Relationship of QRS Amplitude Changes During Exerciseto 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.

SUMMARY Preliminary studies have suggested that QRS-amplitude changes due to exercise-inducedalterations in ventricular volume and function can improve the diagnostic value of the exercise test. To evaluatethis, electrocardiographic data and equilibrium radionuclide angiographic images were recorded simul-taneously in 18 normal subjects and 60 coronary artery disease patients at rest and during supine bicycle exer-cise. In 24 of the 60 coronary artery disease patients, left ventricular volumes were also calculated. Themeasured QRS amplitudes were the R waves in V, X, Y and Z, the Q wave in Z and the sum of amplitudes ofR waves in X and Y and the Q wave in Z (2iR). The mean left ventricular ejection fraction increased signifi-cantly from rest to peak exercise in the normal subjects; however, the mean left ventricular ejection fractionand mean volumes did not change significantly in the coronary patients. There was no significant difference inthe mean QRS-amplitude changes during exercise between the coronary artery disease patients and the normalsubjects in any of the measured leads. The sensitivity and specificity of exercise-induced QRS-amplitudechanges for coronary disease were lower than ST-segment changes. For ST-segment changes, the sensitivitywas 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 sub-jects and coronary patients the correlations were fair (0.50, 0.51 respectively); however, the correlationbetween the magnitude of 2R and ejection fraction change from rest to peak exercise was poor and did not im-prove with any other measured QRS amplitudes or by separating normal subjects from coronary patients withand without previous myocardial infarction. There were also poor correlations between end-diastolic and end-systolic volumes to QRS amplitudes at rest, peak exercise and their magnitude of change from rest to peak ex-ercise. Thus, R-wave amplitude changes during exercise testing have little diagnostic value and are not relatedto exercise-induced changes in left ventricular function or volumes.

R-WAVE CHANGES in CM5 have been reported toincrease the diagnostic value of the exercise test. Anincrease in R-wave amplitude during exercise has beenreported to indicate severe left ventricular dysfunctionand coronary obstruction, while a decrease in R-waveamplitude is consistent with normal left ventricularfunction.1' 2 These changes are thought to be caused byan increase in end-diastolic volume during exerciseand the Brody effect.3 The frequent S-wave increase inV4 in normal subjects during exercise testing waspostulated to be the result of increased myocardialfunction.4 Resting left ventricular ejection fraction ob-tained by contrast angiography has demonstrated apositive correlation with the sum of various Q- and R-wave amplitudes,5 but this has not been studied duringexercise.The gated equilibrium radionuclide angiography

technique correlates well with angiographic studies ofejection fraction and is useful for detecting ejectionfraction responses to exercise.'-" With this method, itis also possible to calculate end-diastolic and end-

From the Division of Cardiology, Department of Medicine andthe 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 ShebaMedical Center, Tel Hashomer, Israel.Address for correspondence: V. Froelicher, M.D., Director of

Cardiac Rehabilitation, Department of Medicine, 225 DickinsonStreet, San Diego, California 92103.

Received March 12, 1979; accepted April 24, 1979.Circulation 60, No. 5, 1979.

systolic volumes independent of geometric assump-tions.'1 Thus, for the first time, QRS-amplitudechanges during exercise could be simultaneously cor-related with exercise-induced changes in left ven-tricular function and volumes. The aim of this studywas to test the diagnostic value of exercise-inducedQRS-amplitude changes and to evaluate thehypothesized mechanisms for these changes.

Methods

Patient Population

Sixty patients with coronary artery disease and 18control subjects were studied. The control subjectswere normal volunteers without clinical or electrocar-diographic evidence of cardiovascular or pulmonarydisease. The mean age was 41 years (range 30-57years). The patients with coronary artery disease weremales, and their mean age was 55 years (range 35-76years). Coronary artery disease was documented byprevious myocardial infarction (MI) at least 3 monthsbefore the study in 35 patients, coronary angiographyin 10 patients, and angina pectoris with abnormalECG and thallium treadmill test in the other 15patients. An abnormal treadmill test was defined asmore than 0.1 mV of flat or downsloping ST-segmentdepression or ST elevation in any of the 15 electrocar-diographic/vectorcardiographic (ECG/VCG) leadsother than aVR. Coronary artery disease was definedangiographically by the presence of at least a 70%decrease in the luminal diameter in one or more of themajor coronary arteries. An abnormal thallium test

1004

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

QRS AMPLITUDES AND LV FUNCTION/Battler et al.

was defined as a perfusion defect at peak exercise fill-ing during the recovery period or a perfusion defectremaining after redistribution. Ten of the coronaryartery disease patients were treated with antianginaldoses of propranolol (80-320 mg) and 15 of thepatients were treated with either nitroglycerin or iso-sorbide dinitrate. All normal subjects and coronaryartery disease patients were studied after they gavewritten informed consent.

Equipment and Imaging Techniques

Multiple ECG gated equilibrium radionuclide an-giography was performed after complete mixingthroughout the vascular space of 20-25 mCi 99mTctagged to human serum albumin, administered via aperipheral arm vein as previously described.6-9-Imag-ing was performed by positioning the detector of anAnger-type single-crystal scintillation camera over thechest in a 40-50° left anterior oblique and 5-10°caudad tilt in order to isolate the left ventricle op-timally from the left atrium. Data acquisition was ac-complished using a commercially available nuclearmedicine computer system, Medical Data System-PAD (MUGA), which divided the RR interval into 28equal time periods (frames) of 20-40-msec duration,depending on heart rate. It then assembled severalhundred heart cycles at corresponding times togenerate composite images throughout the cardiac cy-cle. In a 2-minute acquisition, usually 2500-7000counts could be accumulated in the left ventricularregion of interest at end-diastole in one 40-msec frame(counts corrected for background). To generate atime-activity curve, a rectangular region of interestwas arbitrarily placed around the left ventricle at end-diastole. A computer algorithm (MUGE) auto-matically determined the edge of the left ventricleusing a combination of the second derivative and acount-rate threshold of 5% per element as guidelines.Each subsequent frame was processed at the samethreshold so that throughout all 28 frames a"variable" region of interest was used to determinethe changing count rate within the left ventricle. Acomputer-assigned background region of interest out-side the lower quadrant of the left ventricle was usedto 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 cor-rected for background) and ES, = left ventricularcounts at end-systole (nadir of curve corrected forbackground). The ejection fraction calculated by thistechnique has been shown to correlate well with thosedetermined from biplane cineangiography, with a cor-relation coefficient in our laboratory of 0.91.9 Theinter- and intraobserver variations of ejection fractioncalculated by equilibrium radionuclide angiography inour laborRitory is 0.04 ejection fraction units. Thus, in

any single patient a significant change of ejection frac-tion was considered only when the change was at least+0.05 ejection fraction units. End-systolic and end-diastolic volumes were calculated from 24 coronaryartery disease patients using a radionuclide methodfree of geometric assumptions as previously describedby Slutsky et al.'1 Briefly, both at rest and peak exer-cise, the end-diastolic and end-systolic counts weredivided by the number of processed heart beats andnormalized for counts in milliliters of plasma drawnfrom a peripheral vein while the patient was at restand again at peak exercise. Using this technique thecorrelation of radionuclide angiographic volumes withbiplane contrast angiographic volumes was 0.98."

ECG/VCG Technique

Twelve ECG leads and three orthogonal Frankleads were obtained using 14 electrodes. Consistentelectrode placement was accomplished using theDalhouise square.'2 The chest electrodes were smallsilver 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 andcontinuously stores on floppy discs. In this way wecould obtain accurate peak exercise ECG/VCGchanges. ST-segment changes were considered in anyof the 15 ECG/VCG leads. To measure the QRSamplitudes, the isoelectric line was always defined asthe PR interval and the highest and lowest amplitudeswere averaged from consecutive beats during 10seconds at rest and during the last minute of exercise.The measured QRS amplitudes were R wave in V, Rwave in X, R wave in Y, R wave in Z, Q wave in Z, aswell as 2R,which is equal to RX + RY + QZ.

Study Protocol

Both the normal subjects and patients were studiedfirst in the supine basal state and then during a gradedsupine bicycle exercise with up to three levels of in-creasing work load, each lasting 3 minutes. The testwas 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 imagingtable supporting an electronically braked bicycleergometer (Uniwork 845T, Quinton Instruments) thatkept a constant work load over a pedal speed range of40-70 rpm. This allowed the work done by eachpatient at each level of exercise to be held constantover the 3 minutes. The ejection fraction and the 15scalar leads were recorded in the supine resting state,and then the supine bicycle exercise was performed.Equilibrium data were recorded for 5 minutes at restand continuously throughout the exercise period. Thedata from the last 2 minutes of the peak 3-minutework load level were analyzed to avoid the rapidchanges in heart rate and other hemodynamicvariables that usually occur in the first minute of a newexercise level and affect the accuracy of the ejectionfraction calculations. Blood pressure was taken withan arm cuff at rest and at peak exercise.

1005

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 60, No 5, NOVEMBER 1979

Statistical Analysis

Statistical analysis used was the t test for paired andnonpaired samples with a level of significance of<0.01. The correlation coefficients were significant ata level of <0.05. All the results are presented as themean ±SD. Sensitivity and specificity were defined asdescribed by Yerushalmy. 13 Briefly, sensitivity isdefined as the percentage of coronary disease patientswith an abnormal response TP/(TP + FN);specificity is defined as the percent of subjects withoutcoronary disease with a normal test TN/(TN + FP).To compare the combined sensitivity and specificity ofthe various ECG/VCG variables, Youden's J-summary index'4 was used, which has an algebraicstructure equivalent to the sum of sensitivity plusspecificity minus 1. The best summary index is closestto 1, while negative summary indices indicate littlediagnostic 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 ex-

ercise for the normals, all the coronary artery diseasepatients and the coronary artery disease patients withand without previous MI. The mean heart rate andsystolic blood pressure increased significantly fromrest to peak exercise in all the groups. The mean heartrates at rest were not significantly different betweenthe groups; however, the mean exercise heart rate wassignificantly higher in the normal subjects. In the nor-mals the mean heart rate increased by 72 beats/min,

TABLE 1. Mean (- SD) Heart Rates, Systolic Blood Pressures,

Coronary Artery Disease Patients at Rest and Peak Exercise

but in the coronary artery disease patients it increasedby only 52 beats/min (p < 0.001). There was nosignificant difference in the mean systolic bloodpressure at rest and peak exercise between the normalsand coronary artery disease patients. The mean ejec-tion fraction at rest was significantly higher(p < 0.001) in the normals (0.62 ± 0.06) than in allthe coronary artery disease patients (0.54 ± 0.11) orthe group with previous MI. (0.49 ± 0.11). However,in the coronary artery disease patients withoutprevious MI., the resting ejection fraction was similarto that of normals (0.62 ± 0.06). At peak exercise, thenormal subjects had a mean increase of 0.12 ejectionfraction units from the resting value. The coronaryartery disease patients did not significantly change themean ejection fraction from rest to peak exercise ineither those with or without a previous myocardial in-farction.

ECG/VCG Changes

The mean R-wave amplitudes in V5, X and ERdecreased significantly from rest to peak exercise bothin normals and in coronary artery disease patients.There was no significant difference in the mean R-wave amplitudes in V5 or X either at rest or peak exer-

cise 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 cor-onary artery disease patients (rest 1.98 ± 0.64 mV,peak exercise 1.68 ± 0.64 mV) and in patients withprevious MI (rest 1.79 ± 0.63 mV, peak exercise1.51 ± 0.56 mV), relative to the normal subjects (rest

Ejection Fractions and QRS Amplitudes of the Normal Subjects and

HR(beats/ SBP RVs RX RY RZ QZ ERmin) (mm Hg) EF (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 - 0.81Pex 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 - 0.67p <0.001 <0.001 <0.001 <0.005 <0.001 NS NS NS <0.001

CAD (n = 60)rest 64 - 10 131 - 17 0.54 0.11 1.38 = 0.50 1.06 - 0.34 0.64 0.33 0.75 - 0.30 0.25 0.21 1.98 |0.64Pex 116 = 21 173 - 25 0.55 0.14 1.14 - 0.41 0.87 = 0.31 0.60 0.32 0.80 - C.27 0.22 * 0.20 1.68 i0.64p <0.001 <0.001 NS <0.001 <0.001 NS NS <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 0.63Pex 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 0.56p <0.001 <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 0.58Pex 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

p <0.001 <0.001 NS <0.001 <0.001 NS <0.01 <0.001 <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.

1006 CIRCULATION

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

QRS AMPLITUDES AND LV FUNCTION/Battler et al.

-(DOW

so 0=

Nl CAD

ARX

0

o

-M

NL CAD

ARY

e

_

NI CAD

/ARZ

9

a

-cccm co

on"

NL CAD NI CADFIGURE 1. QRS-amplitude changes (Am V)from rest to peak supine bicycle exercisefor 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 horizontalline. There is no significant difference between normals and coronary artery disease patients in themagnitude 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) andthe patients without previous MI (rest 2.25 0.58mV, peak exercise 1.94 + 0.66 mV). Mean R-waveamplitude in Y did not change significantly from restto peak exercise in any of the groups of patients.However, at rest and peak exercise, the mean R-waveamplitude in Y was significantly lower in the coronaryartery disease patients compared with the normals.Mean R-wave amplitude in Z did not changesignificantly in normals, coronary artery diseasepatients or coronary artery disease patients withprevious MI; however, it did increase significantly inpatients without a previous MI. The trend of the Rwave in Z amplitude was to increase during exercise.The Q-wave amplitude in Z did not change sig-nificantly in the normals, however, it did decrease sig-nificantly in the coronary artery disease patients. Themean resting and peak exercise Q waves in Z were sig-nificantly lower in amplitude in the coronary arterydisease patients than in the normals.

Figure 1 demonstrates the distribution of QRS-amplitude changes from rest to peak exercise betweennormals and coronary artery disease patients. An in-crease or no change in R wave was considered to be a

abnormal response to exercise; a decrease was con-

sidered normal. However, the Q wave in the Frankscalar Z lead corresponds to the R wave in lead V2 ofthe standard ECG, while the R wave of the Frankscalar Z lead corresponds to the S wave in lead V2.Thus for lead Z, any decrease or no change in the Rwave at peak exercise was considered an abnormalresponse and any increase in RZ at peak exercise wasconsidered a normal response. The R wave in V5 in-

creased or did not change at peak exercise comparedwith the preexercise amplitude in only 12 of the 60coronary artery disease patients (for a sensitivity of20%) and decreased in 14 of the 18 normals (for aspecificity of 78%). The R wave in X increased or didnot change in 10 of the 60 coronary artery diseasepatients (for a sensitivity of 17%) and decreased in 14of the 18 normals (for a specificity of 78%). The Rwave in Y increased or did not change at peak exercisein 30 of the 60 coronary artery disease patients (for asensitivity of 50%) but decreased in only six of the 18normal subjects (for a specificity of 34%). The R wavein Z decreased or did not change in 29 of the 60 cor-onary artery disease patients (sensitivity 48%) and in-creased 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 (sensitivity15%) and decreased in 16 of 18 normals (specificity89%). Thirty-four of the 60 coronary artery diseasepatients (sensitivity 57%) had abnormal ST changes atpeak exercise, while none of the normals showedsimilar changes (specificity 100%). Thus, the sen-sitivity and specificity of ST changes at peak exercisewere better than any of the QRS-amplitude changesand according to the summary index, the QRSchanges reflecting the best sensitivity and specificityoccurred in RZ (table 2).

Radionuclide Ejection Fraction Changes

Figure 2 shows the changes in ejection fraction fromrest to peak exercise for each of the 18 normal subjectsand each of the 60 coronary artery disease patients.

1.5 r

0.5 _

EI1

o

-0.5

-1

-1.5

1007

l

00

0

-06 -Q=W-

--cocamomoac-

ow

0

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

Voi 60, No 5, NOVEMBER 1979

TABLE 2. The Sensitivity, Specificity and Summary Index of the QRS Amplitude and ST-segment Criteria forDiagnosing Coronary Disease.

RV. R.X RY RZ ST

Sensitivity 0.20 (12/60) 0.17 (10/60) 0.30 (30/60) 0.48 (29/60) 0.15 (9/60) 0.,7 (34/60)Specificity 0.78 (14/18) 0.78 (14/18) 0.34 (6/18) 0.67 (12/18) 0.89 (16/18) 1.00 (18/18)Summary

index ND ND ND 0.15 0.04 0.57

Abbreviations: SI = summary index; ND = negative values for the SI indicating that the measurementswere nondiagnostic.

The coronary artery disease patients had a sig-nificantly lower mean resting ejection fraction thanthe normals. At peak exercise the mean ejection frac-tion increased significantly in the normals and did notchange in the coronary artery disease patients. Everynormal subject increased the peak exercise ejectionfraction from a normal resting ejection fraction(>0.50) by at least 0.05 ejection fraction units. Of thecoronary artery disease patients, 44 had either adecrease or change in the peak exercise ejection frac-tion from rest.Of the 16 coronary artery disease patients that in-

creased peak exercise ejection fraction by at least 0.05ejection fraction units, four had an abnormal restingejection fraction (<0.50),- three were taking an-tianginal doses of propranolol and three had isolatedright coronary artery stenosis. Of the remaining six,three stopped exercise because of fatigue, and threebecause of angina or significant ST changes.

90 F

.80 K

.70 1

.60 F

NLn 18

CADn = 60

Volume Changes

In figure 3 the end-diastolic and end-systolic volumechanges from rest to peak exercise for 24 of the cor-onary artery disease patients are presented. Both themean end-diastolic and end-systolic volumes did notchange significantly from rest to peak exercise.However, nine patients had increased end-diastolicvolume from rest to peak exercise, of whom seven(78%) had accompanying angina or abnormal STchanges at peak exercise, compared with only five ofthe 15 patients (33%) that decreased or did not change

CAO n= 24220 r-

EDV ESV200 H

180 H

160 V

I1

.50

.40 p <0.001

140 kE

L*J

-j0

I

120 F

100

80

60

40

.30 F

20 L

.20 L

REST PEX REST PEX

FIGURE 2. Ejection fraction changes from rest to peaksupine 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 asmean ± SD.

REST PEX

T7

REST PEX

FIGURE 3. End-diastolic volume (EDV) and end-systolicvolume (ESV) changesfrom rest to peak supine bicycle exer-

cise (PEX) in 24 of the coronary artery disease patients. Theaverage values are presented as the mean ± SD. There is nosignificant change in the mean EDV and ESVfrom rest topeak exercise.

C)

cr

l-l

CC

C-,)LU)

LU

1008 ClIRCULATION

. --I

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

QRS AMPLITUDES AND LV FUNCTION/Battler et al.

NL + CAD n = 78

06

0. rGL0 4-

22

01

Di-

u LBUsBE5BXY 62 2.6R

NL n= 18

RV5 RX RY RZ 2R

CAD n = 60

tL7

= RESTC] PEAK EXERCISEMmA* p 0 05

RVs RX RY 6RZ YR

CAO WITH PREVIOUS Ml n = 35

- RE B 62

RVs RX RY .RZ Y. R

CAD WITHOUT PREVIOUS MI n = 25

RV5 RX RY RZ ' R

FIGURE 4. The correlation coefficients for the least-squares fit regressions of ejection fraction vs QRSamplitudes (R waves in V5, X, Y, Z and ER) at restl and peak exercise and the magnitude ofchange fromrest to peak exercise (A) for both the normal subjects (NL) and coronary artery disease patients (CA D), forthe NL alone, for the CAD alone andfor the CAD patients with and without previous myocardial infarction(MI).

end-diastolic volume at peak exercise. Of the ninepatients who had increased end-diastolic volume atpeak exercise only one also had an increase in one ofthe QRS amplitudes (RV5, from 1.1 mV at rest to 1.2mV at peak exercise); the other seven patientsdecreased all QRS amplitudes measured at peak exer-cise.

Correlation of Ejection Fraction to QRS Amplitudes

Figure 4 shows the correlation coefficients betweenejection fractions and QRS amplitudes at rest andpeak exercise and their magnitude of change from restto peak exercise for normal subjects and coronaryartery disease patients together, for normal subjects,for all the coronary artery disease patients, and for thepatients with and without previous MI. In all thesefive groups there was a fair correlation between mostof the R-wave amplitudes and ejection fraction both atrest and peak exercise. In most of the cases at peak ex-ercise the correlation coefficients were -somewhatlower than at rest. However, the correlation betweenthe magnitude of change of ejection fractions andchange of QRS amplitudes during exercise was neversignificant. The best correlations were obtained in thecombined normals and coronary artery diseasepatients at rest (r = 0.50) and peak exercise(r = 0.51). Figure 5 is an example of the correlationobtained from both the normals and the coronaryartery disease patients for the 2R at rest and peak ex-ercise and the magnitude of changes from rest to peakexercise. At rest, the correlation coefficient was 0.50and at peak exercise it was 0.51. Both these cor-relations were significant (p < 0.001). However, thecorrelation coefficient between the change of ejectionfraction and 2 R amplitude changes from rest to peakexercise was not significant. Fourteen subjects had a2R amplitude of less than 1.5 mV, and only two ofthem were normals; eight of the 12 coronary arterydisease patients in this group also had resting ejectionfractions of less than 0.50 (fig. 5, rest). At peak exer-cise there were 24 coronary artery disease patients andtwo normnal subjects who had 2;R amplitude of lessthan 1.5 mV and 14 of them, all with coronary artery

disease, also had reduced peak ejection fractions ofless than 0.50 (fig. 5, peak exercise).

Volumes and QRS Amplitudes Correlations

In figure 6, the correlation coefficients between end-diastolic and end-systolic volumes and R-waveamplitudes are presented at rest, peak exercise and forthe magnitude of change from rest to peak exercise.All these correlations were poor, although they wereagain somewhat better at rest and peak exercise thanfor the magnitude of change from rest -to peakexercise.

QRS Changes During Treadmill Testing

To test if similar results would be obtained, 25 ofthe coronary artery disease patients had a treadmillexercise test done within 24 hours of the supine bicycleexercise test. R-wave amplitudes in V5 were measuredat rest and at maximal treadmill exercise. During thetreadmill exercise test the mean R-wave amplitude inV5 decreased significantly (p < 0.001) from 1.35 +0.63 mV at rest to 1.16 ± 0.53 mV at peak exercise(fig. 7).

DiscussionRecently, Bonoris et al." 2 recommended the use of

R-wave amplitude changes during exercise to improvethe sensitivity and specificity of the exercise test. Aspreviously hypothesized,3' , 15-18 they suggested thatexercise-induced changes in left ventricular functionand volume cause the QRS-amplitude changes.However, these theories have never been validated,mainly because of technical problems involved insimultaneously studying left ventricular function,volume and QRS amplitude during exercise. Nuclearcardiology and newer ECG techniques make it possi-ble to measure dynamic changes in left ventricularfunction, volumes and the ECG continuously duringexercise. The equilibrium radionuclide angiographytechnique has been extensively validated by us9 andothers6-8, 10 and found to be very accurate. Thus, forthe first time, the relationship of exercise-induced

1009

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 60, No 5, NOVEMBER 1979

PEAK EXERCISE - REST (,A).30 r

.20 -8

0.

o o 8 o

O C) C

50y a 0407 +0074x)2p<0 00

0FCD

z0

C-LLU-

LUJ

.10o

0

-.10 F

-20 1

-30

-3 -2 -1 U 1 2 3mVAIR

.30

.20

A

r= 51y = 0.39+ 0.IIx)102p< 0.01

L

0.5 1 1.5 2 2.5 3 3.5 4 4.532R mV

FIGURE 5. The correlation between ejection fraction and mV E2R amplitude at peak exercise and at rest (left) and themagnitude of change from rest to peak exercise (A) (right). Closed circles represent normal subjects (NL) and open circlesrepresent coronary artery disease patients (CAD).

0.5 EDV ESV0.4 -

0.3

0.2-

0.1

CD,

00.

CC -0.2

C- -0.3r

RV5 RX RY RZ YR RV5 RX RY RZ IRO 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 changefrom rest to peak exercise (A).

r90PEAK EXERCISE

80

70

.60~;z0

4C-, .50

40

30~

20

.90

* O -

..o8°0. #.000.* 0000 0

0S,O§° 0

0 0000

0

REST.80

70

60

z

cr-

C3

.50

* NL n = 780 CAD -0.

r= -0.09

1010 CIRCULATION

0 0 9 c's

c&0 0.

0 ,,

al-<.O 0

...0

- - 8 *" c

)0 0 0 0

L"

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

QRS AMPLITUDES AND LV FUNCTION/Battler et al.

RV5n= 25

2.5

2.0

E 1.5

1.0

0.5

c

p <0.001

REST Peak TMExercise

FIGURE 7. R-wave amplitude changes in V5 (mV) fromrest 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 theelectrocardiogram could be assessed.The magnitude and direction of exercise-induced

volume changes either in normals or coronary patientsis still controversial. Rushmer et al.19 used micro-crystal ultrasound transducers in conscious dogs andconcluded that the heart dimensions often becomesmaller during exertion. However, using the same

technique in swimming dogs, Erickson et al.20 showedan increase in end-diastolic left ventricular diameterduring exercise. Using epicardial metal markers,Braunwald et al.2' showed only small changes in bothend-diastolic and end-systolic dimensions duringsupine exercise. Using the thermodilution technique innormals and patients with a variety of heart diseases,Gorlin et al.22 found variable changes in end-diastolicvolume during supine exercise. Using the echo-cardiographic technique, Stein et al.23 demonstrated a

slight decrease in end-systolic dimension without a sig-nificant change in end-diastolic dimension duringsupine exercise in normal subjects. Using contrastangiography24 and nuclear angiography25 it was dem-onstrated that coronary artery disease patients whodeveloped angina during supine exercise usually in-creased both the end-diastolic and end-systolicvolumes, while patients who did not develop anginausually did not change significantly the left ventricularvolumes. In another study, using erect bicycle exer-cise, both normals and coronary artery diseasepatients increased the end-diastolic volume during ex-

26ercise. Thus, previous studies suggest that differencesin exercise position and severity of coronary diseaseaccount for different volume responses to exercise. Inthe present study the 24 coronary artery diseasepatients in whom we could obtain volumes using theradionuclear angiography technique did not changesignificantly either the mean end-diastolic or end-systolic volumes from rest to peak supine exercise.However, of nine patients that increased end-diastolicvolume during exercise, seven developed angina orsignificant ST changes at peak exercise.

Ejection fraction changes during exercise have beenextensively investigated with radionuclide angiog-raphy. Most investigators agree that ejection fractionincreases during exercise in normal subjects, whilemost coronary artery disease patients show no changeor a decrease.6 9, 25-27 In the present study the normalsubjects increased the ejection fraction significantly bya mean of 0.12 ejection fraction units and each subjectincreased the ejection fraction by at least 0.05; thecoronary artery disease patients did not significantlychange the mean ejection fraction from rest to peakexercise. However, 16 of the 60 coronary artery dis-ease patients increased the ejection fraction at peakexercise by more than 0.05. Of these 16 patients, fourhad an abnormal resting ejection fraction (<0.50),which as an isolated finding is diagnostic of left ven-tricular dysfunction. Of the other 12 patients, threewere taking antianginal doses of propranolol andthree had isolated right coronary lesions, both causesof a false-negative ejection fraction response to exer-cise as shown previously by our group.28' 29 Thus theother six patients (10%) are true false-negativeresponders. Of the six, three patients did not developangina or ST changes during exercise and stoppedbecause of fatigue. We can speculate that in thesepatients the exercise challenge was not enough tocause a mismatch between oxygen demand andsupply.QRS-amplitude changes during exercise are as con-

troversial as volume changes. Bonoris et al.2 measuredexercise-induced R-wave changes in CM5 in a highlyselected group of coronary artery disease patients andin normal subjects. Most of these patients were false-positive or false-negative responders by ST-segmentcriteria. They defined an abnormal response as an in-crease or no change in R-wave amplitude and a nor-mal response as a decrease in R-wave amplitude at

1011

A I by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

VOL 60, No 5, NOVEMBER 1979

peak exercise from rest. Using these criteria, the sen-sitivity and specificity of R-wave amplitude changesduring exercise were better than ST changes. Theyalso compared R-wave changes during exercise to theseverity of angiographic lesions and concluded thatexercise-induced increases in R-wave amplitude oc-curred with left ventricular dysfunction.' Katzeff andEdwards4 predicted an electromechanical associationbetween S-wave amplitudes and cardiac mechanicalfunction. In normal subjects, they reported an in-creasing S-wave voltage in a horizontal bipolar leadwith increasing work load and concluded that theamplitude of the S wave in certain leads acts as an in-dicator of left ventricular function. Kentala et al.'5studied the variation of R wave from V5 duringprogressive bicycle exercise testing as an index topredict the results of physical training in patients withcoronary artery disease. They concluded that the R-wave changes were greatest in the cardiac patientswho responded to exercise training, and the smallerthe dynamic changes of the QRS amplitude during ex-ercise testing the worse the training results. In anotherpaper Kentala and Luurila16 showed that R-waveamplitude changes occurred with postural changesand exercise in healthy subjects and patients survivingacute MI.

In the present study, R-wave changes during exer-cise had the same directions in normal subjects as incoronary artery disease patients. In both normals andunselected coronary artery disease patients, the sen-sitivity and specificity of QRS-amplitude changeswere less than for ST changes. No single lead couldseparate normals and coronary patients using the R-wave amplitude criteria, including RV5 and RX.Among the nine patients who increased end-diastolicvolume at peak exercise, only one also increased theR-wave amplitude in V5 (from 1.1 mV at rest to 1.2mV at peak exercise). A possible explanation for thesedifferent results could be that we used supine bicycleexercise rather than treadmill or erect bicycle exercise.To examine this, 25 of the coronary artery diseasepatients 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 ± .53mV at peak exercise (p < 0.001) and only two patientsincreased the amplitude of R wave in V5 from rest topeak exercise. Another possible explanation is thatothers may only allow their coronary patients withsevere disease to perform submaximal exercise. Dur-ing submaximal exercise, most subjects have an in-creased R-wave amplitude, but decrease it at higherwork loads. As we described previously,30 exercise-induced R-wave changes in the Z vectorcardiogramlead had somewhat better sensitivity and specificitythan amplitudes measured in other leads. Using thesum of R-wave amplitudes from X, Y and Z did notimprove the sensitivity or specificity of the test.The relationship between QRS amplitudes and left

ventricular volumes and function at rest has been ex-tensively investigated. The Brody effect,3 which ex-

plains the effect of intracardiac blood on QRS poten-tials by augmentation of the initial portion of the QRScomplex and attenuation of the latter portion, wasderived from models3' 33 and animal experiments.34However, human studies have not all agreed with thistheory of why an increase in left ventricular end-diastolic volume should increase R-wave amplitude.Baxley et al.3' and Vine et al.36 found a positive cor-relation between QRS voltage and left ventricularend-diastolic volume while Talbot et al.37 found an in-verse 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 ob-tained from the orthogonal Frank system with leftventricular ejection fraction measured by contrastangiography. They found that the sum of RX + RY +QZ correlated significantly with ejection fraction, andthis correlation even improved with the ejection frac-tion of post extrasystolic beats. However, these studieswere not simultaneous ECG-function-volume studiesand were done only at rest. In the present study wefound a fair positive correlation at rest and peak exer-cise between R waves and ejection fractions and, as inthe previous works' 38 we also found that the best cor-relations were obtained using 2R. However, when wecorrelated the magnitude of exercise-induced changesin QRS amplitudes in the various leads to ejectionfraction changes, the correlations were not significant.For left ventricular volumes, there were no significantcorrelations either at rest or peak exercise or betweenthe magnitude of changes between rest and peak exer-cise. We could not find any improvement in the cor-relation coefficients when we separated the normalsfrom coronary patients, and coronary patients with orwithout previous MI. Thus, although we agree thatthere is a positive correlation between R-waveamplitude and left ventricular ejection fraction eitherat rest or peak exercise, this relationship does not ex-plain the changes that occur in the QRS amplitudesduring exercise.Other mechanisms than changes in left ventricular

volume and function must be responsible for exercise-induced QRS-amplitude changes. Changes in wallthickness during exercise could result in alterations insize or number of individual muscle fiber dipoles.Changes in the position of the heart in the thorax dur-ing exercise and in the amount of air in the lungs dur-ing exercise can affect the conductivity of the chest.These are hypothesized mechanisms that requirefurther investigation.

In summary, during maximal exercise there is asignificant increase in ejection fraction in normals, nochange or a decrease in ejection fraction in most of thecoronary artery disease patients, no significant volumechanges in coronary patients who are not developingangina pectoris, and a similar decrease in R-waveamplitude during maximal exercise both in normalsand coronary artery disease patients.

Thus, our study has not confirmed the direction ofexercise-induced QRS changes in coronary patients or

1012 CIRCULATION

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

QRS AMPLITUDES AND LV FUNCTION/Battler et al.

their relationship to left ventricular volume and func-tion. We feel that further research is needed beforeexercise-induced QRS changes can be used to makediagnostic decisions.

Acknowledgment

We are grateful to Steve Walaski and Wilson Rowley for theirtechnical assistance, Betsy Gilpin for the statistical analysis and toRosemary Montoya for the secretarial work.

References1. Bonoris PE, Greenberg PS, Castellanet MJ, Ellestad MH:

Significance of changes in R wave amplitude during treadmillstress 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 versusST-segment depression in stress testing. Circulation 57: 904,1978

3. Brody DA: A theoretical analysis of intracavitary blood massinfluence on the heart-lead relationship. Circ Res 4: 731, 1956

4. Katzeff IE, Edwards H: Exercise stress testing and an elec-tromechanical S wave of the electrocardiogram. Does the Swave voltage change with increasing work rate? S Afr Med J49: 1088, 1974

5. Gottwik MG, Parisi AF, Askenazi J, McCaughan D: Com-puterized orthogonal electrocardiogram: relation ofQRS to leftventricular 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 thenoninvasive evaluation of global and regional left ventricularfunction at rest and during exercise in patients with coronaryartery 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 abnor-malities of left ventricular regional function and ejection frac-tion in coronary artery disease: assessment by radionuclidecineangiography in symptomatic and asymptomatic patients.Circulation 57: 314, 1978

8. Burow RD, Strauss HW, Singleton R, Pond M, Rehn T, BaileyIK, Griffith LC, Nickoloff E, Pitt B: Analysis of left ventricularfunction from multiple gated acquisition cardiac blood pool im-aging: 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 ejectionfraction measured at rest and peak exercise by equilibriumradionuclide angiography using short acquisition times. J NuclMed 20: 484, 1979

10. Folland ED, Hamilton GW, Larson SM, Kennedy JW,Williams DL, Ritchie JL: The radionuclide ejection fraction: acomparison of three radionuclide techniques with contrastangiography. J Nucl Med 18: 1159, 1977

11. Slutsky R, Karliner J, Ricci D, Kaiser R, Pfisterer M, GordonD, Peterson K, Ashburn W: Left ventricular volumes by gatedequilibrium radionuclide angiography: a new method. Circula-tion 60: 556, 1979

12. Rautaharji PM, Wolf HK, Eifler WJ, Backburn H: A simpleprocedure of positioning precordial ECG and VCG electrodesusing an electrode locator. J Electrocardiography 9: 35, 1976

13. Yerushalmy J: Statistical problems in assessing methods ofmedical 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, 195015. Kentala E, Heikkila J, Pyorala K: Variation of QRS amplitude

in exercise ECG as an index predicting result of physical train-ing in patients with coronary heart disease. Acta Med Scand194: 81, 1973

16. Kentala E, Luurila 0: Response of R wave amplitude toposural changes and to exercise. Ann Clin Res 7: 258, 1975

17. Simoons ML, Hugenholtz PG: Gradual changes of ECGwaveform during and after exercise in normal subjects. Circula-

tion 52: 570, 197518. Simoons ML, Van Den Brand M, Hugenholtz PG: Quan-

titative analysis of exercise electrocardiograms and left ven-tricular angiocardiograms in patients with abnormal QRS com-plexes at rest. Circulation 55: 55, 1977

19. Rushmer RF, Crystal DK, Wagner C, Ellis RM, Nash AA:Continuous measurements of left ventricular dimensions in in-tact, unanesthetized dogs. Circ Res 2: 14, 1954

20. Erickson HH, Bishop VS, Kardon MB, Horwitz LD: Left ven-tricular internal diameter and cardiac function during exercise.J Appl Physiol 30: 473, 1971

21. Braunwald E, Goldblatt A, Harrison DC, Mason DT: Studieson 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 ofsupine exercise on left ventricular volume and oxygen consump-tion in man. Circulation 32: 361, 1965

23. Stein RA, Michielli D, Fox EL, Krasnow N: Continuous ven-tricular 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, RainbowRG,Taylor SH: Left ventricular angiography on exercise. Anew method of assessing left ventricular function in ischemicheart disease. Br Heart J 38: 59, 1976

25. Berger HJ, Reduto LA, Johnstone DE, Borkowski H, SandsJM, Cohen LS, Langou RA, Gottschalk A, Zaret BL, Pytlik L:Global and regional left ventricular response to bicycle exercisein coronary artery disease. Assessment by quantitativeradionuclide angiography. Am J Med 66: 13, 1979

26. Rerych SK, Scholz PM, Newman GE, Sabiston DC Jr, JonesRH: Cardiac function at rest and during exercise in normalsand in patients with coronary heart disease. Ann Surg 187: 449,1978

27. Pfisterer ME, Battler A, Swanson SM, Slutsky R, FroelicherVF, Ashburn WL: Reproducibility of ejection fraction deter-minations by equilibrium radionuclide angiography during ex-ercise 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 oralpropranolol of exercise-induced ischemic dysfunction inpatients 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 ejectionfraction response to exercise: propranolol and isolated rightcoronary stenosis. (abstr) Clin Res 27: 2A, 1979

30. Battler A, Pfisterer M, Schuler G, Froelicher V: Exercise in-duced changes in QRS amplitude in the diagnosis of coronaryartery disease. (abstr) Circulation 58: (suppl II): 11-199, 1978

31. Nelson CV, Chatterjee M, Angelakos ET: Model studies on theeffect of the intracardiac blood on the electrocardiogram. AmHeart J 62: 83, 1961

32. Bayley RH, Kalbfleisch JM, Berry PM: Change in the body'sQRS surface potentials produced by alterations in certain com-partments of the nonhomogeneous conductivity model. AmHeart J 77: 517, 1969

33. Voukydis PC: Effect of intracardiac blood on the electrocar-diogram. N Engl J Med 291: 612, 1974

34. Horan LG, Flowers NC, Brody DA: Body surface potential dis-tribution: comparison on naturally and artificially producedsignals as analyzed by digital computer. Circ Res 13: 373, 1963

35. Baxley WA, Dodge HT, Sandler H: A quantitative angiocar-diographic study of left ventricular hypertrophy and the elec-trocardiogram. Circulation 37: 509, 1968

36. Vine DL, Finchum RN, Dodge HT, Bancroft WH Jr, HurstDC: Comparison of the vectorcardiogram with the electrocar-diogram in the prediction of left ventricular size. Circulation43: 547, 1971

37. Talbot S, Kilpatrick D, Jonathan A, Raphael MJ: QRS voltageof the electrocardiogram and Frank vectorcardiogram in rela-tion 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 ejectionfraction. Am J Cardiol 41: 494, 1978

1013

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from

A Battler, V Froelicher, R Slutsky and W Ashburnvolumes and the diagnosis of coronary artery disease.

Relationship of QRS amplitude changes during exercise to left ventricular function and

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1979 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.60.5.1004

1979;60:1004-1013Circulation. 

http://circ.ahajournals.org/content/60/5/1004.citationthe World Wide Web at:

The online version of this article, along with updated information and services, is located on

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer information about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. FurtherEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculationpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:

by guest on May 10, 2018

http://circ.ahajournals.org/D

ownloaded from