PATHOPHYSIOLOGY AND NATURAL HISTORYEXERCISE TESTING

The effect of left ventricular systolic function on

maximal aerobic exercise capacity in asymptomaticpatients with coronary artery disease*ALl A. EHSANI, M.D., DANIEL BIELLO, M.D., DOUGLAS R. SEALS, PH.D., MARK B. AUSTIN, M.D.,AND JOAN SCHULTZ, M.S.

ABSTRACT The purpose of this study was to examine the relationship between maximal 02 uptake(VO2max) and left ventricular systolic function in patients with coronary artery disease. We studied 27patients, age 50 10 years (mean + SD), who were asymptomatic and able to attain true VO2max.VO2max was defined by the leveling-off criterion and/or a respiratory exchange ratio of 1.15 or greater.Left ventricular ejection fraction was determined, by gated cardiac blood pool imaging. In patientswhose ejection fraction decreased with exercise, VO2max was 21 + 4 vs 27 + 4 ml/kg/min in thosewhose ejection fraction increased (p < .001). Systolic blood pressure/end-systolic volume relation wasshifted upward and to the right in the former group in response to peak exercise. In contrast, thepressure-volume relation was shifted upward and to the left in patients whose ejection fraction in-creased with exercise. Ejection fraction at rest did not correlate with VO2max. There was a significantbut weak correlation between peak exercise ejection fraction and VO2max (r .43, p < .025). Leftventricular exercise reserve, i.e., the change in ejection fraction from rest to exercise, correlated withVO2max (r .77, p < .0002), maximal 02 pulse (r = .50, p < .005), and maximal heart rate duringtreadmill exercise (r = .61, p < .001). Maximal heart rate during treadmill exercise correlated withVO2max (r = .70, p < .0002). These data suggest that impaired left ventricular function can limitVO2max and that maximal heart rate and left ventricular exercise reserve are among the variablesaffecting VO2max in patients with coronary artery disease who are not limited by angina.Circulation 70, No. 4, 552-560, 1984.

MAXIMAL 02 uptake capacity (VO2max), the bestavailable objective measure of aerobic exercise capac-ity, 1 is generally lower in patients with coronary arterydisease than in age-matched healthy subjects.2 Manyof these patients also exhibit impaired left ventricularfunction in response to exercise.4 5 However, recentstudies have reported a poor correlation betweenVo2max and left ventricular performance.6-' Further-more, pharmacologic interventions that improve left

From the Section of Applied Physiology and the CardiovascularDivision, Department of Medicine, the Irene Walter Johnson Institute ofRehabilitation, and the Division of Nuclear Medicine, the Edward Mal-linckrodt Institute of Radiology, Washington University School ofMedicine, St. Louis.

Supported by NIH Research Grant HL22215.Address for correspondence: Ali A. Ehsani, M.D., Department of

Medicine, Washington University School of Medicine, 4566 ScottAve., St. Louis, MO 63110.

Received Nov. 30, 1983; revision accepted July 5, 1984.Dr. Seals is a Postdoctoral Research Trainee supported by Institution-

al National Service Award AG-00078.*All editorial decisions for this article, including selection of review-

ers and the final disposition, were made by a guest editor. This proce-dure applies to all manuscripts with authors from the Washington Uni-versity School of Medicine.

552

ventricular function in patients with heart failure maynot necessarily result in increased maximal exercisecapacity.9 This would imply that left ventricular con-tractile function has little if any effect on maximalexercise capacity in cardiac patients. However, sinceVO max is sensitive to changes in cardiac output'0 andleft ventricular contractile function is one of the majordeterminants regulating cardiac output during maximalexercise,"l our hypothesis was that changes in left ven-tricular ejection fraction should influence VO2max. Totest this hypothesis, we believed it was imperative touse true VO2max rather than peak VOX as a marker ofmaximal aerobic exercise capacity because, unlikepeak V07, VO2max is not affected by subjects' moti-vation. Therefore, in this study we examined the effectof left ventricular systolic function on maximal aerobicexercise capacity in patients with coronary artery dis-ease who were able to attain true VO2max.

MethodsPatient selection. Twenty-seven patients (age 50 + 10 years

[mean + SD], 26 men and one woman) who met the following

CIRCULATION

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

PATHOPHYSIOLOGY AND NATURAL HISTORY-EXERCISE TESTING

criteria were selected from among 67 patients with coronaryartery disease. Since the purpose of this study was specificallyto assess the relationship between 'VO2max and left ventricularejection fraction, only those patients who were able to reach trueVO2max were selected. VO2max was defined as (1) attainmentof a plateau of V02 (<100 ml/min change in V02) despite anincrease in work rate and/or (2) a respiratory exchange ratio of1. 15 or greater, indicating marked hyperventilation. 12 Further-more, only those patients who were able to exercise to maxi-mum in the recumbent position, defined as achieving (1) at least80% of the maximal heart rate recorded during treadmill testand/or (2) 95% or greater of the measured maximal treadmillrate pressure product, were included. The rationale for choosing80% of maximum heart rate was that the maximum possible 02uptake during supine bicycle ergometer exercise is generallyless than 80% of the true VO,max even in healthy subjects.'3This was confirmed by V02 measurements made during supinebicycle ergometer exercise tests in a separate group of six pa-tients with ischemic heart disease. The maximal attainable V02for the bicycle exercise test averaged 77 ± 9% of VO2max.Myocardial 02 consumption (MV02), however, was unlikely tobe lower during supine bicycle exercise than treadmill exerciseas reflected by the values for rate pressure product (table 1,figure 1). All patients gave written informed consent. The studyprotocol has been reviewed and approved by the Human StudiesCommittee of Washington University School of Medicine.The clinical characteristics of the patients are summarized in

table 1. All patients had prior myocardial infarctions. The inter-val between myocardial infarction and study was at least 3months (table 1). Of the 15 patients who underwent coronaryarteriography, six had single-vessel disease, three had double-vessel disease, and six had triple-vessel disease. Although threepatients had previous histories of angina, none was limited byangina during maximal exercise tests and all were able to attaintrue MO2max.

Ten of 27 patients were taking propranolol with a total dailydose averaging 109 + 83 mg/day (table 1). Three patients wereon nadolol, three on timolol, and one on atenolol (table 1). Allexercise tests were performed 4 to 6 hr after the last dose of a ,B-blocking drug. We did not discontinue fl-blockade therapy be-fore exercise tests because we believed it would not be justifiedor safe to subject the patients to strenuous exercise necessary toelicit true VO2max (respiratory exchange ratio = 1.27 ± 0. 10)after withdrawal of propranolol.

Treadmill exercise test and M02max. The exercise testingwas performed on a motor-driven treadmill (Quinton Instru-ments, Seattle). Initially each subject underwent a maximalexercise treadmill test according to the Bruce protocol. 14 One or2 weeks later each patient underwent another maximal treadmilltest for measurement of VO2max by means of a modified Bruceprotocol as previously described. 15 The patients breathedthrough a Daniels valve, and expired gases were collected intoneoprene meteorologic balloons at 60 or 30 sec consecutiveintervals. 02 and CO2 were analyzed with a mass spectrometer(Perkin-Elmer MGA1 100). Expired volumes were measuredwith a Tissot spirometer. The exercise end point for all patientswas severe fatigue and dyspnea.

Assessment of left ventricular function at rest and duringexercise. Left ventricular performance was assessed by electro-cardiogram-gated cardiac blood pool imaging with erythrocyteslabeled in vivo by intravenous injection of 7.7 mg of stannouspyrophosphate followed 20 min later by 25 mCi of 99n'Tc givenintravenously. The images were obtained with a standard field-of-view scintillation camera (Siemens LEM) equipped with a0.64 cm thick Nal crystal and with a low-energy, medium-resolution, parallel-hole collimator. The patients were imagedin the supine position with the scintillation camera positioned in

TABLE 1Clinical characteristics of patients

,B-AdrenergicAge (yr)/ Prior MI blockade

Patient sex (mo) (mg/day)

1 42/M 4 Propranolol, 402 52/M 3 Nadolol, 1603 65/M 54 65/M 84 Propranolol, 405 64/M 36 59/M 4 Propranolol, 1607 34/F 4 Propranolol, 808 56/M 15 Atenolol, 1009 49/M 310 38/M 14 Timolol, 2011 36/M A Nadolol, 8012 42/M 5 Timolol, 2013 41/M 414 53/M 315 60/M 6016 39/M 6 Propranolol, 32017 62/M 3 Propranolol, 3018 50/M 6 Propranolol, 8019 37/M 24 Propranolol, 16020 45/M 921 54/M 522 54/M 5 Nadolol, 4023 61/M 1824 35/M 5 Timolol, 2025 54/M 4 Propranolol, 6026 38/M 427 51/M 4 Propranolol, 120

MI myocardial infarction.ARemote myocardial infarction; date is unknown.

the left anterior oblique projection that gave the best separationof the ventricles (35 degrees) with 15 degrees caudal angulationto maximize separation of the left atrium and left ventricle. Adedicated computer system (Technicare VIP 450) was inter-faced to the scintillation camera. Data were collected in theframe mode (32 frames per RR interval) in a 64 x 64 pixelmatrix and were processed off line with a DEC PDP 11/34aminicomputer equipped with a Lexidata display unit as pre-viously described. 16 Ejection fraction was calculated as (EDC- ESC) x 1 00/EDC, where EDC and ESC are the left ventric-ular end-diastolic and end-systolic counts, respectively, correct-ed for background activity. With this method, left ventricularejection fraction correlates well with contrast left ventriculogra-phy and is reproducible.'6The left ventricular end-diastolic volume was calculated by

the standard geometric area-length method17: V = 8 A2/3 1,where V is volume, A is the area, and 1 is the long axis of the leftventricle. Spatial calibration factors for the X and Y axes of thedigital images were generated with a phantom.'8 The area andthe long axis of the left ventricle were determined with the end-diastolic region of interest generated during calculation of ejec-tion fraction. The left ventricular end-systolic volume (ESV)and stroke volume were derived from ejection fraction and leftventricular end-diastolic volume. This scintigraphic method forvolume measurements has been validated in our laboratory andcorrelates well with contrast ventriculography (r - 0.97).

After imaging at rest, each patient performed a graded supine

Vol. 70, No. 4, October 1984 553

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

EHSANI et al.

bicycle exercise test with an electronically braked bicycle er-gometer (Engineering Dynamics Corp.), which maintains a

constant work rate over a wide range of pedaling frequencies.The pedaling rate was between 65 to 70 rpm. The initial workrate selected was 150 kpm/min for most patients and 300 kpml/min for those few whose maximal treadmill exercise aerobiccapacity was relatively high. The work rates were increased by150 kpm/min every 3 min until severe fatigue developed. Theinitial work rate chosen was relatively low, ranging from 25C1c to35% of the measured VO,max, to allow an exercise duration of9 to 12 min for each patient. Images were obtained in the same

modified left anterior oblique projection used at rest in the last 2

min of each stage of the exercise. The heart rate was recordedevery minute and blood pressure was measured with a mercurysphygmomanometer in the last 30 sec of each stage of theexercise test or more frequently if necessary.

Statistical analysis. Student's t test for unpaired observa-tions and least-squares linear regression were used for compari-son of group means of continuous variables and for evaluatingthe relationship between VO,max and various hemodynamicvariables, respectively. Stepwise multiple regression analysiswas performed to assess the independent contribution of theselected physiologic variables in the prediction of VOMmax.Because of the limited sample size, only a small number ofvariables were incorporated. these included peak exercise ejec-

tion fraction, change in ejection fraction from rest to exercise

(z\EF), change in the ratio of systolic blood pressure (SBP) toESV from rest to exercise (ASBP/ESV), and maximal heart rateduring treadmill exercise. Left ventricular end-diastolic vol-ume, ESV, and L\SBP/ESV values were found to conform to lognormal distribution. Therefore they were logarithmically trans-formed in the statistical analysis. The values presented, howev-er, are the original ones. Values are expressed as means ± SD.

ResultsHeart rate and blood pressure responses to exercise. The

highest heart rate recorded during supine bicycle er-

gometer exercise was lower than that obtained at maxi-mal treadmill exercise, averaging 89 + 9C/c of maximalheart rate during treadmill exercise (table 2, figure 1).SBP was, however, significantly higher with bicycleergometer exercise than with treadmill exercise (figure1). The rate pressure product attained with bicycleergometer exercise was higher than that with tread-mill exercise, averaging 1 13 6% of maximal ratepressure product during treadmill exercise (table 2,figure 1).

TABLE 2Selected hemodynamic variables

HRmax RPPiax EF(beats/min) (X10-31 () SBP/ESV

V0,maxPatient T SB T SB (mlJkg/min) R E A R E A

1 163 152 22.82 26.45 27 57 55 2 .43 4.05 1.6'22 108 107 16.20 20.33 12 62 51 - 11 -3 160 140 24.00 27.72 22 16 13 - 3 0.75 1.08 0.334 130 120 17.16 18.00 20 2X 23 - 5 0.79 0.94 0.155 152 150 28.88 24.60 17 38 34 -4 0.99 1.31 0.326 140 125 21.00 21.25 19 43 34 -9 1.21 1.48 0.277 142 134 12.78 17.96 21 64 63 -1 1.88 2.00 0.128 142 130 28.40 27.30 21.5 71 69 32 .47 4.77 1.309 136 133 19.04 26.33 21 65 63 2 1 82 3.19 1.3710 137 138 17.81 24.84 23 60 55 -5 2.07 2.43 0.3611 133 107 16.76 21.40 21 62 59 -3 2.24 2.67 0.4312 125 115 15.75 19.55 24.5 49 47 - 2 1.18 1.57 0.3913 154 120 24.95 24.00 26 57 60 + 3 1.73 2.74 1.0114 190 167 30.40 35.07 29 68 79 + 11 2.60 6.36 3.7615 162 130 25.90 26.00 33 60 70 + 10 2.60 5.26 2.6616 145 115 23.20 24.15 26 50 51 + 1 1.82 2.23 0.4117 168 150 25.20 31.50 29 57 61 +4 2.03 2.96 (0.9318 122 100 15.86 17.00 23 52 54 -'-2 1.77 2.18 0.4119 167 172 30.10 39.56 29 65 74 +-9 2.84 5.61 2.7720 176 158 23.90 30.97 33 55 58 + 3 1.29 2.48 1.1921 174 143 31.32 28.60 31 43 49 +6 1.40 2.99 1.5922 118 97 15.30 15.52 22 64 70 +6 2.83 4.10 1.2723 176 134 28.20 26.80 23 51 57 +6 1.71 3.17 1.4624 158 162 23.70 28.19 27 42 48 +6 1.31 1.79 0.4825 130 116 15.60 19.72 28 62 68 +-6 2.55 4.47 1.9226 176 136 22.88 22.03 28 49 55 +6 1.65 3.06 1.4127 140 95 16.32 16.63 21 43 47 +4 0.94 1.39 (0.45

E exercise: EF = ejection fraction: HRmax =maximal heart rate: R = rest: RPPmax maximal rate pressure product:SB = supine bicycle: T = treadmill.

554 CIRCULATION

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

PATHOPHYSIOLOGY AND NATURAL HISTORY-ExERCISE TESTING

treadmillsupine bicycle

p<O.Ol210

160Fc

-o

I

EE 170

p<O.OOl

K301

m

x

X,

p<o.oo1

201

00; | 3-| |O gziocil ~~~~~~~10111

FIGURE 1. Heart rate (HR) at maximal treadmill exercise was higher than that with peak supine bicycle ergometer exercise.Rate pressure product (RPP x 10-3) at peak supine bicycle ergometer exercise was greater than that with treadmill exercisebecause of the significantly higher SBP attained at peak exercise, suggesting that MVO2 during supine exercise was not lowerthan that during upright exercise. Data are means ± SD.

W)2max. VO2max during treadmill exercise rangedfrom 12 to 33 ml/kg/min, averaging 24 ± 5 ml/kg/minfor the entire group (table 2). Maximal 02 pulse, de-fined as VO2max divided by maximal heart rate, variedbetween 0.11 and 0.20 ml/kg/beat, with an averagevalue of 0.16 ± 0.02 ml/kg/beat. Respiratory ex-change ratio averaged 1.27 ± 0.10, ranging from 1.09to 1.46. Peak bicycle work rate correlated withVO2max (r = .61, p < .001).

Left ventricular contractile function. Left ventricularejection fraction at rest was normal ('50%) in 18 pa-tients and low in nine (table 2). Left ventricular exer-cise reserve (zAEF) was variable; ejection fraction de-creased in 12 patients and increased in the remaining15 patients in response to peak exercise (table 2). Ab-normal left ventricular exercise reserve was arbitrarilydefined as a fall in ejection fraction at peak exercise.Patients who had an abnormal left ventricular exercisereserve exhibited a significantly lower ejection frac-tion at peak exercise than those whose ejection fractionincreased with exercise (47 ± 17% vs 60 ± 10%; p <.05). However, ejection fraction at rest did not differbetween the two groups (51 ± 16% vs 55 + 8%; p =

NS). At rest, ESV and SBP were not significantlydifferent in the two groups (figure 2). At peak exercise,the relation between SBP and ESV was shifted upwardand to the right, with increases in SBP (p < .001) andESV (from 88.5 ± 48 to 96 + 44 ml; p < .025) in thepatients with abnormal left ventricular exercise reserve(figure 2). In contrast, the SBP-ESV relation was shift-ed upward and to the left, with an increase in SBP (p <.001) but a decrease in ESV (from 73 ± 24 to 66 + 26ml; p < .005) in response to peak exercise in thepatients with normal left ventricular exercise reserve(figure 2). I\SBP/ESV was significantly lower in thepatients with abnormal left ventricular exercise reserve

Vol. 70, No. 4, October 1984

than that in patients whose ejection fraction increasedwith exercise (0.61 ± 0.50 vs 1.5 ± 0.90; p < .01).There was a a good correlation between left ventricularexercise reserve (lAEF) and A\SBP/ESV = 0. 13AEF +0.87 (r = .74, p < .0002).

The relation between end-diastolic volume andstroke volume (Frank-Starling mechanism) in responseto peak exercise was also different between the twogroups (figure 3); the patients with normal left ventric-

220r

E

E

L-V)

180F

140p

innlU

047

i Q**

ll,1

40 80ESV (r

Exercise

Rest

* p<O.025** p < 0.005

120 160nl)

FIGURE 2. SBP-ESV relation characterizing differences in left ven-tricular contractile function between patients with abnormal (solid cir-cles and line) and normal (open circles, broken line) left ventricularexercise reserve. In the fornmer, the SBP-ESV relation was shifted up-ward and to the right, with increases in SBP (p < .001) and ESV inresponse to exercise. In contrast, in patients with normal left ventricularexercise reserve, the SBP-ESV relation was shifted upward and to theleft, with an increase in SBP (p < .001) and a decrease in ESV consis-tent with a larger increase in inotropic state during exercise than that inthe other group.

555

- by guest on M

ay 31, 2018http://circ.ahajournals.org/

Dow

nloaded from

EHSANI et al.

A

1401F1001-

601-w

W

-i

0I-

LL)

140 r

0

,,

Peak i

Rest

B

1001-

60 -

20):

A

A PeakA Rest

,, I

90 1

FIGURE 3. Left ventricular ervolume relation (Frank-Starlin1and abnormal (B) left ventricul;ventricular exercise reserve w,

was higher than the resting levdid not change. In patients wh

ejection fraction was reduced with exercise comparedwith 0. 17 + 0.02 ml/kg/beat in those whose ejectionfraction increased (p < .01). Respiratory exchangeratio attained at maximal exercise did not differ be-tween the two groups (1.27 + 0.12 vs 1.27 + 0.09 inpatients with abnormal and normal left ventricular ex-

QxF) 7L ,) ercise reserve, respectively). Endurance time during4w' maximal treadmill exercise (Bruce protocol) was 462

°- --+± 74 sec in the patients whose ejection fraction in-exercise creased with exercise and 351 + 106 sec in those with

a decrease in ejection fraction (p < .01). The highestl I L l heart rate attained during supine exercise was not sig-

nificantly different between the two groups (129 + 14vs 133 ± 25 beats/min in those with abnormal andnormal left ventricular exercise reserve, respectively).However, maximal heart rate during treadmill exercise(VO,max) was lower in the patients with abnormal leftventricular exercise reserve than that in patients with anormal value (139 + 14 vs 157 + 21 beats/min; p <.005). SBP at peak supine exercise was 178 + 21 and191 ± 20 mm Hg in the patients with abnormal andnormal left ventricular exercise reserve, respectively

exercise AA (p - NS). SBP at maximal treadmill exercise did notdiffer between the two groups (144 + 28 vs 149 ± 19

I 1 l mm Hg). There was no significant difference between30 170 210 250 the mean ages of the two groups (51 + 10 vs 49 ± 9

LVEDV (ml) years).Correlation between left ventricular contractile function

id-diastolic volume (LVEDV) and stroke and V02max. At rest, neither ejection fraction nor thegmechanism) in patienlts with normlal (A)grexerchise rrin patienitswithnoseale(A) ratio of SBP/ESV correlated with VO,max. There wasar exercise reserve. In patients whose leftas normal, peak exercise stroke volume a weak but significant correlation between peak exer-cl. Left ventricular end-diastolic volumiie cise ejection fraction and VOmax (r = .43, p < .025;ose left ventricular exercise reserve was figure 4). Peak exercise SBP/ESV did not correlate

abnormal, the average stroke volume at peak exercise did not changefrom the resting level. In this group, stroke volume tended to be lower inpatients with a larger end-diastolic volume than that in patients with asmaller end-diastolic volume.

ular exercise reserve (figure 3, A) showed a modest butsignificant increase in stroke volume in response topeak exercise (85 + 14 vs 95 + 17 ml; p < .001)compared with those with abnormal left ventricularexercise reserve (figure 3, B), whose stroke volumedid not change (80 + 21 vs 79 ± 28 ml). There were nostatistically significant differences in end-diastolic vol-ume between the two groups (figure 3). Body surfacearea was similar in the two groups (1.99 + 0.16 vs1.98 + 0.16; p NS).

Patients with abnormal left ventricilar exercise re-serve had a significantly lower VO,max than patientswhose ejection fraction increased with exercise (21 +4 vs 27 ± 4 ml/kg/min; p < .001). Maximal 0, pulsewas 0.15 + 0.02 ml/kg/beat in the patients whose

556

40rc

c)E

0E

30 F

20 -

101of

0 0

* 0

*0

*

100 ~~~~~0.z0

0

0r - 0.43

pc 0.025

l ,

20PEAK

40 60EXERCISE EF (%)

80

FIGURE 4. Relation between peak exeicise ejection fraction (EF) andVO,max. There was a modest (r = .43) but signiificant (p < .025)correlation between peak exercise ejection fraction and VO,max.

CIRCULATION

1

1

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

PATHOPHYSIOLOGY AND NATURAL HISTORY-EXERCISE TESTING

40 rc

0)

E

0

0>

* _

* 0

*301-

20}

10020

c

-Q

0E

I!

0

.

r 0.77p<0.0002

-10 0 10 20A EF

FIGURE 5. Left ventricular exercise reserve (lAEF) correlated wellwith VO7max (r = .77, p < .0002).

250 r

150 0

*.

50 _

0.- jUs

r : 0.61p<O.O00

-10 02j

10 4'

A EFFIGURE 7. Left ventricular exercise reserve (AEF) correlated modest-ly but significantly with maximal heart rate during treadmill exercise(HRRmax; r = .61, p < .001).

with VO2max (r = .08, p = NS). Left ventricularexercise reserve correlated with VO2max (V09 = 0.67x AEF + 23.5; r = .77, p < .0002; figure 5). Therewas also a significant correlation between ASBP/ESVand VO2max(VO2max = 2.98 X ASBP/ESV + 21.6;r = .67, p < .001; figure 6). Maximal 02 pulse corre-lated with left ventricular exercise reserve (r = .50, p< .005). There was a significant correlation betweenleft ventricular exercise reserve and the highest heartrate attained during maximal treadmill exercise (r =

.61, p < .001; figure 7). Furthermore, there was agood correlation between maximal heart rate duringtreadmill exercise and VOmax (r = .70, p < .0002).Endurance time during maximal treadmill exercise test(Bruce protocol) correlated with left ventricular exer-

45r

25F

51-r,

0

4)

r = 0.67p < 0.0005

1 3

A SBP/ESVFIGURE 6. ASBP/ESV correlated reasonably well with VO7max (r.67, p < .0005).

Vol. 70, No. 4, October 1984

cise reserve (r = .61, p < .001) and ASBP/ESV (r =.59, p < .01). Stepwise multiple regression analysisshowed that among the variables considered, AEF andmaximal treadmill heart rate were the two best varia-bles to predict VO2max in these patients (p = .0025and p = .037, respectively; r2 = .60). Addition ofmore variables did not improve the predictive powerand raised the r2 value only marginally (from .60 to.61).

DiscussionOur findings provide evidence that left ventricu-

lar contractile function can significantly influenceVOmax in patients with coronary artery disease whodo not experience angina. This effect is most likelymediated by the changes in maximal stroke volume,which in turn can affect V02max. Patients whose leftventricular contractile function is abnormal during ex-ercise are likely to have a significantly reduced maxi-mal exercise capacity as evidenced by lower VO2max,maximal 02 pulse, and exercise duration (Bruce proto-col). In the patients who exhibited a fall in ejectionfraction with exercise, VO2max was approximately22% lower than that in patients whose ejection fractionincreased with exercise.The fall in ejection fraction during exercise may be

caused by the following changes during exercise: (1)an inadequate rise in stroke volume with an inappropri-ately large end-diastolic volume, (2) no change instroke volume with a larger end-diastolic volume, or(3) a decrease in stroke volume with a minimal changein end-diastolic volume. Our results suggest that thefall in ejection fraction in our patients was associatedwith a lack of increase in stroke volume from the

557

C

0)'._

E

x0E

S

AL

l l l

l l

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

EHSANI et al.

resting level and a small but statistically insignificantincrease in end-diastolic volume.

The lower V02max observed in the patients whoseejection fraction fell during exercise may be, in part,mediated by a lower arteriovenous 02 difference.However, since none of the patients had congestiveheart failure or engaged in regular exercise training, itis unlikely that a difference in 02 extracting capacity ofskeletal muscles was primarily responsible for the dif-ferences observed in the V02max values. It is likelythat central circulatory mechanisms directly affectingcardiac output played a major role in loweringV02max in patients with abnormal left ventricular ex-ercise reserve as suggested by the differences in therelationship, in response to exercise, between SBP andESV, and between end-diastolic volume and strokevolume in the two groups. In comparison with theresting state, it appears that left ventricular contractilestate was increased at peak exercise in both groups asreflected by a large increase in SBP with only a smallchange in ESV. However, the SBP-ESV relationshowed clear differences in response to peak exercisebetween the two groups. Patients with abnormal leftventricular exercise reserve exhibited a large increasein SBP with a modest increase in ESV, shifting thepressure-volume relation upward and to the right. Thisgroup showed no increase in stroke volume at peakexercise. In contrast, the group whose left ventricularexercise reserve was normal demonstrated a compara-ble increase in SBP but a slight decrease in ESV shift-ing the pressure-volume relation upward and to theleft, suggestive of a better global left ventricular con-tractile function than the other group.'9 Furthermore,these patients showed an increase in stroke volume atpeak exercise even though the changes in preload weresimilar to those in patients with abnormal left ventricu-lar exercise reserve.

Ejection fraction is influenced by contractile state,afterload,20 and to some extent by large variations inpreload.2' It is unlikely, however, that the loading con-ditions could have accounted for differences in theejection fraction response to peak exercise becauseSBP and end-diastolic volume, two major determi-nants of left ventricular wall tension, were not signifi-cantly different between the two groups. Therefore anabrupt decrease in ejection fraction is consistent withsome degree of impairment in left ventricular function.Differences in the response of the SBP-ESV relationsto peak exercise in the two groups support this conclu-sion, since the end-systolic pressure-volume relation isindependent of the loading conditions.'9 22

Ejection fraction response to exercise may deterio-

rate with advancing age, presumably due to the agingprocess and independent of disease.23 However, this isnot a likely explanation for our results because themean age of the patients whose ejection fraction de-creased with exercise was similar to that of patientswho exhibited an increase in ejection fraction withexercise and because no relationship could be estab-lished between age and left ventricular exercise reservein our patients. /3-Adrenergic blockade may concealdeterioration of left ventricular systolic function in pa-tients with coronary artery disease.24 25 However, thisis also an unlikely possibility because the proportion ofthe patients who were taking /3-adrenergic-blockingdrugs was not significantly different between the twogroups (75% vs 53%; p = NS). Moreover, the exer-cise protocols which might influence ejection fractionresponse to exercise, were similar in the two groups.26

Although it may be argued that two different modesof exercise were used for measurement of V02max andejection fraction, MVO2 during supine bicycle exer-cise was unlikely to be lower than that with maximaltreadmill exercise because the rate pressure productwas higher with bicycle ergometer exercise. In addi-tion, left ventricular end-diastolic volume, anothermajor determinant of MVO2, is expected to be largerwith supine than with upright exercise. True V0,maxcan generally be measured during a treadmill exercisetest,' which is not a suitable method for assessment ofleft ventricular function during strenuous exercise.Even maximal attainable V02 during upright bicycleexercise is generally lower than that during treadmillexercise in healthy subjects as well as in patients withcoronary artery disease. 21 Thus, although exerciseintensity in terms of total body VO2 was lower duringthe radionuclide studies than during the treadmill test,myocardial energy expenditure and MVO0 were prob-ably not different between the two modes of exercise.Previous investigators did not find any relationship

between maximal aerobic exercise capacity and leftventricular systolic function at rest or at peak exer-

cise.` Although our results confirm some of the pre-

viously reported results, namely that left ventricularcontractile function at rest has no appreciable effect onmaximal aerobic exercise capacity,28 we found a strongcorrelation between left ventricular exercise reserve

and VO2max and a significant, although weak, corre-

lation between peak exercise ejection fraction andV02max. The apparent disparity between our findingsand those of others is probably because of the follow-ing: In this study we included only those patients whoattained true V02max as defined by objective criteria.In some of the previous reports, exercise duration was

558 CIRCULATION

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

PATHOPHYSIOLOGY AND NATURAL HISTORY-EXERCISE TESTING

used as an index of maximal exercise capacity.6-5 Ourfindings indicate that although exercise duration corre-lated with changes in ejection fraction, the correlationwas far better between V02max and left ventricularexercise reserve. In addition, some investigators haveincluded patients with angina who are unable to reachtheir true V02max' or have studied patients with severeleft ventricular dysfunction at rest.it 28, 29 Others haveused bicycle ergometer exercise tests for measurementof V02max,29 which is not likely to produce trueV02max'3' 27 because of the smaller muscle mass usedduring bicycle exercise compared with treadmill exer-cise.Our findings suggest that maximal chronotropic re-

sponse and left ventricular exercise reserve are amongthe variables that influence true V02max in patientswith ischemic heart disease. The effect of left ventricu-lar exercise reserve on V02max would have to be indi-rect and dependent on its influence on stroke volumeand heart rate at maximal exercise. The reason for abetter correlation between V02max and lAEF than thatbetween V02max and peak exercise ejection fraction isnot clear. It is possible, however, that since V02max isalso influenced by end-diastolic volume, heart rate,and arteriovenous 02 difference, compensatory mecha-nisms altering one or more of these variables couldhave blunted the effect of peak exercise ejection frac-tion on true V02max causing a weak association. Onthe other hand, it appears likely that left ventricularexercise reserve is less affected by these compensatorymechanisms. For example, when there is a decrease inejection fraction with exercise caused by myocardialischemia, compensatory mechanisms are probably nolonger sufficient to maintain cardiac output and there-by 02 uptake capacity at the expected normal level.Thus acute left ventricular dysfunction along with apoor chronotropic response may have accounted, atleast in part, for the lower V02max observed in thesepatients.

Although our observations may not be extrapolatedto the symptomatic population with coronary arterydisease, particularly patients with effort angina, theyindicate that there is a reasonably strong correlationbetween maximal aerobic exercise capacity and leftventricular exercise reserve in those patients who canattain true V02max and that impaired left ventricularsystolic function during exercise is likely to result in amarkedly diminished maximal aerobic exercise capac-ity in patients who are not limited by angina.

References1. Astrand PO, Rodahl K: Textbook of work physiology. New York,

1977, McGraw-Hill Book Co., Inc. pp 365-383

Vol. 70, No. 4, October 1984

2. Detry JM, Vandenbroucke G, Kussumi F, Brasseur LA, BruceRA: Increased arteriovenous oxygen difference after physical train-ing in coronary heart disease. Circulation 44: 109, 1971

3. Clausen JP: Circulatory adjustment to dynamic exercise and effectof physical training in normal subjects and patients with coronaryartery disease. Prog Cardiovasc Dis 18: 495, 1976

4. Borer JS, Bacharach SL, Green MV, Kent KM, Epstein SE, John-ston GS: Real-time radionuclide cineangiography in the noninva-sive evaluation of global and regional left ventricular function atrest and during exercise in patients with coronary artery disease. NEngl J Med 296: 839, 1977

5. Berger HJ, Reduto LA, Johnston DE, Borkowski H, Sands JM,Cohen LS, Langon RA, Gottschalk, Zaret BL: Global and regionalleft ventricular response to bicycle exercise in coronary arterydisease: assessment by quantitative radionuclide angiocardiogra-phy. Am J Med 66: 13, 1979

6. Kramer N, Susmano A, Shekelle RB: The false negative treadmillexercise and left ventricular dysfunction. Circulation 57: 763,1978

7. Benge W, Litchfield RL, Marcus ML: Exercise capacity in patientswith severe left ventricular dysfunction. Circulation 61: 955, 1980

8. Port S, McEwan P, Cobb FR, Jones RH: Influence of resting leftventricular function on left ventricular function during exercise inpatients with coronary artery disease. Circulation 63: 856, 1981

9. Maskin CS, Forman R, Sonnenblick EH, Frishman WH, LeJemtelTH: Failure of dobutamine to increase exercise capacity despitehemodynamic improvement in severe chronic heart failure. Am JCardiol 51: 177, 1983

10. Mitchell JH, Blomqvist G: Maximal oxygen uptake. N Engl J Med284: 1018, 1971

11. Vatner SF, Franklin D, Higgins CB, Patrick T, Braunwald E: Leftventricular response to severe exertion in untethered dogs. J ClinInvest 51: 3052, 1972

12. Issekutz B, Birkhead NC, Rodal K: Use of respiratory quotients inassessment of aerobic work capacity. J Appl Physiol 17: 47, 1962

13. Stenberg J, Astrand PO, Ekblom B, Royce J, Saltin B: Hemody-namic response to work with different muscle groups, sitting andsupine. J Appl Physiol 22: 61, 1967

14. Bruce RA: Exercise testing of patients with coronary heart disease:principles and normal standards of evaluation. Ann Clin Res 3:323, 1971

15. Ehsani AA, Heath GW, Hagberg JM, Sobel BE, Holloszy JO:Effects of 12 months of intense exercise training on ischemic STsegment depression in patients with coronary heart disease. Circu-lation 64: 1116, 1981

16. Biello DR, Sampathkumaran KS, Geltman EM, Briston WA, ScottDJ, Grbac RT: Determination of left ventricular ejection fraction: anew method that requires minimal operator training. J Nucl MedTechnol 9: 77, 1981

17. Dodge HT, Sandler H, Ballew DW, Lord JD Jr: The use of biplaneangiocardiography for the measurement of left ventricular volumein man. Am Heart J 60: 762, 1960

18. Esser PD, Seldin DW, Nichols AB, Anderson PO: Spatial calibra-tion of digital scintigraphic images. Radiology 144: 901, 1982

19. Grossman W, Braunwald E, Mann T, McLourin LP, Green LH:Contractile state of the left ventricle in man as evaluated from endsystolic pressure volume relations. Circulation 56: 845, 1977

20. Ross J Jr: Afterload mismatch and preload reserve: a conceptualframework for the analysis of ventricular function. Prog Cardio-vasc Dis 18: 255, 1976

21. Nixon JR, Murray RG, Leonard PD, Mitchell JH, Blomqvist GC:Effect of large variations in preload on left ventricular performancecharacteristics in normal subjects. Circulation 65: 698, 1982

22. Mitchell JH, Wildenthal K, Mullins CB: Geometrical studies of theleft ventricle using biplane cineangiography. Fed Proc 28: 1334,1969

23. Port S, Cobb FR, Coleman RE, Jones RH: Effect of age on theresponse of the left ventricular ejection fraction to exercise. N EnglJ Med 303: 1133, 1980

24. Battler A, Ross J Jr, Slutsky R, Pfisterer M, Ashburn X, Froe-licher V: Improvement in exercise-induced left ventricular dys-function with oral propranolol in patients with coronary heart dis-ease. Am J Cardiol 44: 318, 1979

25. Marshal RC, Weinberg G, Schelbert MR, Henze E: Effect of oralpropranolol on rest, exercise and postexercise left ventricular per-

559

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

EHSANI et al.

formance in normal subjects and patients with coronary arterydisease. Circulation 63: 572, 1981

26. Foster C, Anholm JD, Hellman CK, Carpenter J, Pollack ML,Schmidt DH: Left ventricular ejection fraction during sudden stren-uous exercise. Circulation 63: 592, 1981

27. Ehsani AA, Martin WH IlI, Heath GW. Coyle EF: Cardiac effectsof prolonged and intense exercise training in patients with coronary

artery disease. Am J Cardiol 50: 246, 198228. Franciosa JA, Park M, Levin B: Lack of correlation between exer-

cise capacity and indices of resting left ventricular performance inheart failure. Am J Cardiol 47: 33, 1981

29. Higginbotham MB, Morris KG. Conn EH, Coleman RE, Cobb FR:Determinants of variable exercise performance among patientswith severe left ventricular dysfunction. Am J Cardiol 51: 52, 1983

CIRCULATION560

by guest on May 31, 2018

http://circ.ahajournals.org/D

ownloaded from

A A Ehsani, D Biello, D R Seals, M B Austin and J Schultzasymptomatic patients with coronary artery disease.

The effect of left ventricular systolic function on maximal aerobic exercise capacity in

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1984 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.70.4.552

1984;70:552-560Circulation. 

http://circ.ahajournals.org/content/70/4/552the 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 31, 2018

http://circ.ahajournals.org/D

ownloaded from