRS changes in coronary patients or
QRS AMPLITUDES AND LV FUNCTION/Battler et al.
their relationship to left ventricular volume and function. We feel that further research is needed before
exercise-induced QRS changes can be used to make
diagnostic decisions.
Acknowledgment
We are grateful to Steve Walaski and Wilson Rowley for their
technical assistance, Betsy Gilpin for the statistical analysis and to
Rosemary Montoya for the secretarial work.
References
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
1. Bonoris PE, Greenberg PS, Castellanet MJ, Ellestad MH:
Significance of changes in R wave amplitude during treadmill
stress testing: angiographic correlation. Am J Cardiol 41: 846,
1978
2. Bonoris PE, Greenberg PS, Christison GW, Castellanet MJ,
Ellestad MH: Evaluation of R wave amplitude changes versus
ST-segment depression in stress testing. Circulation 57: 904,
1978
3. Brody DA: A theoretical analysis of intracavitary blood mass
influence on the heart-lead relationship. Circ Res 4: 731, 1956
4. Katzeff IE, Edwards H: Exercise stress testing and an electromechanical S wave of the electrocardiogram. Does the S
wave voltage change with increasing work rate? S Afr Med J
49: 1088, 1974
5. Gottwik MG, Parisi AF, Askenazi J, McCaughan D: Computerized orthogonal electrocardiogram: relation of QRS to left
ventricular ejection fraction. Am J Cardiol 41: 9, 1978
6. Borer JS, Bacharach SL, 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, 1977
7. Borer JS, Bacharach SL, Green MV, Kent KM, Johnston GS,
Epstein SE: Effects of nitroglycerin on exercise induced abnormalities of left ventricular regional function and ejection fraction in coronary artery disease: assessment by radionuclide
cineangiography in symptomatic and asymptomatic patients.
Circulation 57: 314, 1978
8. Burow RD, Strauss HW, Singleton R, Pond M, Rehn T, Bailey
IK, Griffith LC, Nickoloff E, Pitt B: Analysis of left ventricular
function from multiple gated acquisition cardiac blood pool imaging: comparison to contrast angiography. Circulation 56:
1024, 1977
9. Pfisterer ME, Ricci DR, Schuler G, Swanson SS, Gordon DG,
Peterson KL, Ashburn WL: Validity of left ventricular ejection
fraction measured at rest and peak exercise by equilibrium
radionuclide angiography using short acquisition times. J Nucl
Med 20: 484, 1979
10. Folland ED, Hamilton GW, Larson SM, Kennedy JW,
Williams DL, Ritchie JL: The radionuclide ejection fraction: a
comparison of three radionuclide techniques with contrast
angiography. J Nucl Med 18: 1159, 1977
11. Slutsky R, Karliner J, Ricci D, Kaiser R, Pfisterer M, Gordon
D, Peterson K, Ashburn W: Left ventricular volumes by gated
equilibrium radionuclide angiography: a new method. Circulation 60: 556, 1979
12. Rautaharji PM, Wolf HK, Eifler WJ, Backburn H: A simple
procedure of positioning precordial ECG and VCG electrodes
using an electrode locator. J Electrocardiography 9: 35, 1976
13. Yerushalmy J: Statistical problems in assessing methods of
medical diagnosis with special reference to x-ray techniques.
Pub Health Rep 62: 1432, 1947
14. Youden WJ: Index of rating diagnostic tests. Cancer 3: 32, 1950
15. Kentala E, Heikkila J, Pyorala K: Variation of QRS amplitude
in exercise ECG as an index predicting result of physical training in patients with coronary heart disease. Acta Med Scand
194: 81, 1973
16. Kentala E, Luurila 0: Response of R wave amplitude to
posural changes and to exercise. Ann Clin Res 7: 258, 1975
17. Simoons ML, Hugenholtz PG: Gradual changes of ECG
waveform during and after exercise in normal subjects. Circula-
1013
tion 52: 570, 1975
18. Simoons ML, Van Den Brand M, Hugenholtz PG: Quantitative analysis of exercise electrocardiograms and left ventricular angiocardiograms in patients with abnormal QRS complexes at rest. Circulation 55: 55, 1977
19. Rushmer RF, Crystal DK, Wagner C, Ellis RM, Nash AA:
Continuous measurements of left ventricular dimensions in intact, unanesthetized dogs. Circ Res 2: 14, 1954
20. Erickson HH, Bishop VS, Kardon MB, Horwitz LD: Left ventricular internal diameter and cardiac function during exercise.
J Appl Physiol 30: 473, 1971
21. Braunwald E, Goldblatt A, Harrison DC, Mason DT: Studies
on cardiac dimensions in intact, unanesthetized man. III.
Effects of muscular exercise. Circ Res 13: 460, 1963
22. Gorlin R, Cohen LS, Elliot WC, Klein MD, Lane FJ: Effect of
supine exercise on left ventricular volume and oxygen consumption in man. Circulation 32: 361, 1965
23. Stein RA, Michielli D, Fox EL, Krasnow N: Continuous ventricular dimensions in man during supine exercise and recovery.
An echocardiographic study. Am J Cardiol 41: 655, 1978
24. Sharma B, Goodwin JF, Raphael MJ, Steiner RE, Rainbow
RG,Taylor SH: Left ventricular angiography on exercise. A
new method of assessing left ventricular function in ischemic
heart disease. Br Heart J 38: 59, 1976
25. Berger HJ, Reduto LA, Johnstone DE, Borkowski H, Sands
JM, Cohen LS, Langou RA, Gottschalk A, Zaret BL, Pytlik L:
Global and regional left ventricular response to bicycle exercise
in coronary artery disease. Assessment by quantitative
radionuclide angiography. Am J Med 66: 13, 1979
26. Rerych SK, Scholz PM, Newman GE, Sabiston DC Jr, Jones
RH: Cardiac function at rest and during exercise in normals
and in patients with coronary heart disease. Ann Surg 187: 449,
1978
27. Pfisterer ME, Battler A, Swanson SM, Slutsky R, Froelicher
VF, Ashburn WL: Reproducibility of ejection fraction determinations by equilibrium radionuclide angiography during exercise and in the recovery period. J Nucl Med 20: 491, 1979
28. Battler A, Ross J Jr, Slutsky R, Costello D, Pfisterer M,
Cerreto W, Ashburn W, Froelicher V: Improvement by oral
propranolol of exercise-induced ischemic dysfunction in
patients with coronary heart disease. (abstr) Am J Cardiol 43:
415, 1979
29. Battler A, Ross J Jr, Slutsky R, Gordon D, Ashburn W,
Froelicher VF: Causes for false negative radionuclide ejection
fraction response to exercise: propranolol and isolated right
coronary stenosis. (abstr) Clin Res 27: 2A, 1979
30. Battler A, Pfisterer M, Schuler G, Froelicher V: Exercise induced changes in QRS amplitude in the diagnosis of coronary
artery disease. (abstr) Circulation 58: (suppl II): 11-199, 1978
31. Nelson CV, Chatterjee M, Angelakos ET: Model studies on the
effect of the intracardiac blood on the electrocardiogram. Am
Heart J 62: 83, 1961
32. Bayley RH, Kalbfleisch JM, Berry PM: Change in the body's
QRS surface potentials produced by alterations in certain compartments of the nonhomogeneous conductivity model. Am
Heart J 77: 517, 1969
33. Voukydis PC: Effect of intracardiac blood on the electrocardiogram. N Engl J Med 291: 612, 1974
34. Horan LG, Flowers NC, Brody DA: Body surface potential distribution: comparison on naturally and artificially produced
signals as analyzed by digital computer. Circ Res 13: 373, 1963
35. Baxley WA, Dodge HT, Sandler H: A quantitative angiocardiographic study of left ventricular hypertrophy and the electrocardiogram. Circulation 37: 509, 1968
36. Vine DL, Finchum RN, Dodge HT, Bancroft WH Jr, Hurst
DC: Comparison of the vectorcardiogram with the electrocardiogram in the prediction of left ventricular size. Circulation
43: 547, 1971
37. Talbot S, Kilpatrick D, Jonathan A, Raphael MJ: QRS voltage
of the electrocardiogram and Frank vectorcardiogram in relation to ventricular volume. Br Heart J 39: 1109, 1977
38. Askenazi J, Parisi AF, Cohn PF, Freedman WB, Braunwald E:
Value of the QRS complex in assessing left ventricular ejection
fraction. Am J Cardiol 41: 494, 1978
Relationship of QRS amplitude changes during exercise to left ventricular function and
volumes and the diagnosis of coronary artery disease.
A Battler, V Froelicher, R Slutsky and W Ashburn
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Circulation. 1979;60:1004-1013
doi: 10.1161/01.CIR.60.5.1004
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1979 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/60/5/1004.citation
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/