Preoperative Exercise Capacity in Symptomatic Patientswith Aortic Regurgitation as a Predictor of Postoperative

Left Ventricular Function and Long-term PrognosisROBERT 0. BONOW, M.D., JEFFREY S. BORER, M.D., DOUGLAS R. ROSING, M.D.,

WALTER L. HENRY, M.D., ALAN S. PEARLMAN, M.D., CHARLES L. MCINTOSH, M.D.,ANDREW G. MORROW, M.D., AND STEPHEN E. EPSTEIN, M.D.

SUMMARY Forty-five symptomatic patients with aortic regurgitation underwent graded treadmill exercisetesting before operation. Twenty-seven patients (group A) could not complete stage I of the National Institutesof Health exercise protocol because of limiting symptoms (exercise duration < 22.5 minutes); 18 patients(group B) completed this stage without limiting symptoms (exercise duration > 22.5 minutes). Patients ingroup A had higher resting pulmonary capillary wedge pressures (mean 19 vs 13 mm Hg, p < 0.05) and leftventricular (LV) end-diastolic pressures (mean 24 vs 16 mm Hg, p < 0.05) than those in group B, but did notdiffer with respect to LV systolic dimension or fractional shortening by echocardiography or LV ejection frac-tion at rest or during exercise by radionuclide cineangiography.Among 32 patients with subnormal preoperative LV fractional shortening on echo, nine of 17 in group A and

0 of 15 in group B have died (p < 0.01); seven of the nine deaths were from late congestive heart failure. GroupA patients also had less decrease postoperatively in LV diastolic size by echocardiography (mean decrease 8 vs23 mm, p < 0.001) and less increase postoperatively in LV ejection fraction during exercise by radionuclidecineangiography (mean increase 11% vs 23%,p < 0.05) than group B patients. No group A patient and 60% ofgroup B patients had normal exercise ejection fractions postoperatively (p < 0.01). The differences inpostoperative mortality and function were not predicted by the differences in preoperative hemodynamicsbetween the two groups. Thus, exercise capacity is imprecise in assessing preoperative LV function in symp-tomatic patients with aortic regurgitation, but is useful in predicting long-term survival after operation andreversibility of LV dilatation and systolic dysfunction.

THE MAJORITY of symptomatic patients who un-dergo surgery for aortic regurgitation have symp-tomatic improvement after valve replacement. Others,however, develop progressive congestive heart failure,the most common cause of late postoperative death.In such patients, irreversible myocardial changespresumably occur before or concomitant with thedevelopment of symptoms that lead to operation.1'8To define more precisely a strategy for determiningthe optimal timing of operation in patients with aorticregurgitation, we used echocardiography to assess leftventricular systolic function in symptomatic patientsbefore operation. This provided a sensitive means ofpredicting which patients would die after operationfrom irreversible left ventricular failure.8 Despite theenhanced ability to predict long-term prognosis afteroperation, however, such analyses still identifiedwithin the "high-risk" group a substantial number ofpatients with impaired preoperative systolic functionwho did well after operation. Thus, while sensitivity inpredicting late congestive failure deaths was high,specificity was low.

Exercise testing until exhaustion or the development

From the Cardiology and Surgery Branches, National Heart,Lung, and Blood Institute, National Institutes of Health, Bethesda,Maryland.Address for correspondence: Robert 0. Bonow, M.D., Building

10, Room 7B-15, National Heart, Lung, and Blood Institute,National Institutes of Health, Bethesda, Maryland 20205.

Received November 5, 1979; revision accepted June 25, 1980.Circulation 62, No. 6, 1980.

of symptoms is often used as an index of cardiac func-tion and, as such, in the management of patients witha variety of cardiac disorders. One assumption im-plicit in using the results of exercise testing in this wayis that the response of the heart to stress providesmore information concerning left ventricular func-tional reserve than can be obtained from hemo-dynamic and echocardiographic studies obtained withthe patient at rest.9-12To ascertain whether graded exercise testing would

further aid in the clinical decision analysis of patientswith aortic regurgitation, treadmill exercise tests wereperformed in symptomatic subjects before operation.These subjects were then followed after operation todetermine whether preoperative exercise performancecould predict long-term postoperative survival. Thedata were also analyzed to determine whetherpreoperative exercise performance provided insightinto preoperative left ventricular function (determinedby echocardiography and radionuclide cineangiog-raphy) as well as into reversibility of left ventriculardysfunction.

MethodsPatient Selection

All patients who underwent aortic valve replace-ment for chronic aortic regurgitation at the NationalInstitutes of Health (NIH) between January 1972 andDecember 1978 were evaluated consecutively in aprospective manner. The indication for operation ineach patient was a history of moderate-to-severe

1280

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

EXERCISE CAPACITY IN AR/Bonow et al.

dyspnea on exertion, orthopnea, or paroxysmal noc-turnal dyspnea, an episode of pulmonary edema, or ahistory of angina pectoris or syncope. All patients hadsevere aortic regurgitation as assessed by aortic rootcineangiography (early opacification of the cardiacapex that did not clear during the subsequent cardiaccycle). Patients were excluded from evaluation if theyhad evidence of acute aortic regurgitation, a systolicaortic value gradient greater than 20 mm Hg, or eithermitral valvular disease or aortic root disease that re-quired operative repair at the time of aortic valvereplacement.

Seventeen of the 71 patients (24%) who fulfilled theselection criteria were found to have coexistent cor-onary artery disease, defined as greater than 50%stenosis of the diameter of at least one major coronaryartery. These patients were excluded from graded ex-ercise testing because exercise capacity might belimited by myocardial ischemia caused by inadequatecoronary reserve, making it impossible to determinemaximal functional capacity of the myocardiumper se.

Six other patients were not exercised; two hadhistories of syncope, three were considered clinicallyunstable, and another had severe ankylosing spon-dylitis. Exercise data were unavailable on anotherthree patients. Thus, 45 patients (ages 20-68 years,mean 44 years) made up the study population. Therewere 38 men and seven women. Thirty-four of the 45patients underwent coronary arteriography and eitherhad normal studies or less than 50% stenoses. Theother 11 patients were all younger than 35 years of age(nine were less than 30 years of age) and none com-plained of chest pain. We refrain from performingcoronary arteriography in such patients with aorticvalve disease. Three of these 1 1 patients died fromcongestive heart failure after operation and had nor-mal coronary arteries at necropsy. The other eightpatients are in New York Heart Association func-tional class I after operation.None of the 45 patients had other medical con-

ditions, such as anemia or chronic pulmonary disease,that might limit exercise capacity. Thirty-one of the 45patients were previously reported in the results of aor-tic valve replacement.8

All patients had unequivocal clinical evidence ofsevere aortic regurgitation with blowing diastolic mur-murs and cardiomegaly on chest x-ray. The averagesystemic pulse pressure was 95 mm Hg and was lessthan 60 mm Hg in only one patient. Systemic diastolicpressure averaged 51 mm Hg and exceeded 70 mm Hgin only two patients. Modified Romhilt-Estes pointscores, computed as previously described,5 averaged6.7 and were less than 5.0 in only seven patients.

Preoperative Studies

All patients were evaluated preoperatively byhistory, physical examination, 12-lead electrocar-diography, M-mode echocardiography, left- andright-heart catheterization and graded treadmill exer-cise testing. In addition, 24 patients underwent

radionuclide cineangiography at rest and during exer-cise. This includes all patients who underwent aorticvalve replacement for aortic regurgitation at our in-stitution between April 1976 and December 1978. Allstudies were performed during a 5-day inpatientevaluation, and the maximum time between any twostudies was 4 days, except for 1 1 patients who un-derwent catheterization at other institutions beforetheir referral to the NIH. In these patients, the max-imum time between catheterization and the nonin-vasive studies was 6 weeks.

Left Ventricular Contrast Cineangiography

Single-plane left ventricular cineangiograms wereobtained at catheterization in all patients. End-diastolic and end-systolic volumes were calculated aspreviously described.13 In 18 patients, however, poordye opacification, premature ventricular complexes ortechnical factors prevented optimal determination ofventricular volumes and ejection fraction. Hence, wecould not compare echocardiographic vs cineangio-graphic measurements in all patients. However, areasonable correlation was found between cineangio-graphic left ventricular volumes and echocardio-graphic left ventricular internal dimensions in the 27patients with adequate cineangiographic studies, fordiastole (r = 0.75), systole (r = 0.71), or diastole andsystole combined (r = 0.89). A similar correlation wasfound between angiographic left ventricular ejectionfraction and echocardiographic fractional shortening(r = 0.74).Echocardiography

Echocardiograms were obtained using a 12.5-mmdiameter, 2.25-MHz unfocused ultrasound transducerand either an Ekoline 20A or a Hoffrel 201 ultrasoundtransceiver interfaced with a Honeywell 1856 strip-chart recorder. Echocardiographic measurement ofthe left ventricular dimensions at end-diastole andend-systole and the thickness of the ventricular sep-tum and the left ventricular free wall were obtainedwith the ultrasound beam passing through the left ven-tricle caudal to the tips of the mitral leaflets.8'14, 15Left ventricular fractional shortening was calculatedas the ratio of the difference between left ventriculardiastolic dimension and systolic dimension to the leftventricular diastolic dimension."' In addition, left ven-tricular ejection fraction17 and left ventricular mass18were calculated.

Graded Treadmill Exercise TestingExercise testing was performed in the postabsorp-

tive state. Digitalis and diuretic therapy were discon-tinued at least 3 days before testing. Heart rate andrhythm were monitored continually during testingwith a two-lead ECG with a CM, electrode and amodified posteroinferior electrode. Patients exercisedon a motor-driven treadmill using the NIH exerciseprotocol (table 1). In stage I of this protocol, thetreadmill is driven at a constant speed of 2.2 mph at an

1281

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 6, DECEMBER 1980

TABLE 1. National Institutes of Health Exercise Protocol

Stage I Stage IITime Speed Speed(min) (mph) Grade (mph) Grade

0 2.2 0% 1.9 10%2.5 2.2 2.5% 2.3 12%5.0 2.2 5.0% 2.7 14%7.5 2.2 7.5% 3.1 16%

10.0 2.2 10.0% 3.5 18%12.5 2.2 12.5% 3.9 20%15.0 2.2 15.0% 4.7 20%17.5 2.2 17.5% 5.5 20%20.0 2.2 20.0% 6.3 20%22.5 End of End of 7.1 20%

stage I stage I

initial inclination of 0%. Every 2.5 minutes the inclina-tion is increased 2.5% until a maximum of 22.5minutes elapse. Exercise is continued to markedfatigue or until the patient complains of angina, limit-ing dyspnea or lightheadedness. Twenty-seven patientsdeveloped limiting symptoms during stage I. Patientswho completed the entire 22.5 minutes of stage I

without limiting symptoms were then tested within4-48 hours on stage II of the exercise protocol (table1). In stage II, the treadmill is initially driven at 1.9mph at an inclination of 10%. Every 2.5 minutes thetreadmill speed and inclination are increased by 0.4mph and 2%, respectively. After 15 minutes, the in-clination is held constant at 20% and the speed is in-creased 0.8 mph every 2.5 minutes. The indications forcessation of testing are the same as for stage I of theexercise protocol. Each patient who completed stage I

without symptoms completed at least 8 minutes ofstage II before developing limiting fatigue, dyspnea orangina. Five patients had been followed in our clinicfor several years before the onset of symptoms, andprevious exercise testing had documented good exer-cise tolerance. These five patients were tested directlyon stage II of the exercise protocol without prior test-ing on stage I; all five completed more than 8 minutes(9.5-17.1 minutes) of exercise in stage II before limit-ing symptoms, and were assumed capable of complet-ing stage I without symptoms.

Forty-four patients discontinued exercise testingbecause of fatigue or dyspnea. Only one patientdeveloped angina during exercise; this patient had nor-mal coronary arteriograms. Isolated ventricular ec-topic beats were often observed during exercise, but nopatients showed high-grade ventricular arrhythmiasthat required discontinuation of testing.

In 35 patients, oxygen consumption was measuredat the end of symptom-limited exercise'9 using a con-tinuous flow system and paramagnetic oxygen

analyzer.20 Twelve of these patients completed stage I

without symptoms and developed limiting symptomsonly during stage II; each obtained higher levels ofoxygen consumption during stage II than duringstage I.

Blood Pool Cardiac Scintigraphy

Technetium-99m radionuclide cineangiography wasperformed in the supine position at rest and duringbicycle exercise using a conventional Anger cameraequipped with a high-sensitivity, parallel-hole col-limator oriented in a modified left anterior obliqueposition. Visually interpretable and statisticallyreliable movies of the blood pool were obtained usingthe ECG-gated equilibrium procedure previouslydescribed."', 12, 21-23 Ejection fractions were determinedat rest and during exercise by analysis of count-basedleft ventricular time-activity curves after backgroundcorrection.21-23 During exercise studies, performedwithin 30 minutes of the rest studies, supine bicycle ex-ercise was begun at a work load of 25 W; the load wasincreased at 25-W increments every 2 minutes untilthe development of angina or limiting dyspnea orfatigue. Heart rate and blood pressure (by sphyg-momanometry) were monitored during exercise.Imaging was begun shortly after the onset of exercise,but only the portion of the image series that occurredduring maximal exercise, encompassing the final 2minutes of exercise, was selected for analysis.

Aortic Valve Replacement

At operation, 18 patients received 2320 series Starr-Edwards prostheses, 17 received glutaraldehyde-fixedporcine heterograft prostheses, eight received 2400series Starr-Edwards prostheses, and two receivedBjork-Shiley prostheses. In addition to total bodyhypothermia to 30-31 'C in all patients, myocardialpreservation techniques included direct coronaryartery perfusion with blood cooled to 30°C in 11patients, topical iced saline lavage (4°C) in threepatients, and coronary perfusion plus iced salinelavage in 31 patients.

Postoperative Studies

All patients who survived 6 months after operationunderwent repeat echocardiography and cardiaccatheterization, performed using either the transeptalor left ventricular puncture technique. Echocar-diographic left ventricular systolic dimension and per-cent fractional shortening were not analyzed becauseof abnormal septal motion in most patients afteroperation.8' 24 Hemodynamic data demonstrated peaksystolic gradients across the prosthetic valve of 0-40mm Hg (mean 9 mm Hg). Prosthetic valve areas werenot computed because of the small systolic gradients.Only two patients had prosthetic valve gradientsgreater than 20 mm Hg; both are doing well 51 and 67months after operation. In both, 6-month post-operative echocardiograms demonstrated substantialreduction in left ventricular diastolic dimension com-pared with the preoperative value (54 vs 74 mm and 52vs 72 mm) despite the prosthetic valve gradient.Prosthetic valve regurgitation was demonstrated intwo patients; each had only mild regurgitation on aor-tic root angiography (faint opacification of the leftventricular outflow tract clearing with the next cardiaccycle) and each had substantial reduction in left ven-

CIRCULATION1282

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

EXERCISE CAPACITY IN AR/Bonow et al.

tricular diastolic dimension by echocardiographycompared with the preoperative value (55 vs 79 mmand 45 vs 64 mm). Both patients are asymptomatic 36and 82 months after operation.

Patients in whom radionuclide cineangiography wasperformed before operation underwent repeat evalua-tion 6 months postoperatively. Supine exercise wasperformed according to a protocol identical with thatused before operation for each patient. To achievesimilar heart rates, however, we often had to continueexercise to higher load levels than could be obtainedbefore operation. Therefore, results used for assess-ment of postoperative systolic function were thoseachieved during maximal exercise. The heart rate atmaximal exercise after operation was not significantlydifferent compared with the preoperative value.

Statistical MethodsLinear regression analysis was performed to relate

treadmill exercise performance to preoperativehemodynamic, echocardiographic, and radionuclidecineangiographic data. Patients unable to completestage I of treadmill exercise because of limiting symp-toms were compared with patients who completedstage I without symptoms by t test. Postoperativechanges in left ventricular diastolic size and systolicfunction were analyzed by t test (using paired and un-paired data), Fisher's exact test and the method ofcovariance (to assess relative influences of pre-operative hemodynamic values and preoperative exer-cise capacity), as appropriate. Survival curves wereplotted by the Kaplan-Meier life-table method25 andstatistical analysis of the survival data by the methodof Mantel and Haenszel.26

Results

Patient Experience

Three of the 45 patients died at operation. Sevendied late (5-42 months) after operation from con-gestive heart failure; each had chronic progressivesymptoms of pulmonary venous hypertension and lowcardiac output before death. Two of these sevenpatients died before-the 6-month postoperative evalua-tion. One other late death was not associated with con-gestive failure; 46 months after operation this patientdied while swimming, and we do not know whetherthis represents drowning or sudden cardiac death.Thirty-four patients are alive at a mean follow-up of38 months (range 6-89 months). Postoperative sur-vival did not correlate with age, type of prostheticvalve, method of myocardial preservation at opera-tion, or systolic gradient across the prosthetic valve.

Preoperative Data

Maximum heart rates during treadmill exercisetesting (fig. 1) ranged from 90-205 beats/min (mean157 beats/min). In only one patient did the heart ratenot reach 100 beats/min. This patient could exerciseonly 45 seconds in stage I before the onset of severedyspnea, which occurred at a heart rate of 90 beats/

min and oxygen consumption of 9.7 ml/min/kg. Thispatient had echocardiographic evidence of a dilatedleft ventricle with poor systolic function and atcatheterization had a resting mean pulmonary arterywedge pressure of 26 mm Hg. She died from con-gestive heart failure 5 months after technicallysuccessful valve replacement. Four other patients hadmaximum heart rates less than 120 beats/min (fig. 1).None of these patients could complete stage I of thetreadmill exercise protocol because of symptoms, andtheir maximum oxygen consumptions ranged from12.5-24.5 ml/min/kg. The remaining 40 patients allreached exercise heart rates of 130 beats/min or more.The 27 patients who could not complete stage I

of treadmill exercise because of limiting symptomsranged in age from 20-65 years (table 2) and did notdiffer from the 18 patients who completed stage Iwithout symptoms, who ranged in age from 24-61years. The male to female ratio was also not signifi-cantly different between the two groups. Patients un-able to complete stage I, in addition to demonstratinggreater impairment of exercise capacity as measuredby exercise duration, also achieved significantly lowerlevels of heart rate and oxygen consumption duringtreadmill exercise than the patients who completedstage I (table 2).

Exercise Capacity vs Resting Hemodynamic Data

The resting cardiac index of patients who could notcomplete stage I exercise because of symptoms was

200 h

175 -

zc

.E

-ccr.

:,

150

125

100

75

0

00

0

00

0 600 A* 000

0 *0

0 00 0

STAGE I

25 _

ol

0

00 0

0

000

0 00 c

0

0

0 AliveDeaths0 Operativea Late (sudden)* Late (CHF)

STAGE II

10 15 20 5' 10' 15'

PREOPERATIVE EXERCISE DURATION (minutes)

FIGURE 1. Preoperative exercise performance. Maximumheart rate during treadmill exercise is plotted as a functionof exercise duration at the time of limiting symptoms. Thedashed line at 22.5 minutes indicates the end ofstage I oftheNational Institutes of Health exercise protocol. The twopatients shown at 22.5 minutes developed limiting symptomstoward the end of stage I and were not tested further.CHF = congestive heart failure.

nsU

1283

50 _ by guest on A

pril 18, 2017http://circ.ahajournals.org/

Dow

nloaded from

VOL 62, No 6, DECEMBER 1980

TABLE 2. Exercise Heart Rate and Oxygen Consumption During Treadmill Exercise

Exercise duration < 22.5 min Exercise duration > 22.5 min(Unable to complete stage I) (Completed stage I) p

Age (years) 47 15 (n = 27) 42 t 14 (n = 18) NSMaximum heart rate

(beats/min) 145 25 (n = 27) 167 - 32 (n = 18) < 0.001Maximum oxygen con-sumption (ml/min/kg) 21 - 6 (n = 23) 31 7 (n = 12) < 0.001Data are mean 1 SD.

0

<22.1<22.5

50-p <0.05 p <0.05

E 403 OWE 0

ccLU0

30 0 010H 0

00Z O0~~~~0

>22.5 522.5 >22.5

o AliveDeaths

0 Operativea Late (Sudden)* Late ICHF)

FIGURE 2. Preoperative hemodynamicdata. Resting cardiac index, meanpulmonary artery wedge pressure and leftventricular end-diastolic pressure in the 27patients unable to complete stage I oftread-mill exercise because of limiting symptoms(< 22.5 minutes) are compared to the 18patients who completed stage I withoutsymptoms (> 22.5 minutes). Large circleswith horizontal bars indicate mean values.One patient who died after operation fromcongestive heart failure (CHF) did not havepreoperative determination ofcardiac index.

2.5 ± 0.6 1/min/m2 (mean ± SD) and did not differfrom that of patients who completed stage I, 2.6 ± 0.41/min/m2 (fig. 2). Patients unable to complete stage I,

however, did have significantly higher resting meanpulmonary artery wedge pressures (19 ± 9 vs 13 ± 6mm Hg, p < 0.05) and left ventricular end-diastolicpressures (24 ± 11 vs 16 ± 9 mm Hg, p < 0.05) thanthe group with greater exercise capacity (fig. 2). Pul-monary artery systolic and diastolic pressures werealso significantly higher (both p < 0.05) in the groupunable to complete stage I exercise, but systemicarterial systolic and diastolic pressures and pulsepressure were not significantly different between thetwo groups.

Exercise Capacity vs Left Ventricular AngiographicVolumes

Among the 27 patients with angiographic dataadequate for volume analysis, the 16 patients unableto complete stage I exercise because of symptoms hadsmaller left ventricular end-diastolic volumes (fig. 3)than the group with greater exercise capacity(282 ± 79 vs 350 ± 61 ml, p < 0.05). The two groupsdid not differ with respect to left ventricular end-systolic volume or left ventricular ejection fraction.

Exercise Capacity vs Echocardiographic andECG Data

The angiographic volume data were paralleled bythe echocardiographic measurements. The 27 patientsunable to complete stage I exercise because of symp-toms had significantly smaller echocardiographic leftventricular end-diastolic dimensions (fig. 4) than the

18 patients with greater exercise capacity (73 ± 8 vs77 ± 7 mm, p < 0.05). Calculated left ventricularmass, however, was not significantly different betweenthe two groups. The two groups also did not differ withrespect to left ventricular end-systolic dimension or

left ventricular fractional shortening (fig. 4). Althoughthe four patients with the lowest fractional shortening

Diastole

500 r

W

Uo

CL >< cc

o D_ Uz CC

zW

U-WJ

80

9

0

p <0.05

<22.5 >22.5

400 F

300

200

100 _

Systole

0

0

(§(30

N.S.

< 22.5 > 22.5

PREOPERATIVE EXERCISE DURATION (minutes)FIGURE 3. Preoperative left ventricular contrastangiographic data. Left ventricular volumes at end-diastoleand end-systole in patients unable to complete stage I(< 22.5 minutes) are compared with patients who completedstage I (> 22.5 minutes).

5O r

E

c

-1

x

0z

Cr

C.)

- 40X

a: Ew Ei_C W 30< cc

O>. 20CCU)o CL 20

JW

B 10

o<22.5 >22.5

PREOPERATIVE EXERCISE DURATION (minutes)

1284 CIRCULATION

_°_

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

EXERCISE CAPACITY IN AR/Bonow et al.

90

80

Ez0U)zM

zH

zccW

z

cr

Js

70

60

LVD (DIASTOLE)0

_;0

0

- 80p 0

8o 8

CM-M c

50 _

40

30F

20

10

0

p <0.05

:22.5 >22.5 <22.5 >22.5

%FS

(DQ

zz

0XU)

z0

:XCR

a:C-)

uw

<22.5 >22.5

PREOPERATIVE EXERCISE DURATION Iminutes)

FIGURE 4. Preoperative echocardiographic data. Left ven-

tricular dimension (LVD) at end-diastole and end-systoleand left ventricular fractional shortening (%FS) in patientsunable to complete stage I (< 22.5 minutes) are comparedwith patients who completed stage I (> 22.5 minutes). Thestippled area identifies the normal range of %FS for our

laboratory (29-45%). Symbols and abbreviations are ex-

plained in figures 1 and 2.

(below 15%) could not complete stage I without limit-ing symptoms, 10 of 13 patients (77%) with normalfractional shortening were also unable to completestage I without symptoms. Other echocardiographicvariables, including ventricular septal thickness, leftventricular free wall thickness, aortic root dimension,and left atrial dimension, as well as the calculatedelectrocardiographic Romhilt-Estes point score, didnot differ between the two groups.

Exercise Capacity vs Left VentricularEjection Fraction

Echocardiographic left ventricular fractionalshortening was compared with resting radionuclideleft ventricular ejection fraction. Figure 5 is a com-

parison of fractional shortening and resting ejectionfraction data for the 24 patients in this study and foran additional 39 patients with aortic regurgitationstudied in our laboratory by these techniques. Therewas good correlation between these two independentmeasures of left ventricular systolic function for all 63patients (r = 0.85) and for the 24 patients included inthis study (r = 0.82). The relation between ventricularsize by echocardiography and resting and exerciseejection fractions by radionuclide cineangiography hasbeen reported separately.27

(Dzz

cr_0I

_i,z0

C:)

40!r = 0.85

30 L

20 H o--Q 00@0

10 H

(C0 10 20 30 40 50 60 70 80

LV EJECTION FRACTION

FIGURE 5. Echocardiographic left ventricular (LV) frac-tional shortening compared with resting radionuclide LVejection fraction. Data are shown for the 24 patients in thisstudy (solid circles) and for 39 other patients with aorticregurgitation studied by these two techniques (open circles).The dashed line indicates the lower limits ofnormalfor eachvariable. The correlation coefficient was 0.82 for the 24patients and 0.85 for the 63 patients.

Radionuclide left ventricular ejection fractions atrest in the 24 patients in this study ranged from29-59% (mean 42%) and were significantly lower thanvalues previously reported for normal subjects.'2Twelve patients had normal ejection fractions at rest(at least 45%). Ejection fractions significantly de-creased during maximal supine exercise (fig. 6) andwere subnormal in all patients (less than 55%), rang-ing from 14-52%. Despite greater exercise capacity asassessed by upright exercise studies, the 15 patientswho completed stage I of treadmill exercise withoutsymptoms had ejection fractions at rest and during ex-ercise that were not significantly different from thenine patients unable to complete stage I of treadmillexercise because of limiting symptoms (fig. 6). Eightpatients who could not complete stage I and 10patients who completed stage I had severely depressedleft ventricular ejection fractions during exercise (lessthan 40%).

Postoperative Survival

Preoperative echocardiographic evidence of leftventricular systolic dysfunction, manifested by sub-normal left ventricular fractional shortening (less than29%) or left ventricular end-systolic dimension greaterthan 55 mm, identified patients at high risk ofoperative death or death from congestive heart failureduring long-term postoperative follow-up (fig. 4), aswe have previously demonstrated.8 Prediction ofoperative or late death was considerably improved,however, when exercise capacity was also considered.Thus, no deaths occurred in any of the 18 patients whocould complete stage I of treadmill exercise beforeoperation, despite the fact that many of these patientshad evidence of systolic dysfunction by echocar-diography (fig. 4). Analysis of the 32 patients withsubnormal preoperative left ventricular fractionalshortening (fig. 7) indicated that the 15 patients who

1285

'i

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 6, DECEMBER 1980

70

C.)

h-01)

z0

1=

0LLz0F-0Ww

0

H

zW

tLW_j

60

50

40

30

20

10

0

<22.5 min >22.5 min

PREOPERATIVEEXERCISE DURATION

FIGURE 6. Preoperative left ventricular ejection fractionsat rest and during maximal supine exercise obtained bytechnetium-99m radionuclide cineangiography. Patientswho could not complete stage I of treadmill exercise (< 22.5minutes) did not significantly differ at rest or during exercise(Ex) from patients who completed stage I (> 22.5 minutes).The dashed line at 55% indicates the lower limit of normalfor ejection fraction during exercise.

completed stage I without symptoms before operationhad significantly better survival after operation thanthe 17 patients unable to complete stage I (100% sur-vival at 3 years vs 53% survival, p < 0.01). Similaranalysis of the smaller subgroup of patients withpreoperative left ventricular systolic dimensiongreater than 55 mm (20 patients) also revealed bettersurvival in the nine patients who completed stage I

before operation than in the 11 patients unable tocomplete stage I (100% survival at 3 years vs 38% sur-vival, p = 0.05).

5 4

z

D(I)

zLl

:-40

20 F

Exercise Duration >22.5 min

5

Exercise Duration < 22.5 min4

p < 0.01

1 2 3 4 5

TIME FOLLOWING OPERATION IN YEARS

FIGURE 7. Survival after operation in the 32 patients withsubnormal preoperative left ventricular fractional shorten-ing (less than 29%) plotted by the method of Mantel andHaenszel.24 The vertical lines with bars indicate the standarderror of the technique. The 15 patients who completed stageI before operation (exercise duration > 22.5 minutes) hadsignificantly better survival than the 17 patients who couldnot complete stage I (exercise duration < 22.5 minutes),with a 3-year survival of 100% compared to 53%. Thenumber of patients at risk is indicated at each interval.

When the survival of all 45 patients was analyzed(life-table method) independent of echocardiographicassessment of left ventricular function, exercisecapacity alone was helpful in determining relativepostoperative prognosis: 100% survival at 3 years inthe 18 patients who could complete stage I withoutsymptoms before operation, and 68% survival in the27 who could not (p < 0.05).

Postoperative Left Ventricular Function

Left Ventricular Diastolic Size

Among the 32 patients with subnormal preoperativeleft ventricular fractional shortening, preoperativeechocardiographic left ventricular end-diastolicdimension was not significantly different betweenpatients who could complete stage I of treadmill exer-cise and those who could not (fig. 8). The patients whocompleted stage I before operation, however, achievedsignificantly greater reduction in end-diastolic dimen-sion 6 months after operation than patients unable tocomplete stage I (mean decrease 23 ± 7 vs 8 ± 9 mm,p < 0.001), such that their end-diastolic dimensionswere significantly lower (mean 56 ± 8 vs 68 ± 11 mm,p < 0.005). Of the 32 patients, all but five haddecreases in left ventricular end-diastolic dimensionafter operation of 10 mm or greater; these five (whosechange in diastolic dimension ranged from a 4 mmdecrease to a 5 mm increase) all had limited exercisecapacity before operation and could not completestage I because of symptoms. Eight patients had per-sistent left ventricular diastolic dilatation of 70 mm orgreater at 6-month postoperative study (fig. 8) and five

nt

CIRCULATION1286

u

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

EXERCISE CAPACITY IN AR/Bonow et al.

90

E

z0C,,zW

C-)0aU_i-

0

cl

6zW

C-)

c-zW

80

70

60

50

40

30

20

10

0<22.5 min >22.5 min

PREOPERATIVE EXERCISE DURATIONFIGURE 8. Changes in echocardiographic left ventricularend-diastolic dimension as a result of operation in the samepatients shown in figure 4. Patients who completed stage Ibefore operation (> 22.5 minutes) had a greater decrease indiastolic size than patients who could not complete stage I(< 22.5 minutes). The dashed line at 70 mm indicates thevalue of postoperative diastolic dimension above whichpatients have been shown to be at high riskfrom subsequentdeath from congestive heart failure.8 Symbols and ab-breviations are explained in figure 1.

Left Ventricular Ejection FractionThe trends encountered in postoperative left ven-

tricular diastolic size by echocardiography wereparalleled by the changes in left ventricular systolicfunction by radionuclide cineangiography (fig. 9).Resting left ventricular ejection fraction improved in20 of 22 patients after operation, and was normal(45% or greater) in 18 patients. Patients who com-pleted stage I of treadmill exercise before operationdid not have significantly higher postoperative ejectionfractions at rest than patients unable to complete stageI. However, while ejection fraction during exercise in-creased after operation in 22 of 23 patients, thepatients who completed stage I of treadmill exercisebefore operation manifested greater improvement andhad higher exercise ejection fractions after operationthan patients unable to complete stage I (mean58 ± 15% vs 42 ± 8%,p < 0.01). Moreover, althoughsix of eight patients unable to complete stage I beforeoperation had normal ejection fractions after opera-tion at rest, ejection fraction during exercise remainedsubnormal in all eight (fig. 9). In contrast, nine of 15

REST

c

C0C.Lz0

U

ccI-

LA.z0

&I-

z

-LJ

<22.5 min >22.5 min

EXERCISE

S22.5 min >22.5 min

of the eight died from congestive heart failure. Sevenof these eight patients (including the five late deaths)could not complete stage I exercise before operation(p < 0.05). Two additional patients (not shown infigure 8), who had subnormal left ventricular frac-tional shortening and could not complete stage I exer-cise, died from congestive heart failure before 6-month postoperative study.

PREOPERATIVE EXERCISE DURATIONFIGURE 9. Changes in technetium-99m radionuclide leftventricular ejection fractions at rest and during maximalsupine exercise. Patients who completed stage I oftreadmillexercise before operation (> 22.5 minutes) did not differ inresting ejection fraction from patients who could not com-plete stage I (< 22.5 minutes), but had significantly higherpostoperative exercise ejection fractions. Dashed lines in-dicate the lower limit of normalfor ejection fraction at rest(45%) and during exercise (55%).

1287

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 6, DECEMBER 1980

patients (60%) who completed stage I before operationhad normal exercise ejection fractions after operation(p < 0.01). Of 10 patients who completed stage Ibefore operation despite severely depressed exerciseejection fractions (less than 40%), five had normal ex-ercise ejection fractions after operation (ranging from58-74%) and three were near normal (52%, 52% and54%).

Exercise ejection fraction failed to improve afteroperation in only one patient (fig. 9). This patient hadevidence of intraoperative myocardial damagemanifested by new anteroseptal Q waves on post-operative ECG and new septal dyskinesia on post-operative radionuclide study. This patient is also theonly patient who completed stage I exercise beforeoperation whose postoperative echocardiographic leftventricular diastolic dimension failed to decreasebelow 70 mm (fig. 8).

Influence ofPreoperative Hemodynamic Values

Patients who could not complete stage I exercisebefore operation because of symptoms hadsignificantly higher preoperative pulmonary arterywedge and left ventricular end-diastolic pressures thanpatients who could complete stage I (fig. 2). A mul-tivariate analysis, however, demonstrated that thedifferences in postoperative diastolic size and systolicfunction between the two groups were not significantlyaltered after adjustment for differences in preoperativehemodynamics, and that the preoperative hemo-dynamic data did not further enhance the accuracy ofpreoperative exercise capacity in predicting revers-ibility of left ventricular dysfunction. For example,postoperative echocardiographic left ventricular end-diastolic dimensions, and changes in this dimensionfrom before to after operation, did not significantlydiffer between patients with pulmonary artery wedgepressures less than 15 mm Hg and those with wedgepressures of 15 mm Hg or greater, whether thisanalysis was performed on all patients, on patientswho could not complete stage I exercise before opera-tion, or on patients who could complete stage I.Moreover, a multivariate analysis of the postoperativesurvival data relative to preoperative pulmonaryartery wedge or left ventricular end-diastolic pressurefailed to enhance the significant correlations providedby exercise capacity alone (fig. 7).

DiscussionPatients with aortic regurgitation may remain

asymptomatic for many years despite the burden of alarge volume load on the left ventricle.', 2, 28, 29 Themajor dilemma in managing these patients is deter-mining the proper timing of operation. Valve replace-ment in all asymptomatic patients would result inoperative deaths or prosthetic valve complications inmany patients who would otherwise have been asymp-tomatic for many years. On the other hand, whensymptoms alone are used as the criteria for operation,many patients who survive valve replacement will

develop progressive left ventricular decompensationand die many months after operation from congestiveheart failure.'-8 Such patients have insidiouslydeveloped irreversible myocardial dysfunction whileasymptomatic or only mildly symptomatic. Thus,more objective and prognostically relevant measuresof left ventricular function are required to determinethe optimal timing of aortic valve replacement.

In a recent study of 50 patients with aorticregurgitation who underwent valve replacement at theNIH, of which 31 patients in the present series form apart, we found that preoperative echocardiographicindexes of left ventricular systolic dysfunction (leftventricular fractional shortening less than 25% andleft ventricular dimension at end systole greater than55 mm) are highly sensitive in identifying patients atrisk of dying after operation because of congestiveheart failure.8 We also recently reported radionuclidecineangiographic studies demonstrating that patientswho had severely reduced left ventricular ejection frac-tions during exercise (less than 40%) before operationusually have persistent left ventricular dysfunctionafter operation, and only rarely does ejection fractionreturn to normal at rest or during exercise.30

However, although all patients (with the exceptionof the occasional patient sustaining intraoperativemyocardial damage) who die after operation fromcongestive heart failure could be identified by detec-tion of preoperative systolic dysfunction, manypatients with impaired function did well after opera-tion. Moreover, other institutions have reported thatreturn of normal left ventricular systolic function fre-quently occurs after valve replacement." 31-33 Wetherefore sought to document whether maximum exer-cise testing in symptomatic patients with aorticregurgitation provided additional insights and predic-tive information regarding preoperative left ven-tricular function, reversibility of preoperative left ven-tricular dysfunction, and long-term survival afteroperation.The results of the present study demonstrate that

preoperative treadmill exercise testing alone does notidentify patients who have severe left ventriculardysfunction. Exercise capacity did not correlate withindexes of resting left ventricular systolic functionderived from echocardiographic studies, or with leftventricular ejection fractions at rest or during exercisederived from radionuclide cineangiographic studies.In fact, some patients with the greatest impairment ofleft ventricular systolic function exercised muchlonger, and to higher levels of heart rate and oxygenconsumption, than other patients with apparently nor-mal systolic function. Moreover, patients who com-pleted stage I of the NIH exercise protocol, sur-prisingly, showed greater left ventricular dilatation bycontrast angiography and by echocardiography thanthose who could not complete stage I because ofsymptoms (figs. 3 and 4). These findings document theclinical impression that certain patients with aorticregurgitation develop severe ventricular dysfunctionwhile virtually free of symptoms and with apparentlynormal exercise capacity.2

1288 CIRCULATION

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

EXERCISE CAPACITY IN AR/Bonow et al.

Since objective determination of preoperative exer-cise capacity is not useful in assessing left ventricularfunction in symptomatic patients with aorticregurgitation, it would appear to be a poor method, initself, with which to prospectively screen or selectpatients for operation. Despite the lack of correlationwith echocardiographic or radionuclide indexes of leftventricular function, however, preoperative exercisecapacity appears capable of enhancing our ability topredict long-term postoperative survival beyond thatprovided by documentation of preoperative left ven-tricular systolic dysfunction,8' 3 and to predict revers-ibility of left ventricular dysfunction.

Regarding long-term survival, none of the patientswith subnormal preoperative fractional shortening orend-systolic dimension greater than 55 mm who couldcomplete stage I of the NIH exercise protocol withoutsymptoms died at or after operation (fig. 7). In con-trast, over 50% of the patients with subnormalpreoperative fractional shortening or end-systolicdimension greater than 55 mm who could not com-plete stage I exercise because of limiting symptomsdied at operation or during long-term postoperativefollow-up (figs. 4 and 7). Likewise, while all patientswho completed stage I of the exercise protocol im-proved significantly after operation in left ventriculardiastolic size, with reduction to less than 70 mm in allbut one patient (fig. 8), seven of 13 patients (54%) whocould not complete stage I of the exercise protocol hadpotstoperative diastolic dimensions of 70 mm orgreater. The improvement in left ventricular end-diastolic dimension is important because post-operative echocardiographic evidence of persistent leftventricular dilatation (70 mm or greater) aftertechnically successful valve replacement strongly cor-relates with subsequent late death from congestiveheart failure.8

Regarding reversibility of left ventricular dysfunc-tion, we found that although all patients who couldcomplete stage I of the exercise protocol had subnor-mal left ventricular ejection fractions during exercisebefore operation, nine of 15 (60%) had return of exer-cise ejection fraction to the normal range, includingfive of 10 patients (50%) with preoperative exerciseejection fractions of less than 40%. In contrast, noneof the eight patients who could not complete stage I ofexercise had normal exercise ejection fractions afteroperation (fig. 9).

Thus, in symptomatic patients with aorticregurgitation and left ventricular dysfunction, goodpreoperative exercise capacity appears to indicate thatirreversible myocardial dysfunction has not de-veloped; the converse is true for patients with leftventricular systolic dysfunction who have poorpreoperative exercise capacity. These results suggestthat subtle factors that influence exercise capacity, butwhich are not reflected in usual indexes of left ven-tricular systolic function at rest or exercise, are impor-tant determinants of reversibility of myocardial ab-normalities.Our data support the concept that patients with aor-

tic regurgitation should undergo operation once left

ventricular systolic dysfunction is detected under rest-ing conditions,8' 3 even though appreciable deteriora-tion in symptoms and exercise capacity is not yet evi-dent; once the latter occurs, irreversible myocardialdamage is likely. These findings also demonstrate thatthe incidence of reversibility of left ventriculardysfunction after operation for aortic regurgitation atany institution depends to a large extent on the degreeof symptomatic limitation of the patients at the timeof operation.

AcknowledgmentThe authors acknowledge the excellent and thorough technical

assistance of Cora Burn, Joyce McKay and Estelle Cohen in per-forming the echocardiographic and treadmill studies, and BonnieMack and Susan Farkas in performing the radionuclide studies. Theassistance of Dr. John Hyde of the Mathematics and AppliedStatistics Branch of the Division of Heart and Vascular Diseases,National Heart, Lung, and Blood Institute, was also valuable.

References1. Gault JH, Covell JW, Braunwald E, Ross J: Left ventricular

performance following correction of free aortic regurgitation.Circulation 42: 773, 1970

2. Goldschlager N, Pfeifer J, Cohn K, Popper R, Selzer A: Thenatural history of aortic regurgitation: a clinical andhemodynamic study. Am J Med 54: 577, 1973

3. Cohn PF, Gorlin R, Cohn LR, Collins JJ Jr: Left ventricularejection fraction as a prognostic guide in surgical treatment ofcoronary and valvular heart disease. Am J Cardiol 34: 136,1974

4. Isom OW, Dembrow JM, Glassman E, Pasternack BS, SacklerJP, Spencer FC: Factors influencing long-term survival afterisolated aortic valve replacement. Circulation 49 (suppl II): II-154, 1974

5. Hirshfeld JW, Epstein SE, Roberts AJ, Glancy DL, MorrowAG: Indices predicting long-term survival after valve replace-ment in patients with aortic regurgitation and patients with aor-

tic stenosis. Circulation 50: 1190, 19746. Barnhorst DA, Oxman HA, Connolly DC, Pluth JR, Danielson

GK, Wallace RB, McGoon DC: Long-term follow-up ofisolated replacement of the aortic or mitral valve with the Starr-Edwards prosthesis. Am J Cardiol 35: 228, 1975

7. Smith HG, Neutze JM, Roche AHG, Agnew TM, Barratt-Boyes BG: The natural history of rheumatic aortic regurgita-tion and the indications for surgery. Br Heart J 38: 147, 1976

8. Henry WL, Bonow RO, Borer JS, Ware JH, Kent KM,Redwood DR, McIntosh CL, Morrow AG, Epstein SE: Obser-vations on the optimum time for operative intervention for aor-

tic regurgitation. I. Evaluation of the results of aortic valvereplacement in symptomatic patients. Circulation 61: 471, 1980

9. Lewis RP, Bristow JD, Griswold HE: Exercise hemodynamicsin aortic regurgitation. Am Heart J 80: 171, 1970

10. Bolen JL, Holloway EL, Zener JC, Harrison DC, AldermanEL: Evaluation of left ventricular function in patients with aor-

tic regurgitation using afterload stress. Circulation 53: 132,1976

11. 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

12. Borer JS, Bacharach SL, Green MV, Kent KM, Henry WL,Rosing DR, Seides SF, Johnston GS, Epstein SE: Exercise-induced left ventricular dysfunction in symptomatic and asymp-tomatic patients with aortic regurgitation: assessment withradionuclide cineangiography. Am J Cardiol 42: 351, 1978

13. Marcus ML, Schuette WH, Whitehouse WC, Bailey JJ, GlancyDL: An automated method for the measurement of ventricularvolume. Circulation 45: 65, 1972

14. Feigenbaum H: Echocardiography, 2nd ed. Philadelphia, Lea

1289

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

VOL 62, No 6, DECEMBER 1980

& Febiger, 1976, p 46415. Henry WL, Clark CE, Epstein SE: Asymmetric septal hyper-

trophy (ASH): echocardiographic identification of thepathognomonic anatomic abnormality of IHSS. Circulation47: 225, 1973

16. Fortuin NJ, Hood WP Jr, Sherman ME, Craige E: Determina-tion of left ventricular volume by ultrasound. Circulation 44:575, 1971

17. Popp RL, Harrison DC: Ultrasound cardiac echography fordetermining stroke volume and valvular regurgitation. Circula-tion 41: 493, 1970

18. Troy BL, Pombo J, Rackley CE: Measurement of left ven-tricular wall thickness and mass by echocardiography. Circula-tion 45: 602, 1972

19. Epstein SE, Beiser GD, Stampfer M, Robinson BF, BraunwaldE: Characterization of the circulatory response to maximal up-right exercise in normal subjects and patients with heart disease.Circulation 35: 1049, 1967

20. Epstein SE, Beiser GD, Stampfer M, Braunwald E: Exercise inpatients with heart disease: effects of body position, type and in-tensity of exercise. Am J Cardiol 23: 572, 1969

21. Green MV, Ostrow HG, Douglas MA, Myer RW, Scott RN,Bailey JJ, Johnston GS: High temporal resolution ECG-gatedscintigraphic angiocardiography. J Nucl Med 16: 95, 1975

22. Bacharach SL, Green MV, Borer JS, Douglas MA, OstrowHG, Johnston GS: A real-time system for multi-image gatedcardiac studies. J Nucl Med 18: 79, 1977

23. Green MV, Brody W, Douglas MA, Borer JS, Ostrow HG,Line BR, Bacharach SL, Johnston GS: A count based methodfor determination of ejection fraction from radionuclideangiograms. J Nucl Med 19: 880, 1978

24. Burggraf GW, Craige E: Echocardiographic studies of left ven-tricular wall motion and dimensions after valvular heart sur-gery. Am J Cardiol 35: 473, 1975

25. Kaplan EL, Meier P: Non-parametric estimation from in-complete observations. J Am Stat Soc 53: 457, 1958

26. Mantel N: Chi-square test with one degree of freedom: exten-sions of the Mantel-Haenszel procedure. J Am Stat Soc 58:690, 1963

27. Henry WL, Borer JS, Bonow RO, Rosing DR, Bacharach SL,Green MV, Epstein SE: Functional adaptations of the left ven-tricle to chronic aortic regurgitation. (abstr) Am J Cardiol 43:412, 1979

28. Segal J, Harvey WP, Hufnagel C: A clinical study of one hun-dred cases of severe aortic insufficiency. Am J Med 21: 200,1956

29. Spagnuolo M, Kloth H, Taranta A, Dolyle E, Pasternack B:Natural history of rheumatic aortic regurgitation: criteriapredictive of death, congestive heart failure and angina in youngpatients. Circulation 44: 368, 1971

30. Borer JS, Rosing DR, Kent KM, Bacharach SL, Green MV,McIntosh CL, Morrow AG, Epstein SE: Left ventricular func-tion at rest and during exercise after aortic valve replacement inpatients with aortic regurgitation. Am J Cardiol 44: 1297, 1979

31. Schuler G, Peterson KL, Johnson AD, Francis G, Ashburn W,Dennish G, Daily PO, Ross J: Serial noninvasive assessment ofleft ventricular hypertrophy and function after surgical correc-tion of aortic regurgitation. Am J Cardiol 44: 585, 1979

32. Gaasch WH, Andrias CW, Levine HJ: Chronic aorticregurgitation: the effect of aortic valve replacement on left ven-tricular volume, mass, and function. Circulation 58: 825, 1978

33. Clark DG, McAnulty JH, Rahimtoola SH: Valve replacementin aortic insufficiency with left-ventricular dysfunction. Circula-tion 61: 411, 1980

34. Henry WL, Bonow RO, Rosing DR, Epstein SE: Observationson the optimum time for operative intervention for aorticregurgitation. II. Serial echocardiographic evaluation ofasymptomatic patients. Circulation 61: 484, 1980

1290 CIRCULATION

by guest on April 18, 2017

http://circ.ahajournals.org/D

ownloaded from

and S E EpsteinR O Bonow, J S Borer, D R Rosing, W L Henry, A S Pearlman, C L McIntosh, A G Morrow

predictor of postoperative left ventricular function and long-term prognosis.Preoperative exercise capacity in symptomatic patients with aortic regurgitation as a

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1980 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.62.6.1280

1980;62:1280-1290Circulation. 

http://circ.ahajournals.org/content/62/6/1280the 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 April 18, 2017

http://circ.ahajournals.org/D

ownloaded from