LABORATORY INVESTIGATIONMITRAL REGURGITATION

A Doppler-two-dimensional echocardiographicmethod for quantitation of mitral regurgitationKATHRYN J. ASCAH, M.D., WILLIAM J. STEWART, M.D.,* LENG JIANG, M.D.,** J. Luis GUERRERO,JOHN B. NEWELL, B.S., LINDA D. GILLAM, M.D. AND ARTHUR E. WEYMAN, M.D.

ABSTRACT A noninvasive method to accurately quantitate the severity of mitral insufficiencywould be of major clinical value. In theory, in the absence of confounding variables, regurgitant mitralflow should represent the difference between forward mitral blood flow and aortic blood flow. SinceDoppler-two-dimensional echocardiographic (D2DE) methods for measuring transvalvular mitral andaortic flow have been validated, it should be possible to use mitral and aortic flows derived by thismethod to calculate regurgitant mitral flow. To assess the validity and accuracy of this combinedapproach for quantitation of regurgitant flow, we developed an open-chest canine preparation in whichwe could simulate, vary, and accurately measure degrees of mitral regurgitation. Seven animals were

anesthetized and prepared to allow controlled right heart output. Mitral regurgitation was then simulat-ed by placing a flexible conduit incorporating a one-way valve and electromagnetic flowmeter betweenthe left ventricular apex and left atrium. Flow through the tube (effective mitral regurgitation) was

varied between 0.2 and 1 .8 liters/min and forward cardiac output ranged between 0.5 and 4 liters/min.Transmitral and transaortic flows were calculated by previously reported Doppler methods. Doppler-derived estimates of forward flow through the aortic valve correlated well with the flow measured byflowmeter (r = .92), and regurgitant flow and regurgitant fraction calculated by the D2DE approachalso compared well with those measured by flowmeter (r = .84 and .83, respectively). This studydemonstrates that mitral regurgitant flow and regurgitant fraction calculated by the D2DE methodprovide an acceptable measure of both absolute regurgitant flow and the regurgitant fraction in theexperimental setting.Circulation 72, No. 2, 377-383, 1985.

AN ASSESSMENT of the severity of mitral regurgita-tion is often important in the management of patientswith mitral valvular disease.' The degree of mitralregurgitation may be gauged angiographically with theuse of the degree of left atrial opacification, as suggest-ed by Sellers et al.) Unfortunately, this approach issubject to considerable interobserver variability andthe inherent qualitative nature of the data makes itdifficult to assess changes in the degree of mitral regur-gitation over time. The more rigorous Sandler-Dodgemethod, in which the regurgitant volume is calculatedas the difference between the angiographic left ventric-

From the Cardiac Ultrasound Laboratory, Division of Cardiology,Massachusetts General Hospital, and Harvard Medical School, Boston.

Address for correspondence: Arthur E. Weyman. M.D.. CardiacUltrasound Laboratory, Massachusetts General Hospital, Boston, MA02114.

Received Feb. 22, 1985; revision accepted May 9, 1985.Dr. Ascah is a fellow of the Canadian Heart Association.*Current address: Department of Cardiology. Cleveland Clinic Foun-

dation, Cleveland, OH 44106.**Current address: Department of Cardiology. Shanghai Cardiovas-

cular Institute, Shanghai. People's Republic of China.

Vol. 72, No. 2, August 1985

ular output and forward cardiac output by the thermo-dilution, Fick, or dye-dilution technique is thus usuallyused to provide quantitative information.3 The accura-cy of this approach, however, is limited by the nonsi-multaneous acquisition of the angiographic and for-ward cardiac outputs, the errors inherent in each of thecardiac output measurements, and the presence of as-sociated regurgitant lesions and arrhythmias. Finally,both the Sandler-Dodge and Sellers methods may pro-vide somewhat misleading information in that changesin loading conditions and contractility induced by thecatheterization procedure may lead to underestimationor overestimation of the basal severity of the lesion.

Several noninvasive techniques for assessing mitralregurgitation have also been reported. With M modeand cross-sectional echocardiography, the presence ofregurgitation may be inferred from a large left ventricleand atrium, in association with abnormalities of mitralvalve structure and/or function.4 An increased left-to-right ventricular stroke volume ratio on the radionu-clide angiogram is suggestive of left-sided valvular

377

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

ASCAH et al.

insufficiency. More recently, Doppler echocardio-graphic assessment of the spatial extent of systolicfrequency dispersion within the left atrium has beenshown to provide a semiquantitative measure of theseverity of mitral regurgitation.' 9 Although use ofthese methods has the advantages of being noninva-sive and providing information about function in thebasal state, none provide truly quantitative data. Thus,to date, there is no completely satisfactory means,invasive or noninvasive, of quantitating mitral insuffi-ciency.

Combined Doppler-two-dimensional echocardio-graphic (D2DE) techniques for measuring flows atspecific points within the heart have recently been vali-dated'0- ' and the relationship of Doppler-derived sys-temic and pulmonic flows has been successfully usedto quantitate intracardiac shunting. "-" Since regurgi-tant flow through the mitral valve theoretically equalsthe difference between forward mitral and aorticflows, it should, likewise, be possible to calculate re-gurgitant volume as the difference between the Dop-pler-two-dimensional echocardiographically calculat-ed flows through these valves.

The purpose of this study was to test the validity ofthis combined D2DE approach to the quantitation ofmitral regurgitation with use of an open-chest caninepreparation of mitral insufficiency in which both for-ward cardiac output and mitral regurgitant flow couldbe precisely controlled.

MethodsExperimental preparation. To create a preparation of effec-

tive mitral regurgitation, a preparation of controlled right heartcardiac output previously employed in our laboratory was modi-fied by the addition of a valved conduit between the left ventri-cle and left atrium.'', 19

Seven mongrel dogs weighing 16 to 23 kg were studied. Thedogs were anesthetized with sodium pentobarbital (35 mg/kg)and mechanically ventilated with a pressure-cycled respirator(Bird Corporation, Palm Springs. CA). The chest was openedthrough a midline stemotomy, and the heart suspended in apericardial cradle. The superior and inferior venae cavae werecannulated to collect systemic venous return, and then severedproximal to the site of the cannula. A third cannula was insertedthrough the inferior vena caval stump into the coronary sinusand sutured in place to collect coronary sinus return. All threevenous cannulac emptied into a bubble oxygenator and rollerpump, which had been calibrated with use of a stopwatch andgraduated cylinder. The output of the roller pump was returnedto the right atrium through the stump of the superior vena cava(figure 1). Total forward cardiac output could thus be preciselycontrolled by the roller pump by regulating the amount of oxy-genated blood infused into right atrium. Respiration was thensuspended for the remainder of the experiment. To maintain aconstant heart rate, the sinus node was crushed and atrial pacingwas established with right atrial epicardial electrodes. An elec-tromagnetic flow probe was placed snugly around the ascendingaorta and calibrated to the roller pump. The femoral arteries

378

-- FORWARD FLOW

t'<<--Ar~-L&L '1 t x)PACING

PA (

ELECTRO -MAGNETIC RA-FLOWMETER RA, 4

SUC

1 IUC

K VENOUS RETURN

LA PRESSURECATHETER

CLAMP

ELECTROMAGNETIC LV-LAFLOWMETER SHUNT-WAY VALVE

LV PRESSURECATHETER

\ -TO RA

TO Ao- ROLEPUMPS

FIGURE 1. Diagram of the preparation. Systemic venous return iscollected and directed to a membrane oExygenator. and returned to theright heart at a controlled rate. A second roller pump intuses or with-draws blood from the femoral arteries to alter aortic impedence. Theheart is paced via atrial electrodes. An electromagnetic flowmeter isplaced on the ascending aorta (Ao). A flexible tube. in which a flow-meter and one-way valve are incorporated, is sutured in place betweenthe left ventricle (LV). left atrium (LA). such that blood passes from theLV to the LA simulating mitral insufficiency. RA = rigLht atriumn: RV

right vcntricle.

were cannulated and the cannulae were attached to a secondroller pump that permitted blood to be infused into or withdrawnfrom the aorta. This permitted maintenance of a constant aorticpressure (mean 75 mm Hg) at low flows, as well as variation ofaortic impedence and therefore shunt flow during the experi-mental studies (see below). Left ventricular and central aorticpressure were continuously monitored with fluid-filled polyeth-ylene catheters connected to Statham P-23 PD pressure trans-ducers positioned at the mid-thoracic level.To simulate regurgitant flow, a flexible polyethylene conduit

( /2 inch diameter) into which a one-way valve and electromag-netic flow probe were incorporated was inserted through stabwounds in the left ventricular lateral wall near the apex and theposterior left atrial wall and sutured in place (figure 1). Theconduit valve permitted flow only from the left ventricle to theleft atrium. The shunt flowmeter was calibrated before insertionagainst timed volumes measured by graduated cylinder. Flowthrough the shunt was regulated by variable occlusion of thetubing with an adjustable clamp as well as by varying aorticimpedence throughout the infusion or withdrawal of blood fromthe aorta through the roller pump attached to the femoral arter-ies. A fluid-filled catheter was positioned in the left atriumthrough the shunt to monitor left atrial pressure. Aortic, leftatrial, and left ventricular pressures, a monitoring (modifiedchest lead) electrocardiogram. and aortic and shunt electromag-netically measured flows were recorded on a multichannel re-corder.

Echocardiographic imaging and acquisition of Dopplerdata. All D2DE studies were performed with a combined two-dimensional echocardiographic/range-gated Doppler instru-ment, ATL MK600 (Advanced Technology Laboratories,Bellevue, Washington), equipped with a mechanical transducerusing a 3.0 MHz carrier frequency. Doppler velocity spectra,echocardiographic images. and simultaneous single-lead elec-trocardiograms were recorded on ]/2 inch videotape with VHS

CIRCULATION

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

LABORATORY INVESTIGATION-MITRAL REGURGITATION

format on a Panasonic NV8200 Omnivision 11 recording sys-tern.A saline bath was suspended over the heart to optimize the

intertace between the transducer and the heart. Care was takenthroughout the study not to apply direct pressure to the heart orto distort cardiac anatomy. Complete exposure of the anteriorsurface of the heart allowed the transducer to be positionedappropriately so as to allow optimal structural imaging andDoppler velocity recordings.

Echocardiographic image acquisition. Calculation of car-diac output by the Doppler technique requires knowledge of thecross-sectional area of the vessel or valve through which bloodis flowine and the linear velocity of flow. In this study, theaortic cross-sectional area was calculated froom the vessel diam-eter, assuming a circular configuration. This diameter was mea-sured from a long-axis recording of the aorta in which the scanplane was aligned to include the aortic valve and proximalascending aorta (figure 2).

Calculation of the effective mitral valve flow area was morecomplex due to the normal variation in size of the mitral orificeduring diastole. A mean diastolic mitral valve area was there-fore computed, as illustrated in figure 3, following the methodof Fisher et al.1`1

Doppler velocity recordings. Aortic flow velocities wererecorded with the transducer placed dircctly over the cardiacapex. The scan plane was directed toward the base of the heartand rotated to include the left ventricular outflow tract andascending aorta. The Doppler cursor was positioned as nearlyparallel to flow in the left ventricular outflow tract as possibleand the sample volume was placed superior to the aortic valve.From this starting point, the system was switchcd to the Dopplermode and the beam scanned in a tight radial pattern with slightchanges in the axial depth of sampling until the velocity profileswith the largest Doppler shifts were recorded (see figure 2).

AVA = 92)

Aortic Diameter (D)

Sample Volume PoSition S,Y$soli Velaoitiflegra

ABF AVA x VIR-R (cos8)

FIGURE 2. Aortic valve calculation of cardiac OLutpUt with Doppleiechocardiography. Top left, Measurement of the artIc dialieter at theaortic valve anulus. Top right, Formula used to calculate aiea of theaortic valve. Bottom leift, Positioning of sample volume afor recordine ofaortic velocity integral. Bottom, Formula for calculation of cardliacoutput through the aortic valve.

Maximum Valve Area Mitral Mean-o-Max Rato

Sample Volume Position Diastolic Velocity Integral

(MVA)(Meon -To-Maximum Rotio) (Velocity Integrol)MtITRAL FLOW -------------------------- -- -------> >

(R-R Interval) (Cosine 8)

FIGURE 3. Mitral valve calculation of cardiac output with Dopplerechocardiography. Top left. Measurement of the mitral valve area at thetips of the leaflets. Top right. Measuremiient of the diastolic mitral area,

e e. and d-c lengths from the M minode echocardiogram of the mitralvalve. Bottom left, Positioning of sample volume for recording mitralvelocity integral. Bottomii, Formula for calculation of cardiac outputthrough the mitral valve.

Mitral inflow velocity recordings were obtained from theequivalent of the standard apical four-chamnber or apical two-chamber view. The Doppler cursor was initially aligned parallelto the apparent direction of flow and the sample volume waspositioned at the depth of the mitral anulus in diastole. In Dopp-ler mode, the flow profile with the highest apparent velocitieswas searched for in a manner analogous to that described for theaortic valve (see figure 3). At the end of each flow recording, acareful search was made to exclude native mitral and aorticvalvular insufficiency.

Echocardiographic/Doppler analysis. All measurementswere made with an off-line computer graphics system (EasyView I1, Micerosonics, Indianapolis). Each cchocardiographicand Doppler measurement was obtained in nine different cardi-ac cycles, and the average was used in subsequent analysis.The systolic diameter of the aortic valve was measured at the

point of insertion of the valve leaflets, in the field immediatelyfollowing valve opening, from inner edge to inner edge. Themitral valve area was measured in short axis from the frameshowing maximal diastolic opening by tracing the midpoint ofthe leaflet echoes. The ee' distance, the d-c interval, and thetotal area between the anterior and posterior leaflets duringdiastole were measured from the M mode echocardiogram. Themean leaflet separation (area between the leaflets divided by thed c interval) was divided by the maximal leaflet separation(ee') to give the imean-to-max' ratio. To determine a meandiastolic area of valve orifice, the maximal valve area derivedfrom the two dinmensional study was then multiplied by thisratio.

Doppler measurements were made directly from the velocityspectra with the computer graphics system. The Doppler spectrachosen for analysis were those that demonstrated the greatestvelocities and that had profiles with a narrow frequency band-width (representative of undisturbed "laminar" flow). The area

Vol. 72, No. 2, August 1985 379

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

ASCAH et al.

under the systolic (aortic) or diastolic (mitral) velocity profilewas traced through the modal components of the frequencyspectrum. The modal velocity, which is determined visually asthe brightest component of the spectral display (highest ampli-tude), represents the velocity at which the greatest number ofred cell scatterers are traveling. The enclosed area (velocityintegral), in units of distance (cm), is the integral of the instanta-neous velocities over time and is proportional to stroke volume.The cardiac cycle length (RR interval) was measured directlyfrom the video tracing of the simultaneous electrocardiogram.

Quantitative flow calculations. Forward flow through eachvalve was then calculated from the product of the effectivecross-sectional area and mean linear velocity. Mean linear ve-locity was determined by dividing the velocity integral by theRR interval. Aortic flow was calculated with the formula

Flow = lr(D

2>< SVI

l2 w RR

where D = the diameter of the valve anulus; SVI = the systolicvelocity integral; RR = the RR interval. Mitral valve flow wascalculated from the formula

Flow = MVO, x DmeanDnmax

x DVIRR

where MVO,maX) = the mnaximal area of the mitral valve orifice;(Dmean/Dn,ax) = ratio of mean to maximal M mode valve diame-ter. Since the flows resulting from these calculations were incubic centimeters per second they were converted to liters perminute by multiplying 60 sec/min x 1 liter/l000 cc.

In all velocity recordings, the angle between the Dopplercursor and the apparent direction of flow on the two-dimension-al echocardiographic image was negligible, and no correctionwas made for the effect of angle on the recorded flow velocity.

Doppler-determined regurgitant volume was calculated as thedifference between mitral (DMF) and aortic flow (DAF) asfollows:

DRV DMF - DAF

The Doppler-determined regurgitation fraction (DRF) wascalculated as

DRF = (DMF - DAF)/DMF x 100cI

In cases in which the aortic blood flow was calculated to begreater than the mitral flow by the D2DE method (n = 7). theregurgitant volume and regurgitant fraction were set to zero,since this situation is not physiologically possible.

Analysis of hemodynamic data. Forward flow through theaortic valve and shunt flow were also determined directly withthe aortic and shunt electromagnetic flowmeters. Total mitralvalve flow. MF,,11, was taken as the sum of the shunt flow(SFemnf) and aortic flow (AFc,,-,). Regurgitant fraction was calcu-lated as RF,,, = (SF,1,i-/(SFc-nf + AFCn,,-)) x 100%.

Experimental protocol. To test the ability of the Dopplermethod to record consistently accurate control flows through themitral and aortic valves in this preparation, as well as to assessthe ability of this method to quantitate varying levels of shuntflow, aortic and mitral flows were determined by Doppler echo~cardiography according to the following general protocol: ( I)Control flows were recorded in all animals at roller pump flowsof 1.5 to 2 liters/min, with aortic pressure maintained constantand the shunt closed. (2) With forward flow maintained con-stant, the shunt was opened. (3) With forward flow constant andthe shunt opened, flow into the aorta was increased by thefemoral artery roller pump to increase aortic impedance. Therate of aortic inflow was increased until an obvious increase inshunt flow was noted on the electromagnetic shunt flow record-

ing. (4) The shunt was then closed, aortic inflow was reduced toreturn aortic pressure to the control level, and roller pump out-put was increased to from 2.5 to 4.0 liters/minm and a secondseries of control values was then recorded at the higher forwardflow. (5) At an increased forward flow (between 2.5 and 3.5liters/ min depending on the size of the animal). the shunt wasreopened to permit the highest flow the animal could toleratewithout developing left ventricular failure (visibly apparent leftventricular dilatation with increasing left ventricular end-dia-stolic pressure). Fine adjustments in both forward and shuntflows were then made to stabilize the preparation. (6) Whenpossible, aortic pressure was again increased to increase shuntflow. (7) Finally. to create the highest possible regurgitant frac-tion, forward output was reduced to 0.75 to I liter/min theshunt was fully opened, and aortic impedance was increased.The experiment was then terminated and the animal was killedwith an overdose of pentobarbital. In applying this protocol, itwas necessary to adjust actual forward and shunt flows accord-ing to size of the animal and the ability of the left ventricle totolerate the resulting volume load. This resulted in a wide raneof forward and shunt flows across animals.

Reproducibility of measurements. Interobserver and in-traobserver variations in measurements of mitral and aorticflows by Doppler echocardiography for our laboratory havebeen previously examined and reported.'' The relative correla-tions for two observers with a roller pump were r = .99 and .94(n = 10) for the aortic valve and r = .97 and .96 (n = 10) forthe mitral valve. The experimental design precludes meaningfulassessment of observer variation for the regurgitant volume andregurgitant fraction.

Statistical analysis. Correlations between Doppler-derivedflows and those recorded by electromagnetic flow probe weremade by the least squares method of linear regression. Theexpected error for the derived mitral regurgitant volume wascalculated as V,. V. + V. where V is the expected varn-ance, and VM and Vd are the variance of the mitral and aorticflows, respectively. A first-order approximation was used tocalculate the anticipated mitral regurgitant fraction error.

Results

Forward flows. Aortic and mitral flows were mea-sured by the Doppler method and with the electromag-netic flow probe at 35 experimental stages with andwithout the shunt open. As illustrated in figure 4, anexcellent correlation was observed between the twoflow measurements over a range of forward outputsfrom 0.5 to 4 liters/min (r = .92, y = 1.06x + 0. 16).Doppler-determined mitral valve flow recorded withthe shunt closed at 10 experimental stages showed anequally good correlation with forward cardiac outputdetermined by electromagnetic flowmeter (r = .92,y =1 .05x + 0.08; figure 5). In each case, the slope ofregression was close to unity. The correlation betweenDoppler-determined and electromagnetic mitral flow(aortic plus shunt flow by electromagnetic flow probe)was similar with the shunt open and closed (y 0.88x+ 0.59 vs y = 1.05x + 0.08 respectively, p NS).

Regurgitant flows. Effective mitral regurgitant flowmeasured by the shunt flowmeter ranged from 0.2 to1.8 liters/min. The relationship between directly mea-

CIRCULATION380

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

LABORATORY INVESTIGATION-MITRAL REGURGITATION

2.0 r

y = 1.04x + 0.16p = .ooo

4 r = .92SD = .36

0

.

0S

0

0

S

.-1

14J1

Q.Q.0

0o

0 0

1.5 1

1.0 -

0.5l

By = .97 x + 0.05p=.0001 0 /r = .84

SD = .35

*s00

Oh 2r/- z~~~~~~~'

_

0 0.2 0.4 0.6 0.8 1.0 1.2

EMF MR (L /min)

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

EMF (L/min)

FIGURE 4. Comparison of Doppler-determined aortic cardiac output

with that measured by electromagnetic flowmeter.

sured shunt flow and regurgitant volume calculated bythe Doppler method is illustrated in figure 6. Despitethe relatively small range of shunt flows achievable inthis preparation, a good correlation between the twomethods was still apparent (r = .84, y = 0.97x +0.05). The regurgitant fraction derived from the D2DEdata likewise showed a good correlation with that cal-culated with the flows measured by electromagneticflowmeter (r = .83, y = 0.75x + 5.25).

4

-4i

0)

0

- y = 1.05x +0.08p = .001

3- r = .92SD = .35

0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2

EMF (L/min)

FIGURE 5. Comparison of Doppler-determined mitral cardiac output,in the absence of mitral regurgitation. with aortic cardiac output deter-

mined by electromagnetic flowmeter.

Vol. 72, No. 2. August 1985

FIGURE 6. Regurgitant flow measured by the D2DE method vs that

measured by electromagnetic flowmeter.

Discussion

Doppler echocardiography is uniquely able to non-

invasively measure volumetric flow at multiple loca-tions within the heart and great vessels. In the normalflow state, this capability permits the determination offorward cardiac output from flow data derived fromeach of the four cardiac valves. When flow is disturbedas a result of shunt or valvular regurgitation, quantita-tion of the disturbance should theoretically also bepossible by comparison with flow volumes at serialpoints along the path of normal blood flow through theheart.The accuracy of the Doppler method in defining

forward cardiac output,'0 13 the effects of sampling siteon these measurements,3 and its accuracy in assessingshunt flow'4 '8 have been extensively studied. Lessattention, however, has been given to the use of thistechnique for quantifying regurgitant flow.20We have recently presented preliminary clinical

data suggesting that a significant correlation (p =

.001) exists between the Doppler-derived regurgitantfraction and the hemodynamic/angiographic mitral re-

gurgitant fraction calculated by the method of Sandlerand Dodge.3 The clinical method of Sandler andDodge, however, is based on the difference betweenforward cardiac output as measured by thermodilution,Fick, or dye-dilution techniques and the angiographicstroke volume. Each of these measurements containssignificant inherent sources of error.2' Furthermore,although this method is theoretically attractive, it hasnot to our knowledge been independently validated.Therefore, we believed that to properly interpret andunderstand our clinical data, it was important to definethe accuracy of the Doppler method of calculating re-

381

5

3

2

-.c

k.L4-.0.0

14 1.6 1.8 2.01) 9. 0

j

l

0 '00,

'000, '00, k-X = y

1

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

ASCAH et al.

gurgitant fraction in a more rigidly controlled experi-mental preparation in which forward and regurgitantflows could be precisely measured.Our preparation was a modification of one with con-

trolled right heart output that we had previously used tostudy the accuracy and sources of variability in Dopp-ler flow measurements through the mitral, aortic, andpulmonic valves."1 To simulate mitral regurgitation,the basic preparation was modified by placing a valvedconduit between the left ventricular apex and left atri-um. To exclude the possibility that the addition of theconduit would alter left ventricular or mitral valve ge-ometry in a manner that would adversely affect flowmeasurements, we initially compared the Doppler-de-termined aortic flow and mitral (shunt closed) mea-surements to those obtained with the aortic electro-magnetic flow probe. As noted, the correlationbetween the Doppler- and electromagnetically deter-mined aortic flow was good and similar to that ob-tained in a previous study in which no valved conduitwas used." Likewise, the baseline mitral flows withthe shunt closed correlated well with simultaneouselectromagnetically measured aortic flows. Finally,since flow through the conduit should alter the pattern,volume, and pressure of return to the left atrium, wecompared electromagnetically calculated absolute mi-tral flow (conduit plus aortic) when the shunt was openwith absolute Doppler-determined mitral flow. Again,the correlation was good (r = .82), and the line of bestfit did not differ significantly from that observed withthe shunt closed.

Having validated the accuracy of the basic Dopplermeasurements in this preparation of modified flow, wethen tested the accuracy of the D2DE quantitation ofmitral regurgitant flow, which was calculated as thedifference between total left ventricular inflow andforward cardiac output. In this study a good correlation(r = .84) was obtained between D2DE and electro-magnetic estimates of regurgitant flow and the slope ofthe line of best fit approximated unity (0.97). Thesedata therefore suggest that this method can accuratelymeasure regurgitant volume in the ideal experimentalsetting. The fact that the correlation coefficient be-tween Doppler- and electromagnetically based calcula-tions of regurgitant volume was weaker than that be-tween the "raw" aortic and mitral valve flowsdetermined by the two methods may simply reflect therelatively narrow range of regurgitant volumes (0 to1.8 liters/min) obtainable in this experimental prepara-tion as opposed to the broader range of forward cardiacoutputs (0.5 to 4.0 liter/min). For example, if electro-magnetic and D2DE estimates of aortic flow are com-

382

pared over a range comparable to that in the regurgitantvolume study (i.e., 2 to 4 liters/min), the r value fallsbetween .92 and .78.

While one might expect that the errors in the rawdata would be compounded in the regurgitant volumeand fraction correlations, the variance for both thesecalculated variables is less than predicted from theerrors in the directly measured variables (observed =.35, 11%; predicted = .50, 27%). Thus, it appearsthat some of the inherent error is canceled in the calcu-lation of the regurgitant volume or fraction.

In seven instances, Doppler-derived aortic bloodflow was greater than Doppler-calculated mitral bloodflow by an average of 175 ml. This difference is withinthe error of the method. In all but one case, in whichthe shunt flow was 0.45 liter/min, the shunt was closedwhen this occurred.

"The application to disease states of data derivedfrom acute canine experiments which simulate suchstates, is necessarily limited."'" Despite this, whentesting any new technology, it is important first todefine its capabilities and limitations in an ideal settingwhere it can be tested against established standards ofknown accuracy. Experimental validation of a newtechnique is particularly important in areas in whichthe clinical "sgold standard' has significant inherentsources of error. The information derived from thevalidation studies provides the framework that is nec-essary to interpret the results from clinical studies inwhich few physiologic parameters can be controlled.Thus, our preparation, while admittedly different fromthe human situation in which the mitral valve itself isresponsible for the regurgitant flow, hemodynamicallysimulates mitral insufficiency while allowing regula-tion and precise measurement of both forward andeffective regurgitant flow. In addition, in this prepara-tion, all valves are intrinsically normal and the open-chest setting provides optimum conditions for Dopplerrecording and echocardiographic imaging. This en-sures any that errors due to intrinsic valvular regurgita-tion or difficulty in measurement of valve orifices orvelocity integrals are minimized. Thus, we were ableto validate our method of calculating regurgitant flowand evaluate its accuracy in the absence of confound-ing variables.When applying these results to the clinical environ-

ment, a number of problems can be expected. First, inpatients with mitral regurgitation, the mitral and aorticvalves are frequently deformed, making measurementof valve areas and velocity integrals more difficult.Second, coexisting aortic insufficiency may cause anoverestimation of the true forward flow and may affect

CIRCULATION

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

LABORATORY INVESTIGATION-MITRAL REGURGITATION

the measurement of mitral valve area and velocity inte-gral as a result of impingement of the regurgitant aorticjet on the mitral valve. Third, differences in heart ratemay make measurement more complex. All of ourstudies were performed at a paced heart rate of 120beats/min to maintain a constant cardiac output duringacquisition of the D2DE data. Since our method forcalculation of cardiac output does include the effects ofheart rate, the results for individual measurementsshould not be significantly influenced by the rate.However, large variations in heart rate or changes inrhythm may affect both forward flow and regurgitantvolume, and if there is an interval between the mea-surement of aortic and mitral flows it may affect thecalculation of regurgitant volume.

Finally, there appear to be species-specific differ-ences in the geometry of the mitral valve anulus thathave produced conflicting opinions as to the appropri-ate method for measuring mitral flow. In the dog, themethod of Fisher et al. "' has been extensively validatedand its accuracy has been demonstrated. When appliedto human studies, however, others have found thisapproach less accurate and several alternative methodsfor measuring the effective mitral orifice in the clinicalsetting have been suggested. Although this methodo-logic point is of major importance, it can be assumedthat once the optimal clinical method for measuringmitral valve area is determined, the concepts and meth-ods validated in this study can be easily adapted to thederived clinical formula.

In conclusion, we have described a method for thenoninvasive measurement of regurgitant mitral flow.This method employs combined D2DE techniques tomeasure aortic outflow and mitral inflow and is basedon the theory that the difference between total mitralinflow and forward flow through the aortic valveequals the regurgitant flow. We have demonstratedthat this method can provide accurate quantification ofregurgitant mitral flow in the experimental setting.This technique, being noninvasive, repeatable, and ac-curate has significant advantages over angiographyand may prove to be the method of choice in the clini-cal assessment of mitral regurgitation.

We gratefully acknowledge the secretarial assistance of Kath-leen Lundgren.

References1. Braunwald E: Heart disease: a textbook of cardiovascular medi-

cine. Philadelphia, 1980, W.B. Saunders and Company, p 11 172. Sellers RD. Levy MJ. Amplatz K. Lillehei CW: Left retrograde

cardioangiography in acquired cardiac disease. Am J Cardiol 14:437, 1964

3. SandIer H, Dodge HT, Hay RE, Rackley CE: Quantitation ofvalvular insufficiency in man by angiography. Am Heart J 65: 501,1963

4. Weyman AE: The left ventricular inflow tract, part 1: the mitralvalve. In Weyman AE, editor: Cross-sectional echocardiography.Philadelphia, 1982, Lea and Febiger, pp 139-192

5. Rigo P, Alderson PO, Robertson RM, Becker LC, Wagner HN Jr:Measurement of aortic and mitral regurgitation by gated cardiacblood pool scans. Circulation 60: 306, 1979

6. Patel AK, Rowe GG, Thomsen JH, Dhanani SP, Kosolcharoen P,Lyle LE: Detection and estimation of rheumatic mitral regurgita-tion in the presence of mitral stenosis by pulsed Dopplerechocardiography. Am J Cardiol 51: 986. 1983

7. Abbassi AS, Allen MW, DiCristofaro D, Ungar I: Detection andestimation of the degree of mitral regurgitation by range-gatedpulsed Doppler echocardiography. Circulation 61: 143, 1980

8. Quinones M, Young JB, Waggoner AD, Ostojic MC, RibeiroLGT, Miller RR: Assessment of pulsed Doppler echocardiographyin detection and quantification of aortic and mitral regurgitation. BrHeart J 44: 612, 1980

9. Veyrat C, Ameur A, Bas S, Lessana A, Abitbol G, Kalmanson D:Pulsed Doppler echocardiographic indices for assessing mitral re-gurgitation. Br Heart J 51: 130, 1984

10. Fisher DC, Sahn DJ, Friedman MJ, Larson D, Valdes-Cruz LM,Horowitz S, Goldberg SJ, Allen HD: The mitral valve orificemethod for non-invasive two-dimensional Doppler determinationof cardiac output. Circulation 67: 872, 1983

11. Stewart WJ, Jiang L, Guerrero JL, Weyman AE: Variable effectsof changes in flow rate through the aortic. pulmonic, and mitralvalves on valve area and flow velocity: Impact on quantitativeDoppler flow calculations. J Am Coll Cardiol (in press)

12. Freidman MJ, Sahn DJ, Larson D. Flint A: Two-dimensional echo-range gated Doppler measurements of cardiac output and strokevolume in open-chest dogs. Circulation 62: 101. 1980

13. Fisher DC, Sahn DJ, Freidman MJI Larson D. Valdes-Cruz LM,Horowitz S, Goldberg SJ, Allen HD: The effect of variations onpulsed Doppler sampling site an calculation of cardiac output: Anexperimental study in open chest dogs. Circulation 67: 370, 1983

14. Valdes-Cruz LM, Horiwitz S, Mesel E. Sahn DJ, Fisher DC,Larson D, Goldberg SJ, Allen H: A pulsed Doppler echocardio-graphic method for calculation of pulmonary and systemic flow:accuracy in a canine model with ventricular septal defect. Circula-tion 68: 597, 1983

15. Meijboom EJ, Valdes-Cruz LM, Horowitz S. Sahn DJ, Larson DF,Young KA, Lima CO. Goldberg SJ. Allen HD: A two-dimensionalDoppler echocardiographic method for calculation of pulmonic andsystemic blood flow in a canine model with variable-sized left-to-right extracardiac shunt. Circulation 68: 437. 1983

16. Sanders SP, Yaeger S, Williams RG: Measurement of systemic andpulmonary and QP/QS ratio using Doppler and two-dimensionalechocardiography. Am J Cardiol 51: 952, 1983

17. Goldberg SJ, Sahn DJ, Allen HD. Valdes-Cruz LM. Hoenecke H,Carnahan Y: Evaluation of pulmonary and systemic blood flow bytwo-dimensional Doppler echocardiography using fast Fouriertransform spectral analysis. Am J Cardiol 50: 1398, 1982

18. Barron JV, Sahn DJ. Valdes-Cruz LM, Lima CO, Goldberg SJ,Grenadier E. Allen HD: Clinical utility of two-dimensional Dopp-ler echocardiographic techniques for estimating pulmonary to sys-temic blood flow ratios in children with left to right shunting atrialseptal defect, ventricular septal defect or patent ductus arteriosis. JAm Colt Cardiol 3: 169, 1984

19. Braunwald E. Welch GH Jr. Samoff SJ: Hemodynamic effects ofquantitatively varied experimental mitral regurgitation. Circ Res 5:539, 1957

20. Nichol PM, Boughner DR. Persaud S: Non-invasive assessment ofmitral insufficiency by transcutaneous Doppler ultrasound. Circu-lation 54: 956, 1976

21. Grossman W, Dexter L: Profiles in valvular heart disease. In:Grossman W, editor: Cardiac catheterization and angiography.Philadelphia. 1980. Lea and Febiger, p 315

Vol. 72. No. 2, August 1985 383

by guest on April 17, 2017

http://circ.ahajournals.org/D

ownloaded from

K J Ascah, W J Stewart, L Jiang, J L Guerrero, J B Newell, L D Gillam and A E Weymanregurgitation.

A Doppler-two-dimensional echocardiographic method for quantitation of mitral

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1985 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.72.2.377

1985;72:377-383Circulation. 

http://circ.ahajournals.org/content/72/2/377the 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 17, 2017

http://circ.ahajournals.org/D

ownloaded from