disparity ejection and ventricular in -...

LABORATORY INVESTIGATIONVALVULAR HEART DISEASE

Disparity between ejection and end-systolic indexesof left ventricular contractility in mitralregurgitationBARBARA BERKO, M.D., WILLIAM H. GAASCH, M.D., NAOSHI TANIGAWA, M.D.,DAMON SMITH, B.M.E., AND ERNEST CRAIGE, M.D.

ABSTRACT To examine left ventricular function in mitral regurgitation (MR), we compared theejection phase indexes of left ventricular contractility with maximal systolic elastance (Emax) in anexperimental preparation of MR. In eight anesthetized open-chest dogs, pressure-volume loops werederived during afterload manipulation with methoxamine and nitroprusside from simultaneous leftventricular pressure and dimensional (sonomicrometry techniques) data before and after creation ofMR. From these data maximal systolic elastance (Emax), the end-systolic pressure-volume relation-ship (ESPVR), and the end-systolic stress-volume relationship (ESSVR) were determined by linearregression analysis. After creation of MR, end-diastolic volume increased significantly (40 ± 13 to 53+ 18 ml, p < .001); likewise end-systolic volume increased (28 ± 11 to 33 ± 15 ml, p < .05).Ejection fraction increased after MR (35 ± 6% to 44 + 8%, p < .005), as did the mean velocity offiber shortening (0.62 + 0.20 to 1.02 ± 0.39 sec- 1, p < .02). In contrast, Emax declined significantly(4.63 ± 2.5 to 3.54 1.94 mm Hg/ml, p < .05); ESPVR and ESSVR showed similar directionalchanges. An inverse relationship was found between systolic elastance and end-diastolic volume inboth control and MR states. When Emax, ESPVR, and ESSVR were normalized to end-diastolicvolume, they were unchanged after MR. These results suggest that either there was a decline in leftventricular contractile state after MR, or that contractility was unchanged (if elastance is normalized forthe increase in heart size). In this preparation of MR the observed increase in shortening was not due toincreased contractility, but occurred as a consequence of increased preload with no significant changein afterload.Circulation 75, No. 6, 1310-1319, 1987.

EXPERIMENTAL and clinical studies have demon-strated exaggerated myocardial shortening in the pres-ence of mitral regurgitation (MR) and it has been sug-gested that this is due to favorable left ventricularloading conditions imposed by the regurgitant leak. 1-4The standard ejection or shortening phase indexes ofleft ventricular contractility are known to be influencedby alterations in preload and afterload5- and thus it islikely that the loading conditions present in MR canmask significant myocardial dysfunction that may be-come manifest after valve replacement.8'-5 The leftventricular end-systolic pressure-volume relationship

From the Departments of Medicine (Cardiology), University of NorthCarolina, Chapel Hill, and the Veterans Administration Medical Cen-ter, Boston.

Supported by a grant-in-aid from the American Heart Association,North Carolina Affiliate, with funds contributed in part by the Greens-boro Chapter of the Affiliate.

Address for correspondence: Barbara Berko, M.D. CardiovascularSection, 948 W. Gates Building, Hospital of the University of Pennsyl-vania, 3400 Spruce St., Philadelphia, PA 19104.

has been proposed as an index of contractile state thatis independent of acute changes in preload and after-load. This relationship is a linear function with a slope(maximal systolic elastance or Emax) that is sensitiveto changes in inotropic state.16-20 Because of its Iloadindependence," systolic elastance may provide a moreaccurate assessment of left ventricular function in thepresence of MR than standard ejection phase indexesof contractility. We therefore developed an animalpreparation in which to compare the ejection phaseindexes with Emax in MR. We then tested the hypoth-esis that the widely observed increase in shorteningmerely reflects altered loading conditions present inthis setting. Since we did not anticipate a change in leftventricular contractile state, we expected that Emaxwould remain unchanged in the presence of MR.

MethodsExperimental preparation. Ten mongrel dogs weighing 20

to 25 kg were sedated with morphine sulfate (2.0 mg/kg im).

CIRCULATION1310

by guest on May 18, 2018

http://circ.ahajournals.org/D

ownloaded from

http://circ.ahajournals.org/

LABORATORY INVESTIGATION-VALVULAR HEART DISEASE

Anesthesia was induced with intravenous sodium pentobarbitol(10 mg/kg) and morphine sulfate (60 mg iv). Thereafter, asurgical level of anesthesia was maintained with continuousintravenous infusion of morphine sulfate (20 to 25 mg/hr). Res-piration was controlled with a mechanical ventilator with inter-mittent positive pressure and supplemental 02 to maintain anarterial Po2 in excess of 90 mm Hg. A left thoracotomy wasperformed through the fifth intercostal space and the heart wassuspended in a pericardial cradle. To maintain intravascularvolume each dog was transfused with 100 to 200 ml of wholeblood.Two pairs of pulse-transit ultrasonic (piezoelectric) trans-

ducers2' were sewn to the epicardial surface of the left ventricleto measure the external long-axis (DL) and short-axis (Ds) di-mensions. Equatorial wall thickness was obtained from a pair ofsmall cylindrical ultrasonic crystals placed across the ventricu-lar wall in the area between the anterior papillary muscle and theseptum. The orientation of these crystals is shown in figure 1.The sampling rate of the sonomicrometer is 250 samples/sec.All three-dimensional signals obtained from the sonomicrom-eter were calibrated and were simultaneously recorded on di-rect-current input channels of a multichannel physiologicrecorder (Electronics for Medicine, model VR 12) with a fre-quency response bandwidth of 250 Hz. Aortic, left ventricular,and left atrial pressures were measured with high-fidelity micro-manometer-tipped catheters (Model PC-350, Millar Instru-ments) that were balanced and calibrated from 0 to 200 mm Hgjust before insertion. Calibrations were verified at the end ofeach experiment to ensure minimal drift. The aortic catheter waspositioned just above the aortic valve via the carotid artery, theleft ventricular catheter was introduced into the ventricular cav-ity from the left ventricular apex, and the left atrial catheter wasplaced via the atrial appendage. The first derivative of leftventricular pressure (dP/dt) was obtained by electronic differ-entiation of the left ventricular pressure signal. Simultaneousrecordings of the electrocardiogram, aortic and left ventricularpressures, dP/dt, and the three-dimensional signals were allobtained on the multichannel physiologic recorder at a paperspeed of 100 mm/sec.

Study protocol. After the dogs were instrumented, atropine(0.5 mg iv) and propranolol (0.15 mg/kg iv) were administeredto blunt reflex responses to arterial pressure manipulations.Baseline measurements of pressure and dimension were made;examples of control and baseline MR records are shown infigure 2. Afterload was then manipulated to obtain multiplepressure-volume loops (derived from pressure-dimension data).All dogs received methoxamine by continuous intravenous infu-sion at a rate sufficient to increase systolic pressure by at least 20mm Hg. In six dogs nitroprusside was also administered toreduce systolic pressure. For each animal, simultaneous pres-sures and dimensions were recorded at multiple pressure levels.

In eight dogs MR was created in the following manner. Apurse-string suture was sewn around the left atrial appendageand a metal hook placed through the appendage was used tosever chordae tendineae. Severe MR was indicated by the devel-opment of a systolic thrill over the left atrial wall, marked leftatrial distention, and left atrial regurgitant (v) waves of 25 to 50mm Hg. Animals were then allowed to stabilize for 90 min;inspired oxygen was adjusted to maintain physiologic arterialblood gases. Pressure and dimensional data were then obtainedat baseline and during afterload manipulation (just as in thecontrol state). At the end of each experiment the heart wasdissected to check the position of the endocardial thicknesscrystal and to confirm that chordae tendineae or mitral leafletswere torn without significant damage to the ventricular wall.Two dogs were used in sham experiments to determine that

the preparation would not deteriorate over time. In these ani-

mals the experimental preparation and baseline measurementswere performed exactly as described for the dogs with MRexcept that MR was not created. Instead, measurements weremade during a "baseline" period and again 90 min later; datawere obtained during afterload manipulation just as in the MRdogs.

Calculations. Analog data were recorded on a multichannelphysiologic recorder and were analyzed with an Apple II com-puter with an x-y digitizing tablet. The internal volume of theleft ventricle was calculated according to a prolate ellipsoidmodel with the formula:

V = (7m/6) (b-2h)2 (a-1. lh)

where b is the measured external Ds, a is the external DL, and his the equatorial wall thickness. This method has been validatedby Rankin et al.22 The ejection fraction was calculated from theend-diastolic volume (EDV) and the end-systolic volume(ESV), defined as the largest and smallest attained volumes.The left ventricular volumes and ejection fractions in our anes-thetized open-chest dogs are lower than those reported in studiesof conscious closed-chest dogs with the use of similar sonomi-crometric techniques.22 This is probably the result of alterationsin left ventricular size and geometry known to occur in open-chest preparations.23An index of the left ventricular ejection time was derived

from the interval between peak positive dP/dt and the smallestsystolic minor-axis dimension (ESD). The normalized mean

DD

Th

DLFIGURE 1. Schematic representation of the orientation of the ultrason-ic transducers on the left ventricle. DL represents epicardial crystalspositioned at the base between the aorta and pulmonary artery and theapex to measure external long-axis dimension. Ds are crystals placed onthe anterior and posterior epicardial surfaces near the interventricularseptum to measure external short-axis dimension. Th represents a pairof cylindrical crystals on the anterior endocardial and overlying epicar-dial surfaces to measure wall thickness.

Vol. 75, No. 6, June 1987 1311



ownloaded from


CONTROL MR0

1 11

dP/dt

/,-- ' ~~10o5

rz

0D -AotCS' N/1-- -I

E 12E

9

' am

LVP

r J

1L 1 LXLLL 12 [1:L 4jl Jrl:;1!1 14. I4 4~4FIGURE 2. Baseline pressure and dimensional data from a typical experiment. Left, Control data; right, data after creation ofMR. Simultaneous recordings of left ventricular (LVP) and aortic (Ao) pressures, external DL and Ds, and wall thickness (Th)are shown. The dP/dt is recorded above the other signals and the electrocardiogram (ECG) below. After creation of MR thesystolic pressure is slightly lower, wall thickness is decreased, and DL and Ds are increased.

velocity of circumferential fiber shortening (VCF, sec- 1) wascalculated as the endocardial fractional shortening (largest inter-nal diastolic short-axis dimension [EDD] minus ESD, dividedby EDD) divided by the left ventricular ejection time.24 A "ratecorrected" mean VCF was also determined with use of the rate-corrected ejection time.25

Circumferential midwall stress (in mm Hg) was calculated byMirsky's formula for a prolate ellipsoid26:

Stress =- 1-[2a]

h 2b 2a2

where P = pressure; h = wall thickness; a = midwall semima-jor axis (DL - 1. lh)/2; b = midwall semiminor axis (Ds -

2h)/2. Peak systolic stress was taken as the maximal circumfer-ential stress achieved during systole. "End systolic" wall stresswas calculated at the time of the aortic incisura; it was alsodetermined from coordinates at the time of smallest systolicvolume. Mean systolic stress was determined by integrating thewall stress throughout the period of ejection and dividing by thetime interval; mean developed systolic stress was derived bysubtracting end-diastolic stress.

Left ventricular volume was calculated from the dimensionaldata, and for each animal eight to 20 pressure-volume loopswere plotted during the control state and again after creation ofMR (figure 3). Simultaneous left ventricular pressure (or stress)and volume coordinates at the time of the incisura were identi-fied. Because of uncertainties regarding a precise definition of

end-systole in the presence of MR, we used three differentmethods of calculating the slope of the "end-systolic" pressure-volume relationship:

(1) Emax: The slope of the pressure-volume relationship wasdetermined from linear regression analysis of multiple pressure-volume coordinates taken at the time of maximal pressure-volume ratio (maximal elastance).

(2) End-systolic pressure-volume relationship (ESPVR): Theslope of the pressure-volume relationship was determined fromlinear regression analysis of multiple pressure-volume coordi-nates taken at the time of the aortic incisura.

(3) End-systolic stress-volume relationship (ESSVR): Theslope of the stress-volume relationship was determined fromlinear regression analysis of multiple stress-volume coordinatesat the time of the aortic incisura.Linear regression analysis by the least squares method was usedto determine the three elastance variables listed above. An in-verse relationship was found between elastance and baselineEDV for each state before load manipulation (see Results), suchthat a linear relationship existed between elastance and I/EDV(linear correlation coefficients for Emax vs I/EDV, r = .87;ESPVR vs 1/EDV, r = .92; ESSVR vs 1/EDV, r - .81). Thisimplies a hyperbolic function relating elastance and EDV.Based on this apparent dependence of elastance on heart size,elastance was normalized for heart size according to this rela-tionship. Thus, normalized elastance was calculated by multi-plying the value for manifest elastance by the baseline EDV foreach of the two states. This is mathematically equivalent to

CIRCULATION

BERKO et al.

200 r E 7Ez0 67.U)zW 62

li52.,

160f

120F

80-

401

LIEEWW:

al-

1%i

At

'i.i

oL

1312



ownloaded from



MR

15 20 25 30 35 40 45 50

Slope=4.29=-4.62

r= 0.885

15 20 25 30 35 40 45 50

VOLUME (ml)

FIGURE 3. Left ventricular pressure-volume loops from a representative experiment before (left) and after (right) creation ofMR. The broken line represents baseline data (before administration of methoxamine). The slope, intercept (V0), and correlationcoefficients (r) are given for the end-systolic pressure-volume relationship (Emax in this example) in each state. After MR, thelinear relationship between end-systolic pressure and volume was shifted to the right and the slope (Emax) decreased.

dividing the systolic volume of each of the pressure-volumecoordinates by the baseline EDV (i.e., each of the ESVs ob-tained by load manipulation is divided by the same EDV).

Diastolic pressure-volume relationships were analyzed bylinear regression analysis of pressure-volume data from individ-ual baseline cardiac cycles from the time of minimal left ventric-ular pressure to the next R wave. Pressure-volume coordinatesfor each state were fit to a linear relationship and to a simpleexponential function such that P = Bekv, where P = left ven-tricular pressure; V = volume; e = base of the natural loga-rithm, and B and k are constants. These constants were derivedfrom the relationship InP = kV + InB. Data could be equallywell fit to either the linear (r - .93 + .05, control; r = .96 +.04, MR) or exponential (r - .92 ± .06, control; r = .92 +.09, MR) relationships. To compare the diastolic pressure-vol-ume curves before and after MR in each animal, these relation-ships were derived from baseline data in each state, and the lineswere compared with the use of statistical tests for parallelismand common intercepts.27

Statistical analysis. All results are reported as the mean +one SD. A two-tailed Student's t test for paired data was used tocompare control and MR data and a p value of less than .05 wasconsidered indicative of a statistically significant difference.

ResultsPressure-volume loops from a single representative

experiment are shown in figure 3. Each loop was gen-erated from approximately 60 pressure-volume coordi-nates. In this example, EDV and stroke volume in-creased, the range of systolic pressure was lower, andthere was an increase in shortening after creation ofMR. The end-systolic pressure-volume coordinateswere shifted down and to the right and Emax declinedfrom 5.19 to 4.29 mm Hg/mi.

Immediately after creation of MR, peak systolicwall stress declined from 186 + 82 to 129 + 67 mmHg (p < .005); after the 90 min stabilization periodwall stress had nearly returned to control levels. Allcontrol and MR data (after the equilibration period) arelisted in tables 1 and 2. There was a tendency for leftventricular systolic pressure to decline, but not to astatistically significant extent; the average value for

TABLE 1Hemodynamic variables before and after mitral regurgitation

Pressure (mm Hg) Stress (mm Hg) Volume (ml) Ejection indexes

PSP ESP LVEDP PSS MDSS ESWS EDV ESV EF (%) VCF (sec -')

Control 134+15 95+14 7+2 184+76 101+34 114+51 40+13 28+11 35+6 0.62+0.20MR 117+11 95+14 19+8 184+73 92+34 118+71 53+18 33+15 44+8 1.02+0.39p valueA NS NS <.005 NS NS NS <.001 <.05 <.005 <.02

All values are mean + SD.PSP = peak systolic pressure; ESP = end-systolic pressure at the incisura; LVEDP = left ventricular end-diastolic pressure; PSS = peak systolic

wall stress; MDSS = mean developed systolic stress; ESWS = end-systolic wall stress at the incisura; EF = ejection fraction.

Apaired t test of control vs MR; NS = not significant at p > .05.

CONTROL

131 3Vol. 751 No. 6, June 1987



ownloaded from


BERKO et al.

TABLE 2Systolic elastance before and after MR

ESPVR ESSVR Emax

Slope V0 r value Slope V0 r value Slope V0 r value

Control 5.33 2.33 8 ± 9 .959 + .030 9.43 + 2.67 17±+9 .984-+.008 4.63 ±+2.15 - 4±+ 14 .967 ±.016MR 4.16+1.93 3±33 .912±.032 7.18±2.71 15+13 .953+.025 3.54+1.94 -11±31 .857+ .150p valueA <.05 NS <.005 NS <.05 NS

All values are the mean + SD.Slope slope of linear regression line for each relationship (in mm Hg/ml); V0

.volume intercept (in ml) at zero pressure or stress; r = linear

correlation coefficient.APaired t test of MR vs control.

end-systolic pressure did not change. Left atrial pres-sure increased significantly and large V waves devel-oped after MR, with a wave/v wave pressures increas-ing from 9 ± 4/11 + 5 mm Hg in the control state to20 ± 6/31 ± 12 mm Hg after MR. Left ventricularend-diastolic pressure increased, and both EDV andESV increased significantly. Accompanying this vol-ume increase the ventricular chamber became morespherical, as manifested by a decline in the ratio of themajor-to-minor internal dimensions from 1.69 ± 0.20in the control state to 1.61 ± 0.16 after MR (p < .01).On visual inspection, the diastolic pressure-volumerelationship was shifted to the right after MR. Thisdifference in the diastolic curves was confirmed bystatistical comparison of the lines relating lnP and vol-ume (or pressure and volume) throughout diastole inthe control state and after MR. The intercepts (B) weresignificantly different after MR in all animals, and theslopes (k) were significantly different in six of eight.

There was a tendency for mean systolic stress todecline after MR, but the average value for systolic

EF

60 - p<O.00555

50

40 T

35-

30

25_If 1

C MR

a)

en

.U

.0

0

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0

VCF

C MRFIGURE 4. Ejection phase indexes before and after the creation of MR.Solid circles represent individual data for each dog in the control state

(C) and after creation ofMR. Open circles represent the mean values forall dogs, with one SD shown. Ejection fraction (EF) and VCF were bothsignificantly increased after MR.

stress was not significantly different from the controlvalue. Peak systolic stress, mean systolic stress, andend-systolic stress (whether defined at the aortic inci-sura or at the time of smallest systolic volume) werenot significantly different after MR from the controlstate. Ejection fraction and mean VCF increased signifi-cantly (figure 4). Because of the increase in heart rate(113 ± 15 to 150 + 26 beats/min, p < .01), VCF wasnormalized for heart rate; the rate-corrected VCF alsoincreasedfrom0.46 ± 0.16toO.65 + 0.22sec '(p<.05). Peak positive dP/dt was unchanged after MR(control, 1886 + 241 mm Hg/sec; MR, 1796 + 353mm Hg/sec, NS).The average results for the three end-systolic index-

es are shown in table 2 and figure 5. In all animals themaximal pressure-volume ratio, used to derive Emax,occurred within 30 msec before aortic valve closure inboth the control and MR states. All of the end-systolicrelationships were highly linear in the control state andafter creation of MR, but there was a tendency for thelinear correlation coefficient to be lower after MR.This trend was least apparent when stress was used inthe calculation of elastance. All three methods re-vealed a significant decline in the slope of theend-systolic pressure (or stress)-volume relationship,and there was no significant change in the volumeintercept.

Based in part on the suggestion of Sagawa2t andothers29 that Emax should be normalized for heart orbody size, we plotted the slope of the pressure-volumerelationship against the baseline (before pharmacolog-ic afterload manipulation) EDV for the control andMRstates to determine the relationship between elastanceand heart size. The data shown in figure 6 indicate aninverse relationship between Emax and EDV, whichcan be described by a hyperbolic function.29 This in-verse relationship, which was present in the controland MR states, forms the basis for normalizing systolicelastance by the baseline EDV; our rationale for such

CIRCULATION

c0

0

a)LL

LLI

er,

1314



ownloaded from



ESPVR

p<O.05

[ MC MR

ESSVR

C MR C

FIGURE 5. Systolic elastance before and after creation of MR. Solid circles represent individual data from each dog in thecontrol state (C) and after MR. Open circles represent mean data, with one SD shown. Data are shown for three different methodsof determining systolic elastance (see text for definitions of ESPVR, ESSVR, and Emax). Data shown are the slopes of eachrelationship in mm Hg/ml. The slopes of all three relationships decreased significantly after MR.

normalization is given in the discussion below. WhenEmax was normalized for the baseline EDV, the nor-

malized values were essentially equal in the controland MR states (table 3). This was also true for ESPVRand ESSVR (table 3).To confirm the stability of our anesthetized open-

chest dog preparation we performed two sham experi-ments. The hemodynamic variables were measuredduring a control state and again 90 min later (at a timeequivalent to that for MR data acquisition). In the firstdog the ejection fraction values were 42% vs 42%,ESPVR values were 5.61 vs 6.39 mm Hg/ml, andEmax values were 4.97 vs 4.92 mm Hg/ml. In thesecond dog the ejection fraction values were 27% vs

10

EN1.1

I

E

Ei

8-

6-

4-

2-

20 40 60 80 100

LV END DIASTOLIC VOLUME (ml)

FIGURE 6. The relationship of Emax and EDV. Emax is plottedagainst the baseline EDV before load manipulation for control (closedcircles) and MR (arrow) in each dog. The curve represents a least

squares best fit (hyperbolic function) of all 16 coordinates. In seven of

eight experiments, the decline in Emax was associated with an increase

in EDV.

29%, ESPVR values were 3.34 vs 3.80 mm Hg/mI,and Emax values were 2.61 vs 3.33 mm Hg/mi.

DiscussionThe problems encountered in the assessment of

myocardial function in the presence of an incompetentmitral valve have been appreciated for many years.Early studies by Braunwald et al.' in a preparation ofacute MR demonstrated a substantial increase in totalleft ventricular output and stroke volume and it wassuggested that this might be due to facilitated ejectioninto the lower pressure left atrium during systole. Sub-sequent studies by Urschel et al.2 and others 3' 4 3 14revealed an increase in both the extent and velocity offiber shortening after MR, which was attributed to anincrease in preload and reduction in late systolic walltension. Under these conditions, they found no changein Vmax (an isovolumetric index of contractility) andthey concluded that myocardial contractility is un-

TABLE 3Normalized elastance before and after MR

ESPVRn ESSVRn Emax,,Control 1.92 -+0.41 3.84 2.20 1.67 - 0.44MR l.99+0.72 3.90+2.81 1.61 +0.53p valucA NS NS NS

All values represent the mean + SD; ESPVRn, ESSVR , and Emaxnrepresent normalized elastance (slope normalized for baseline EDV);units are mm Hg x 102.

APaired t test of control vs MR.

Vol. 75, No. 6, June 1987

16

14

121

LJ

-Jn-

10

8

6

4

2

Emax

p<0.05

IMR

0

1 & i -1 1- 1- 1

1 315



ownloaded from


BERKO et al.

changed in acute MR; they also suggested that theejection phase indexes overestimate contractile state inthe presence of this lesion. Alterations in preload areknown to affect Vmax'6 and the use of an isovolumetricindex is at best questionable when no true isovolume-tric phase exists. For these reasons V.ax has not prov-en to be a reliable index of left ventricular contractilityin this disorder.To assess left ventricular function in the setting of

MR it is necessary to use an index of contractile statethat is not affected by altered loading conditions. Theend-systolic pressure-volume relationship has beenproposed as such an index. '9 In the isolated canine leftventricle Suga et al. 16 found that the maximum slope ofthe end-systolic pressure-volume relationship wasindependent of acute changes in preload and afterload,but was sensitive to changes in inotropic state, becom-ing steeper with increased contractility. 16 1" While theunique load-independence of this relationship wouldseem to make it ideal for examining ventricular func-tion in the setting of MR, its use is complicated bycontinued ejection of blood into the left atrium afteraortic valve closure,30' 31 thus precluding a simple defi-nition of end-systole. Early studies by Suga et al.'6indicate that the slope of the instantaneous relationshipof pressure and volume increases with time throughoutsystole, reaching a maximum value near the end ofejection, nearly coincident with the uppermost left cor-ner of the pressure-volume loop. Several approxima-tions of this relationship have been developed, themost commonly used being the end-systolic pressure-volume relationship determined at the aortic inci-sura32 33 and the simple ratio of pressure or wall stressover volume at end-ejection.34 3 Since "end systole" isnot clearly synonymous with "end ejection" in the set-ting of MR, we chose to evaluate three different meth-ods of approximating systolic elastance. Using a widerange of loading conditions we determined Emax andtwo other end-systolic indexes, the pressure-volumeand the stress-volume relationships at the time of theaortic incisura. It is of interest that in all animals themaximal pressure-volume ratio (used to define coordi-nates for Emax) occurred before aortic valve closure inboth the control state and after MR, suggesting thatcontinued ejection into the left atrium after aortic valveclosure with MR has no significant impact on systolicelastance and that aortic valve closure closely approxi-mates other definitions of end-systole.Our results are consonant with previous studies

demonstrating a significant increase in both the extentand velocity of fiber shoiAening after the creation ofMR. In sharp contrast are the results of the systolic

elastance analysis: regardless of the method used todetermine systolic elastance, a significant decrease inthis variable was observed after the development ofMR. Systolic elastance is known to be sensitive tochanges in left ventricular contractile state,6- 8 andthus data indicate a decline in contractility with MR.We therefore considered what factors might result in adecrease in contractility and/or a decrease in chamberelastance with no change in contractility in our prep-aration of MR.

Although the animals became transiently hypoxe-mic and on occasion mildly acidotic with the creationof MR, these abnormalities were corrected during theequilibration period, and the studies were performedwithout hypoxemia or acidosis. Temperature was care-fully monitored and no dog became hypothermic dur-ing the course of an experiment. The stability of ourpreparation over time was documented in the two shamexperiments; no decline in maximum elastance wasfound during the course of these experiments. Like-wise, the effect of changes in heart rate would not beexpected to cause a decrease in contractile state. If ithas any effect, increased heart rate should result in anincrease in inotropic state,36 not a decline. Within therange of heart rates observed in our study, others havenot found a major effect of heart rate on maximumelastance.36 3 Therefore, the observed decrease in sys-tolic elastance cannot be attributed to "external" fac-tors resulting from experimental design.

It is widely accepted that the end-systolic pressure-volume relationship is insensitive to acute changes inloading conditions, but under experimental circum-stances this relationship does exhibit some loaddependence. For example, at a constant end-systolicvolume, the end-systolic pressure declines as the ejec-tion fraction increases; this can result in a decrease insystolic elastance.38 The mechanisms underlying thisphenomenon are not well understood, but it is probablydue, at least in part, to shortening deactivation. Themagnitude of the dependence of maximum elastanceon shortening is said to be quite small relative to itsdependence on left ventricular contractility. There-fore, while the effect of increased shortening may con-tribute to a decline in elastance, it is unlikely thatincreased shortening alone could explain the consider-able decrease in systolic elastance that was observed inour study.

Another possible cause of the decline in systolicelastance is the mechanical effect of ventricular dilata-tion itself. It is conceivable that short-term mechanicalstresses sufficient to cause ventricular dilation couldproduce myocardial injury resulting in a depression of

CIRCULATION1316



ownloaded from


LABORATORY INVESTIGATION-vALVULAR HEART DISEASE

myocardial function. Although one cannot definitivelyrule out the possibility that some myocardial damagehas occurred to depress contractile state, there are nopublished data to support this notion. Therefore, weconsidered the effect of a change in heart size as anindependent factor that might influence systolic elas-tance. It is apparent that a larger ventricular chamberwill result in a shift of the pressure-volume loop to theright, although myocardial contractile state may beunchanged. In fact, we found an inverse relationshipbetween systolic elastance and baseline EDV (figure 6)similar to that observed by Bogen et al.29 The observa-tion that elastance is lower in larger hearts prompted usto normalize the elastance variables for the baselineEDV before and after creation of MR. This type ofnormalization is not appropriate for very acute changesin preload in which the ventricle merely moves up ordown a single diastolic pressure-volume curve. How-ever, in our study the MR data were obtained 1 to 2 hrafter creating MR, during which time several impor-tant changes in the left ventricle had occurred. Thelarge increase in EDV afterMR was accompanied by arightward shift in the diastolic pressure-volume rela-tionship, a more spherical left ventricular geometry,and an increase in systolic stress (returning to controllevels) after an initial decline. This constellation ofvariables suggests that the ventricular chamber hasundergone considerably greater changes than simplyan acute increase in preload; the pressure-volume rela-tionship may already be reset to a different baseline,comparable to that in subacute or chronic MR. Thus, itseems reasonable to normalize elastance to account forthe change in heart size, much as one might if thesechanges were more chronic. We have considered whatother methods of normalization might be used to ac-count for heart size changes, including VO, the extrapo-lated volume at zero pressure, and left ventricularmass. Neither VO nor left ventricular mass changed inthese short-term experiments, and therefore there is norationale for such normalization.When the baseline EDV for each state was used, the

normalized values of all three elastance parameterswere unchanged, implying that there may be no changein left ventricular contractile state with MR (table 3).This conclusion is supported by our observation thatpeak positive dP/dt did not change after creation ofMR. These results are similar to those reported byTajimi et al.,39 who reported data from dogs with MRthat were studied repeatedly over several weeks. Theyfound a progressive increase in ventricular volume andejection fraction, but the end-systolic pressure-volumerelationship decreased. However, when they normal-

Vol. 75, No. 6, June 1987

ized their data for the baseline EDV, no significantdifference was found between the slopes of the ESPVRbefore and after MR. They interpreted these results asindicating that there is no change in contractility afterMR when the dependence of systolic elastance onbaseline EDV is considered.

It seems possible, therefore, that the decline inmanifest (non-normalized) systolic elastance could bethe result of adaptive mechanical changes such as fiberrearrangement and slippage.4 40 Although such adap-tive changes are known to occur in chronic dilatation,it is not known whether they play a significant role inshort-term experiments such as those conducted in thepreparation ofMR we used. In this study, an equilibra-tion period of 90 min was allowed after the creation ofsevere MR, and during this period progressive ventric-ular dilatation occurred; consequently, systolic wallstress, which was initially decreased, returned to base-line levels. This would suggest adaptation to a newmechanical baseline, as in compensated subacute orchronic MR. 12, 41, 42 Fiber rearrangement, slippage, andother phenomena, such as stress relaxation and creep,can occur over a very short time period,43 and it islikely that such processes occurred in our experiments,although pathologic confirmation is not available.Thus, the rightward shift in the end-systolic pressure-volume relationship may be a result of ventricular dila-tion and remodeling, and not necessarily a change inthe contractile state of the myocardium.We interpret our data as supporting the hypothesis

that augmented fiber shortening in MR does not reflectan increase in left ventricular contractility; this is thecase whether normalized or non-normalized elastanceis used in the analysis. The increased extent and veloc-ity of shortening were due to increased preload; after-load, as assessed by systolic wall stress, had returnedto baseline levels and thus a systolic unloading effectdid not contribute significantly to the observed in-crease in shortening.

Clinical implications. These findings have several im-plications for the clinical use of systolic elastance as ameasure of left ventricular function in the presence ofMR. Although the pressure-volume coordinates can bereasonably fit by a linear function, we observed slight-ly greater scatter after creation of MR. If one were toassume linearity and rely on only two or three pres-sure-volume coordinates to derive systolic elastance,considerable error could be introduced. The need toobtain multiple coordinates to ensure an accurate as-sessment of elastance is thus apparent."4 Previousstudies",45 have also shown that single point pressure-volume ratios are, in fact, load dependent. Incorpora-

1317



ownloaded from


BERKO et al.

tion of wall stress values in the analysis would seemespecially relevant to the assessment of patients withchronic MR in whom myocardial hypertrophy is alsopresent.The observed disparity between the ejection phase

indexes and Emax has clear implications in the clinicalsetting. These data support the concept that ventricularfunction is overestimated by ejection indexes whenMR is present. Numerous clinical studies demonstrate

significant decrease in ejection fraction in patientsundergoing mitral valve replacement for severeMR8 11,42 suggesting that irreversible myocardial dys-function may become unmasked when the regurgitantlesion (favoring enhanced shortening) is no longerpresent. While this is especially common in patientswith decompensated ventricles,14' 15 46 there is evi-dence that a postoperative decline in fiber shortening isnot always due to an increase in afterload.42 In anycase, ejection indexes are poor predictors of myocardi-al function; based on our results and the clinical data ofothers,35 systolic elastance may provide a better defini-tion of left ventricular function in the setting of MR.However, the dependence of this relationship on heartsize requires further elucidation before this relation-ship is applied to the clinical evaluation of left ven-tricular function in patients with MR. Likewise, aconsideration of EDV may be important in aortic re-gurgitation, dilated cardiomyopathy, and even in thecomparison of data from children and adults.

We are very grateful to Jeannette Forte for her excellentsecretarial assistance in the preparation of this manuscript.

References1. Braunwald E, Welch GH, Sarnoff SJ: Hemodynamic effects of

quantitatively varied experimental mitral regurgitation. Circ Res 5:539, 1957

2. Urschel CW, Covell JW, Sonnenblick EH, Ross J Jr, Braunwald E:Myocardial mechanics in aortic and mitral valvular regurgitation:the concept of instantaneous impedance as a determinant of theperformance of the intact heart. J Clin Invest 47: 867, 1968

3. Braunwald E: Mitral regurgitation: physiologic, clinical, and surgi-cal considerations. N Engl J Med 281: 425, 1969

4. Ross J Jr, McCullagh WH: Nature of enhanced performance of thedilated left ventricle in the dog during chronic volume overloading.Circ Res 30: 549, 1972

5. Mahler F, Ross J Jr, O'Rourke RA, Covell JW: Effects of changesin preload, afterload and inotropic state on ejection and isovolumicphase measures of contractility in the conscious dog. Am J Cardiol35: 626, 1975

6. Quinones MA, Gaasch WH, Alexander JK: Influence of acutechanges in preload, afterload, contractile state and heart rate onejection and isovolumic indices of myocardial contractility in man.Circulation 53: 293, 1976

7. Weber KT, Janicki JS, Reeves RC, Hefner LL: Factors influencingleft ventricular shortening in isolated canine heart. Am J Physiol230: 419, 1976

8. Schuler G, Peterson KL, Johnson A, Francis G, Dennish G, UtleyJ, Dailey PO, Ashburn W, Ross J Jr: Temporal response of leftventricular performance to mitral valve surgery. Circulation 59:1218, 1979

9. KennedyJW, Doces JG, Stewart DK: Left ventricular functionbefore and following surgical treatment of mitral valve disease. AmHeartJ 97: 592, 1979

10. Boucher CA, Bingham JB, Oskbakken MD, Okada RD, StraussHW, Block PC, Levine FH, Phillips HR, Pohost GM: Earlychanges in left ventricular size and function after correction of leftventricular volume overload. AmJ Cardiol 47: 991,1981

11. Phillips HR, Levine FH, Carter JE, Boucher CA, Osbakken MD,Okada RD, Akins CW, Daggett WM, Buckley MJ, Pohost GM:Mitral valve replacement for isolated mitral regurgitation: analysisof clinical course and late postoperative left ventricular ejectionfraction. AmJ Cardiol 48: 647, 1981

12. Eckberg DL, Gault JH, Bouchard RL, Karliner JS, Ross J Jr:Mechanics of left ventricular contraction in chronic severe mitralregurgitation. Circulation 47: 1252, 1973

13. RankinJS, Nicholas LM, Kouchoukos NT: Experimental mitralregurgitation: effects on left ventricular function before and afterelimination of chronic regurgitation in the dog. J Thorac Cardio-vasc Surg 70: 478, 1975

14. Spratt JA, Olsen CO, Tyson GS, Glower DD, DavisJW, RankinJS, Sabaston DC: Experimental mitral regurgitation: physiologicaleffects of correction on left ventricular dynamics. J Thorac Cardio-vasc Surg 86: 479, 1983

15. Wong CYH, Spotnitz HM: Systolic and diastolic properties of thehuman left ventricle during valve replacement for chronic mitralregurgitation. Am J Cardiol 47: 40, 1981

16. Suga H, Sagawa K, Shoukas AA: Load independence of the instan-taneous pressure-volume ratio of the canine left ventricle and ef-fects of epinephrine and heart rate on the ratio. Circ Res 32: 314,1973

17. Suga H, Sagawa K: Instantaneous pressure-volume relationshipsand their ratio in the excised, supported canine left ventricle. CircRes 35: 117, 1974

18. Weber KT, Janicki JS, Hefner LL: Left ventricular force-lengthrelations of isovolumic and ejecting contractions. Am J Physiol231: 337, 1976

19. Sagawa K, Suga H, Shoukas AA, Bakalar KM: End-systolic pres-sure/volume ratio: a new index of ventricular contractility. Am JCardiol 40: 748, 1977

20. Weber KT, Janicki JS: Instantaneous force-velocity-length rela-tions: experimental findings and clinical correlates. Am J Cardiol40: 740, 1977

21. Van Trigt P, Bauer BJ, Olsen CO, Rankin JS, Wechsler AS: Animproved transducer for measurement of cardiac dimensions withsonomicrometry. Am J Physiol 240: H664, 1981

22. Rankin JS, McHale PA, Arentzen CE, Ling D, Greenfield JC,Anderson RW: The three-dimensional dynamic geometry of the leftventricle in the conscious dog. Circ Res 39: 304, 1976

23. Rushmer RF: Shrinkage of the heart in anesthetized, thoracoto-mized dogs. Circ Res 2: 22, 1954

24. Karliner JS, Gault JH, Eckberg D, Mullins CB, Ross J Jr: Meanvelocity of fiber shortening. A simplified measure of left ventricu-lar myocardial contractility. Circulation 44: 323, 1971

25. Colan SD, Borow KM, Neumann A: Left ventricular end-systolicwall stress-velocity of fiber shortening relation: a load-independentindex of myocardial contractility. J Am Coll Cardiol 4: 715, 1984

26. Mirsky I: Left ventricular stresses in the intact human heart.Biophys J 9: 189, 1969

27. Kleinbaum DG, Kupper LL: Applied regression analysis and othermultivariate methods. North Scituate, 1978, Duxbury Press, pp97-106

28. Sagawa K: The ventricular pressure-volume diagram revisited.Circ Res 43: 677, 1978

29. Bogen DK, Ariel Y, McMahon TA, Gaasch WH: Measurement ofpeak systolic elastance in intact canine circulation with servo-pump. Am J Physiol 249: H585, l985

30. Sagawa K: The end-systolic pressure-volume relation of the ventri-cle: definition, modifications, and clinical use. Circulation 63:1223, 1981

31. Reichek N, Wilson J, St. John Sutton M, Plappert TA, Goldberg S,Hirshfeld JW: Noninvasive determination of left ventricular end-systolic stress: validation of the method and initial application.Circulation 65: 99, 1982

32. Grossman W, Braunwald E, Mann T, McLaurin LP, Green LH:

CIRCULATION1318



ownloaded from



Contractile state of the left ventricle in man as evaluated from end-systolic pressure-volume relations. Circulation 56: 845, 1977

33. Mehmel HC, Stockins B, Ruffmann K, Olshausen KV, Schuler G,Kubler W: The linearity of the end-systolic pressure-volume rela-tionship in man and its sensitivity for assessment of left ventricularfunction. Circulation 63: 1216, 1981

34. Nivatpumin T, Katz S, Scheuer J: Peak left ventricular systolicpressure/end-systolic volume ratio: a sensitive detector of left ven-tricular disease. Am J Cardiol 43: 969, 1979

35. Carabello BA, Nolan SP, McGuire LB: Assessment of preoper-ative left ventricular function in patients with mitral regurgitation:value of the end-systolic wall stress-end-systolic volume ratio.Circulation 64: 1212, 1981

36. Maughan WL, Sunagawa K, Burkhoff C, Graves WL, Hunter WC,Sagawa K: Effect of heart rate on the canine end-systolic pressure-volume relationship. Circulation 72: 654, 1985

37. Alyono D, Larson EV, Crumbley AJ III, Anderson RW: Is Emaxindependent of heart rate in the conscious animal? Circulation 66(suppl II): II-347, 1982 (abst)

38. Sagawa K, Sunagawa K, Maughan WL: Ventricular end-systolicpressure-volume relations. In Levine HJ, Gaasch WH, editors: Theventricle. Boston, 1985, Martinus Nijhoff Publishing, p 79

39. Tajimi T, Widmann TF, Matsuzaki M, Lee JD, Peterson KL: Leftventricular end-systolic pressure-volume and stress-volume rela-

tions in experimental mitral regurgitation. Circulation 70 (suppl II):II-354, 1984 (abst)

40. Ross J Jr, Sonnenblick EH, Taylor RR, Spotnitz HM, Covell JW:Diastolic geometry and sarcomere lengths in the chronically dilatedcanine left ventricle. Circ Res 28: 49, 1971

41. Wisenbaugh T, Spann JF, Carabello BA: Differences in myocardi-al performance and load between patients with similar amounts ofchronic aortic versus chronic mitral regurgitation. J Am Coll Car-diol 3: 916, 1984

42. Zile MR, Gaasch WH, Levine HJ: Left ventricular stress-dimen-sion-shortening relations before and after correction of chronicaortic and mitral regurgitation. Am J Cardiol 56: 99, 1985

43. Little RC, Wead WB: Diastolic viscoelastic properties of active andquiescent cardiac muscle. Am J Physiol 221: 1120, 1971

44. Hess OM, Maass D, Studer N, Carroll JD, Turina M, KrayenbuehlHP: Left ventricular end-systolic pressure-diameter ratio. Is it use-ful for assessing myocardial contractility? Circulation 66(suppl 11):11-347, 1982 (abst)

45. Wisenbaugh T, Yu G, Evans J: The superiority of maximum fiberelastance over maximum stress-volume ratios as an index of con-tractile state. Circulation 72: 648, 1985

46. Ross J Jr: Afterload mismatch in aortic and mitral valve disease:implications for surgical therapy. J Am Coll Cardiol 5: 811, 1985

Vol. 75, No. 6, June 1987 1319



ownloaded from


B Berko, W H Gaasch, N Tanigawa, D Smith and E Craigemitral regurgitation.

Disparity between ejection and end-systolic indexes of left ventricular contractility in

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1987 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.75.6.1310

1987;75:1310-1319Circulation.

http://circ.ahajournals.org/content/75/6/1310the 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:



ownloaded from

http://circ.ahajournals.org/content/75/6/1310

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

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


disparity ejection and ventricular in -...

Documents