Br Heart J 1980; 44: 612-20

Assessment of pulsed Doppler echocardiographyin detection and quantification of aortic and mitralregurgitation*

MIGUEL A QUINONES, JAMES B YOUNG, ALAN D WAGGONER,MIODRAG C OSTOJIC, LAIR G T RIBEIRO, RICHARD R MILLER

From the Section of Cardiology, Department of Medicine, Baylor College of Medicine andThe Methodist Hospital, Houston, Texas, USA

suMMARY Pulsed Doppler echocardiography was employed to detect disturbed or turbulent flowdiagnostic of aortic or mitral regurgitation. Sensitivity, specificity, diagnostic accuracy, and predictivevalue were assessed by the independent interpretation and comparison of aortic root angiograms (91patients) and left ventriculograms (94 patients) to the time interval histogram display of the pulsedDoppler. Sensitivity of Doppler in detecting mitral regurgitation was 94 per cent, with specificity89 per cent, predictive value 81 per cent, and diagnostic accuracy 90 per cent (32 patients with, 62without regurgitation). In aortic regurgitation, sensitivity was also 94 per cent, specificity 82 per cent,

predictive value 94 per cent, and the diagnostic accuracy was 91 per cent (69 patients with, 22 withoutaortic regurgitation). Additionally, no Doppler evidence of mitral or aortic regurgitation was present in20 normal subjects. The aetiology of left-sided valvular regurgitation varied widely, with prostheticvalvular insufficiency being the cause of mitral and aortic regurgitation in seven and 10 patients, respec-

tively. Sixteen of 17 (94%) paraprosth:tic leaks were correctly identified by pulsed Doppler. In patientswith aortic regurgitation the flow-velocity curve recorded in the ascending aorta frequently showed a

negative (or reversed) diastolic component, the magnitude of which (expressed as percentage negativearea) correlated significantly with angiographic severity of regurgitation. Thus, pulsed Doppler echo-cardiography is a highly accurate and objective non-invasive technique for detecting mitral and aorticregurgitation. In aortic regurgitation, estimation of severity is possible from inspection of the Dopplerascending aortic flow velocity curve.

Pulsed Doppler echocardiography is a new non-invasive technique that has proved useful in thedetection of valvular regurgitation or stenosis'-3 aswell as structural abnormalities such as atrial orventricular septal defects.4 6 The addition of pulsedDoppler studies to routine M-mode echocardio-graphy enhances the likelihood of localising thesource of cardiac murmurs and in certain instancesquantifying the severity of valvular dysfunction. Inthis regard, we have recently shown6 that analysisof the time interval histogram of the pulsed Dop-pler can be of use in discriminating betweenpatients with severe aortic stenosis (aortic valvearea less than 1.0 cm2) and those with less severe* Supported in part by the American Heart Association, TexasAffiliate, research grant, and the National Heart, Lung and BloodVessel Research and Demonstration Center, Baylor College ofMedicine.Received for publication 24 March 1980

stenosis (aortic valve area greater thaa 1.0 cm2).Additionally, we have been able to quantify theamount of tricuspid regurgitation and relate theextent of regurgitation to right ventricular haemo-dynamics.7

Earlier reports concerning the diagnostic ac-curacy of pulsed Doppler echocardiography indetecting left-sided valvular regurgitation havedepended primarily on a subjective interpretationof the audio output from the instrument to identifythe presence of systolic turbulence in the leftatrium, a sign of mitral regurgitation,8 or diastolicturbulence in the left ventricular outflow tract, asign of aortic regurgitation.3 The obvious need fora graphic display of the data has led to the develop-ment of a time interval histogram for displayingthe changes in frequency of the ultrasound.9The objectives of this investigation were to

612

Doppler detection of aortic and mitral regurgitation

determine the specificity, sensitivity, predictivevalues, and diagnostic accuracy of the time intervalhistogram in detecting the mitral and aortic regur-gitation. Additionally, an attempt was made tocorrelate measurements derived from the timeinterval histogram with the angiographic severityof valvular insufficiency.

Materials and methods

TECHNICAL DESCRIPTIONSThe principles of pulsed Doppler echocardio-graphy have been previously discussed by Johnsonet al.' and Baker et al.2 and will be reviewed onlybriefly. The frequency of a reflected sound waveis altered when the reflecting target is in motion.When an ultrasound beam is directed towardsmooth, non-turbulent or laminar blood flow, thereflected sound waves exhibit a fairly uniform shiftin frequency or "Doppler shift". This shift maybe graphically represented as either a positive ornegative deflection depending on direction of flowtoward or away from the transducer. The magni-tude of the shift is related to both velocity of flowand the angle of incidence between flow and sound.When the ultrasound beam is directed at turbulentblood flow, the reflected sound waves demonstratea wide and random shift in frequency because ofthe varying direction and velocity of blood cells inthe turbulent area.

Pulsed Doppler echocardiograms were performedin this study using a commercially available Echo-Doppler unit (advanced Technology Laboratory,Bellevue, Washington) that employs repetitivepulses of sound in the megaHertz range similar totraditional M-mode echocardiography. The advan-tage of pulsed Doppler over continuous waveDoppler is that the Doppler shifts produced bythe velocity of blood flow can be recorded inlocalised intracardiac positions at variable depthsby means of a range-gating system in conjunctionwith a standard M-mode echocardiogram, therebyallowing accurate placement of the sample volumewithin the cardiac chambers. An audio signal isgenerated, which is represented as tonal qualitychanges. Non-turbulent blood flow results in anarrow frequency band width producing a musicalsound in contrast to turbulent flow which generatesa wide band width pattern producing a harshrasping sound.

In addition to the aural presentation, a per-manent display of the Doppler shift is achievedusing a time interval histogram. The time intervalhistogram plots the frequency shifts of the reflectedultrasound using a zero-crossing technique thathas previously been described by Lorch et al.9 Inbrief, a pattern is derived by measuring the timebetween successive zero crossings of audible signalwave forms. A zero crossing is defined as theinstant in time when the Doppler shift signal

A (LSB)R MM

. ...-if..

'l mode

ECG

TIHT

L i M

B(Apex)

mY

C(SSN)*, _ ~~~~~~~~~~~~~~~~~~. ^S..............

AO j~~I......^-'...

Fig. 1 Examples of a normal pulsed Doppler echocardiographic examination with a compressed M-rmode on topand a time interval histogram (TIH) below. The sample volume is placed in the left ventricular outflow tract(LVOT) when searching for aortic regurgitation (panel A) and posterior to the mitral valve (MV) whenassessing mitral regurgitation (panel B). Panel C illustrates a normal ascending aortic flow velocity curve.The transducer location is shown in each panel in parentheses. LSB, left sternal border; SSN, supra-sternal notch;AO, ascending aorta; IVS, interventricular septum; LA, left atrium; RPA, right pulmonary artery;ECG, electrocardiogram.

613

Quinones, Young, Waggoner, Ostojic, Ribeiro, Miller

passes through its zero intensity level. A series ofdots are printed above (positive shift) or below(negative shift) a baseline point. Each dot repre-sents a zero crossing event with the distance awayfrom baseline reflecting the magnitude of theDoppler shift frequency (the greater the frequencyshift, the shorter the zero crossing time interval).Smooth laminar flow produces a narrow cluster ofdots while disturbed or turbulent flow produces awidely dispersed dot pattern. The ultimate displayconsists of the M-mode echo with a superimposedline indicating the sample volume location, anelectrocardiogram, and the time interval histogram(Fig. 1).The technique employed in this study consisted

of placing a 3 0 MHz transducer along the leftsternal border in a fashion similar to a routineM-mode echocardiographic study (Fig. 1 and 2).The mitral and aortic valves as well as the leftventricular outflow tract and left atrium wereidentified with the M-mode echocardiogram.Threshold and gain settings were adjusted with thesample volume placed in the left atrium. Thresholdwas set at the lowest level that allowed a slightbackground noise on the baseline (any furtherreduction in threshold would produce wide disper-sion of the dots giving an appearance of turbulentflow even though no real turbulence would bedetected by the aural analysis). Gain was set at thehighest level that allowed only a slight backgroundnoise to be audible.

Fig. 2 Diagrammatic illustration of the differenttransducer positions and sample volume locations (A-D)used in this study. See text for details. As AO,ascending aorta: other abreviations as in Fig. 1.

Mitral regurgitation was sought from the leftsternal border by placing the sample volume in theleft atrium posterior to the anterior leaflet of themitral valve or posterior to the aortic root whilesearching for systolic turbulence. In all cases thetransducer was also placed at the cardiac apex withthe patient in the left lateral recumbent position,

the mitral valve was visualised, and the samplevolume placed in the left atrium posterior to theanterior leaflet. When searching for turbulence,the operator relied on the audio output from theinstrument while moving the sample volumearound the left atrium making appropriate record-ings when turbulence was heard. In all cases,however, the diagnosis of mitral regurgitation wasultimately based upon an objective analysis of thetime interval histogram where turbulence wastaken as a frequency dispersion greater than 10 mm(absolute measurement of dot distribution aboveand below the zero frequency shift baseline asshown in Fig. 3). When systolic turbulence waspresent, its maximal amplitude was measured inabsolute millimetres from lines drawn outlining itsupper and lower border (Fig. 3B). Measurementswere averaged over the five cardiac cycles displayingwidest frequency dispersion.

Aortic regurgitation was sought from the leftsternal border by placing the sample volume in theleft ventricular outflow tract (Fig. 4) just below theaortic valve, and searching for diastolic turbulence.As in mitral regurgitation, the diagnosis of aorticregurgitation was ultimately made from inspectionof the time interval histogram. Special precautionin patients with left ventricular inflow obstruction(mitral stenosis or prosthetic mitral valves) wasused to avoid confusion between diastolic turbu-lence produced by inflow through the mitral valvewith that caused by aortic regurgitation. Thispotential source of error was prevented by samplingjust below the aortic valve at a point where theM-mode echogram demonstrated the junction ofthe mitral valve anterior leaflet with the posterioraortic root echo. In all cases the transducer wasalso placed at the suprasternal notch with thesample volume in the ascending aorta for recordingof the ascending aortic "flow-velocity" curve.Normally, the analogue of the time interval histo-gram in the ascending aorta produces a positivesmooth "flow-like" curve during systole returningto baseline during diastole with only minimal earlydiastolic negative deflection (Fig. 1C). In aorticregurgitation without significant stenosis the dia-stolic curve frequently becomes negative conse-quent to retrograde flow into the ventricle (Fig.4B). Ii these cases the areas of positive and nega-tive deflections outlined by the analogue curvewere measured by planimetry in five consecutivecardiac cycles and the results averaged to computean index of severity of aortic regurgitation:

% negative area =negative area

total area (negative pluspositive areas)

x l0c

614

Doppler detection of aortic and mitral regurgitation

A(LS1B) B (LSB)

M mode

.. . ........ 5.v .

TIH{

TUSS

615

L (Apex)t_-........... ......e)^s...... ..._.

F-'IM

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... ....

....

~~~~~~~~~~~~~~~~~~~~~~.... ..

*.. ...,..........

?A

Fig. 3 Examples of systolic turbulence (TURB) indicative of mitral regurgitation with sample volume locatedposterior to mitral valve (panels A and C) or in the left atrium (panel B); transducer position at left sternalborder (panels A and B) or cardiac apex (panel C). The maximal amplitude of turbulence (MAT) is measuredas shown in panel B. See text for details. Abbreviations as in Fig. 1.

During a period of six months all patients whohad a pulsed Doppler examination within five daysof cardiac catheterisation were selected for analysis.The results from the Doppler time interval histo-gram were compared with the left ventricular oraortic root angiogram by independent observers.Left ventricular angiography was performed in allpatients in a 300 right anterior oblique positionwith the injection of 40 to 50 ml Renografin-76while the aortic root angiograms were performed

A (LSB)

in a 450 left anterior oblique projection with injec-tion of similar amounts of contrast material. Mitralregurgitation was diagnosed from the left ven-triculogram when contrast material appeared in theleft atrium during systole in the absence of ventricu-lar or atrial ectopy. The severity of mitral regurgi-tation was judged qualitatively as 1 + (rapid clear-ance of dye from the left atrium with each cycle),2+ (slower clearance of dye from left atrium butgreater opacification of ventricle than atrium), or

B (SSN)

M mode

tillJ~~~~~i

_~~~~~L

ECG +

cmfIH|

-Vwt> twso. .- ^o5

! mj c4g uwaA.

S.- N4.....

..

OS R NgATiVEAA A

,,w-,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..

Fig. 4 Panel A shows an example of diastolic turbulence (TURB) in the left ventricular outflow tract indicativeof aortic regurgitation; transducer positioned at left sternal border. Panel B illustrates an ascending aortic flowvelocity curve from the same patient showing a characteristic diastolic negative deflection. Planimetry of thenegative and positive areas allows assessment of severity of aortic regurgitation (see text).Abbreviations as in Fig. 1.

Quinones, Young. Waggoner, Ostojic, Ribeiro, Miller

3+ (left atrium filled with dye and appearingequal to or more dense than ventricle). Aorticregurgitation was diagnosed from the aortic rootangiogram when contrast material appeared in theleft ventricle during diastole; its severity wasjudged qualitatively from 1 + to 4+: 1 +, minimaldye in the ventricle clearing with each systole;2+, slower clearance of dye -from ventricle withaortic opacification greater than ventricular; 3+,rapid opacification of ventricle with slow clearanceand equal density of dye in both ventricle andaorta; 4+, complete opacification of ventriclewithin first two cardiac cycles with greater densityof dye in the ventricle than in the aorta.Four patients were eliminated from analysis,

two for suboptimal echocardiographic studies andtwo for suboptimal quality of the angiograms. Thedecision to eliminate these patients was madewithout knowledge of the results of the angiogramin the first two or the Doppler examination in thesecond two cases. A total of 94 studies was availablefor comparison with angiographic detection ofmitral regurgitation and 91 for comparison withangiographic detection of aortic regurgitation. Inaddition, 20 normal subjects without evidence oforganic heart disease were studied by pulsedDoppler echocardiography to establish specificityfor mitral and aortic regurgitation in a normalpopulation. The sensitivity, specificity, predictivevalue, and accuracy of the pulsed Doppler echo-cardiogram were calculated as follows:

Sensitivitytrue positive

true positive plus false negatives

Specificity

Predictive value

Accuracy

true negativetrue negative plus false positives

true positive (or negative)true positive (or negative) plus

false positive (or negative)

true positives plus true negativesall subjects

where, true positive =Doppler positive, angiopositive; true negative =Doppler negative, angionegative; false negative =Doppler negative, angiopositive; and false positive =Doppler positive,angio negative. The significance of a differencebetween the means of samples was computed usingStudent's t test with significance established at ap < 0 05 level.

Results

MITRAL REGURGITATIONFig. 3 shows examples of mitral regurgitationdetected from either left sternal border or apicalapproaches. Fig. 5 illustrates the sensitivity, speci-ficity, and predictive values of Doppler in detectingmitral regurgitation. Angiographic evidence ofmitral regurgitation was present in 32 out of 94patients. Table 1 lists the aetiologies of mitralregurgitation in these patients. The pulsed Dopplertime interval histogram detected mitral regurgita-tion in 30 of the 32 patients (94% sensitivity).Mitral regurgitation was absent by Doppler in 55out of 62 patients without angiographic evidence of

- PDE+PDE

Fig. 5 Bar diagramillustrating sensitivity, specificity,and predictive values of pulsedDoppler echocardiography (PDE)in mitral regurgitation (MR).Angio, angiography.

0

0:_

oc

3._

.c (

4-

a-

10040v080_

(/)

2 "

XW*3 '40c

2 20a-0

0Sensitivity

C)

0C

Specificity .Acicr-c 90

616

=

Doppler detection of aortic and mitral regurgitation

regurgitation (89% specificity). The predictivevalues for a positive and a negative Dopplerexamination were 85 per cent and 96 per cent,respectively, while the overall diagnostic accuracywas 90 per cent. Both false negative examinationsoccurred in patients with minimal (1+) angio-graphic mitral regurgitation. The seven falsepositive Doppler studies consisted of four patientswith cardiomyopathy, and one each with aorticstenosis and left ventricular failure, mitral valveprolapse, and ventricular septal defect. None of the20 normal subjects had evidence of mitral regurgi-tation by Doppler.

Table 1 Aetiology of mitral regurgitation in 32 patients

Aetiology Angio MR PDE MR

Rheumatic 13 12Prosthetic valve leak 7 6Myxomatous valve 5 5Cardiomyopathy 3 3Papillary muscle dysfunction 3 3Congenital defect 1 1

Total 32 30

Angio, angiographic; MR, mitral regurgitation; PDE, pulsedDoppler echocardiography.

Mitral regurgitation was detected from a singlelocation in 19 patients and from multiple locationsin 11 patients. In patients having regurgitationnoted from a single location, turbulence wasdetected only in the left atrium in two, behind themitral valve with the transducer at the left sternalborder in eight, and behind the mitral valve withthe transducer placed at the apex in nine. Turbu-

- PDE+PDE

la0cmc

n

.C(0

0-a.&I

).0 _

30_Sestrvt

0

0C'

4-

a.6

L0

a..c]0n

.= (D

c._

0-

lence was located in early systole in 10 patients,mid or late systole in six, and was holosystolic in14. No correlation (p > 005) was observed betweenduration or location of turbulence and angiographicseverity of mitral regurgitation.The maximum amplitude of systolic turbulence

was 15 9 +1.0 mm (mean ±standard error), 18&9+19 mm, and 21-8±14mm in patients with 1+,2+, and 3 + mitral regurgitation, respectively.While a tendency for greater amplitude of turbu-lence to occur with increasing severity of mitralregurgitation was suggested, the overlap betweenindividual values was such that only the comparisonof 1+ versus 3+ (15-9 versus 21-8 mm) achievedstatistical significance at the 0 05 level.

AORTIC REGURGITATIONFig. 4A illustrates an example of aortic regurgita-tion detected with the sample volume in the leftventricular outflow tract. Fig. 6 illustrates thesensitivity, specificity, and predictive values ofpulsed Doppler in detecting aortic regurgitation.Angiographic evidence of aortic regurgitation waspresent in 69 out of 91 patients; the time intervalhistogram of the pulsed Doppler was positive in65 out of 69 patients (94% sensitivity). Aorticregurgitation was absent by Doppler in 18 out of22 patients without angiographic evidence ofaortic regurgitation (82% specificity). The predic-tive value for a positive and negative Dopplerexamination was 94 per cent and 82 per cent,respectively; the overall diagnostic accuracy was91 per cent. The different aetiologies of aorticregurgitation are listed in Table 2.

- Angio; + Angio

4

65 0

nc]

CL0-

cL0

i.:4:::: :

:4-...... ..

Fig. 6 Bar diagramillustrating sensitivity, specificity,and predictive values of pulsedDoppler echocardiography (PDE)in aortic insufficiency (AI).Angio, angiography.

Predictive valuesaccuracy = 91°0/

617

Quinones, Young, Waggoner, Ostoiic, Ribeiro, Miller

The four false negative Doppler examinationsoccurred in patients with minimal (1+) angio-graphic aortic regurgitation. The four false positiveDoppler studies consisted of three patients withrheumatic heart disease and evidence of aorticvalve thickening by M-mode echocardiography,and one patient with a membranous ventricularseptal defect, severe pulmonary hypertension, andsignificant right-to-left shunting. None of the 20normal subjects had evidence of aortic regurgitationby Doppler.

Neither the duration nor the amplitude ofdiastolic turbulence in the left ventricular outflowtract correlated with the angiographic severity ofaortic regurgitation. In contrast, the analogue ofthe time interval histogram "flow-velocity" curverecorded in the ascending aorta with the transducerin the suprasternal notch frequently demonstrateda negative diastolic component (Fig. 4B). Themagnitude of negative deflection, measured as thepercentage negative area in 54 of the 69 patientswith aortic regurgitation, was found to relatesignificantly (p <0 05) with the qualitative angio-graphic assessment of severity of regurgitation. Asshown in Fig. 7, the percentage negative area was89 ±2-9 (mean ±SEM) for patients with 1+,23*6±442 for 2+, 35-7±5 for 3+, and 50-246-5for 4+ aortic regurgitation. A 30 per cent or lessnegative area was noted in 26 out of 31 (84%)patients with 2+ or less aortic regurgitation, whilea negative area greater than 30 per cent was presentin 17 out of 23 (74%) patients with 3+ or moreregurgitation. Thus, using a less or greater than30 per cent negative deflection to discriminatebetween mild (<2+) and severe (>3+) aorticregurgitation, the predictive value of the pulsedDoppler was 81 and 77 per cent, respectively, witha diagnostic accuracy of 80 per cent.

Discussion

Pulsed Doppler echocardiography allows assess-ment of intracardiac and aortic flow-velocity

Table 2 Aetiology of aortic regurgitation in 69 patients

Aetiology Angio AR PDE AR

Rheumatic 36 32Calcific 12 12Prosthetic valve leak 10 10Dilated aortic root 6 6Endocarditis 2 2Congenital defect 2 2Myxomatous valve 1 1

Total 69 65

AR, aortic regurgitation; others as in Table 1.

75r-70 F

65

60

55

50

45

40

0/0 35Negative area 30

25

20

15

10

5

0

0

I MeantSEM

*0

: T-

0~~~~~~~~~I:*

0~~~~~~

-I~~~~~

00000 so0

1+ 2+ 3+ 4+(n=17) (n-=14) (n-= 13) (n= 10)

Angiographic severity

Fig. 7 Scatter diagram illustrating the percentagenegative area of the ascending aortic flow velocity curvein patients with aortic regurgitation when comparedwith angiographic severity.

patterns not provided by standard M-mode echo-cardiography. The recent development of a graphicdisplay of the frequency shifts of the reflectedsound waves has made it possible to record andevaluate objectively flow-velocity patterns, eventhough an accurate measurement of flow velocityis still not possible with the commercially availablesystem. Despite this limitation, the results of thisinvestigation demonstrate a high (90% and 91%)accuracy for this technique in detecting both mitraland aortic valvular regurgitation when angiographywas used as the standard for comparison. Thesensitivity for detection of either lesion was 94 percent with all of the false negative cases havingminimal regurgitation at angiography. The speci-ficity for mitral regurgitation was 89 per cent, withsix of the seven false positive patients having eitherleft ventricular dilatation with failure (five patients)or mitral valve prolapse (one patient), conditionsin which mitral regurgitation may be presentinconsistently.

618

Doppler detection of aortic and mitral regurgitation

The specificity for aortic regurgitation was 82per cent. Three of the four false positive cases hadsignificant aortic valve thickening by echocardio-graphy, with varying degrees of stenosis observedduring cardiac catheterisation. The fourth falsepositive case of aortic regurgitation had a mem-branous ventricular septal defect with pulmonaryhypertension and a significant right-to-left shunt;presumably the diastolic turbulence detected in theleft ventricular outflow tract originated fromdiastolic right-to-left shunting through the septaldefect since the diastolic filling pressures in theright ventricle exceeded that in the left (35 versus8 mmHg). None of the 20 normal subjects testedhad Doppler evidence of mitral regurgitation oraortic regurgitation. If these subjects are con-sidered as true negatives, the specificity of thetechnique for mitral and aortic regurgitationbecomes 91 and 90 per cent, with an accuracy of92 and 93 per cent, respectively. These results arecomparable to previous reports on detection ofmitral and aortic regurgitation by pulsed Dopplerechocardiography.3 8 The previous investigations,however, were solely dependent on the subjectiveinterpretation of the audio output from the instru-ment by the technician performing the study whilethe present study relied on an objective interpreta-tion of the time interval histogram.Although the diagnosis of mitral or aortic regur-

gitation in this investigation was made from anobjective inspection of the time interval histogram,the technical skills of the operator remain animportant factor in obtaining an optimal study.Detection of mitral regurgitation is a clear exampleof this. Mitral regurgitation was detected from asingle location in 19 out of 30 patients, the leftsternal border approach being no more sensitivethan the apex (eight of 19 versus nine of 19); inthe remaining 11 patients mitral regurgitation wasdetected fiom multiple locations. Thus, accuraterecognition of left-sided valvular regurgitation atthe time of the procedure requires use of a stand-ardised search procedure with placement of thesample volume visually guided by the compressedM-mode and aurally directed by the valve and flowsounds indigenous to the various chambers. Inaddition, appropriate recordings must be super-vised carefully by the technicians performing thestudy to ensure accurate reproduction of the auralfindings.A potential limitation of the pulsed Doppler

technique may occur when clinical situationsdemand examination with the sample volumelocated a long distance from the transducer. Thisproblem may be manifest when mitral regurgitationis sought from the cardiac apex. As Fig. 2 indicates,

the sample volume is near the distal limits of therange-gate when placed in the left atrium orimmediately posterior to the mitral valve with thetransducer at the apex. At these deeper sites withinthe heart, the pulse repetition frequency (range3500 to 10 000 pulses per second) must be dimini-shed to avoid ambiguous echoes, the net result beingan increase in the minimum Doppler shift that canbe detected. Theoretically, this might decrease thesensitivity for detection of mitral regurgitationsince milder cases may not create sufficiently largeDoppler shifts to be recognised. Despite thispotential limitation, the sensitivity of pulsedDoppler for mitral regurgitation in this study was94 per cent and, interestingly, this lesion wasdetected solely through the apical approach in ninepatients.The aetiological causes of mitral and aortic

regurgitation in the present study were multipleand included the entire spectrum of valvular heartdisease seen in adult cardiology. Both structurallesions (such as rheumatic, myxomatous, andcalcific disease) as well as functional disturbances(such as left ventricular dilatation with papillarymuscle dysfunction or ascending aortic aneurysmswith loss of aortic valve support) were frequentlyobserved. The accuracy of pulsed Doppler indetecting regurgitation was similar in each of theseentities. Importantly, 17 patients had prostheticvalves with paravalvular leaks, including sevenmitral and 10 aortic prosthetic valves. Of these 17patients, 16 or 94 per cent were accurately detectedby pulsed Doppler when compared with angio-graphy; the one false negative consisted of a patientwith a mitral prosthesis and minimal regurgitation.Accurate assessment of patients with valvularprostheses represents an important contribution ofpulsed Doppler echocardiography to the clinicalevaluation of these patients since paraprostheticleaks are frequently difficult to confirm by eitherphysical examination or by standard M-modeechocardiography.

Optimal determination of the severity of valvularregurgitation by Doppler analysis requires quanti-fication of regurgitant flow. The present instru-mentation, however, does not afford precisemeasurement of flow or velocity of flow. Neverthe-less, if the angle between sound and blood flowremains constant, changes in the area under the"flow-velocity" curve should reflect changes inflow. Using this basic assumption, a comparisonwas made between the degree of reverse, or negative,diastolic flow in the ascending aorta (expressed asper cent negative area) and a qualitative angio-graphic assessment of the severity of aortic regurgi-tation. Using 30 per cent negative area as the

619

Quinones, Young, Waggoner, Ostojic, Ribeiro, Miller

separation point, this simple measurement yieldedan 81 per cent and 77 per cent predictive value fordifferentiating mild (<2+) versus severe (>3+)aortic regurgitation, respectively (Fig. 6). Angio-graphic measurements of regurgitant fractions werenot made in these patients. It is conceivable thatsuch a precise assessment of degree of regurgitationmight even relate to the percentage negative areabg. Doppler better than the qualitative assessment.

Determination of severity of aortic regurgitationechocardiographically has relied primarily onevaluating the degree of the left ventricular dilata-tion and the "volume overload" pattern of wallmotion, both of which are haemodynamic conse-quences rather than a true measurement of regurgi-tation; both are also dependent upon obtaining ahigh quality echocardiographic view of the leftventricle. Our experience has been that a pulsedDoppler ascending aorta flow-velocity curve can berecorded easily in the vast majority of patientsregardless of the difficulty in obtaining an adequateimage by standard M-mode techniques. Theresults presented herein suggest that analysis ofthis curve provides a direct index of aortic regurgi-tation.A rough correlation was observed in the present

study between the qualitative angiographic assess-ment of severity of mitral regurgitation and theachieved maximal amplitude of turbulence detectedbe Doppler. Statistical significance (p < 0 05),however, was achieved only for the comparisonbetween 1 + and 3+ mitral regurgitation, andthus the application of this Doppler measurementto individual patients with mitral regurgitation isvery limited.

Future improvements in the instrumentationand signal processing of the data will furtherenhance the ability of this new non-invasive tech-nique to detect and quantify valvular regurgitation.The results of this investigation, however, indicatethat in its present state, pulsed Doppler echo-

cardiography has become a useful clinical adjunctto the non-invasive evaluation of patients suspectedof having left-sided valvular regurgitation.

References

1 Johnson SL, Baker DW, Lute RA, Dodge HT.Doppler echocardiography; the localisation of cardiacmurmurs. Circulation 1973; 48: 810-22.

2 Baker DW, Rubenstein SA, Lorch GS. PulsedDoppler echocardiography: principles and applica-tions. Am J Med 1977; 63: 69-80.

3 Ward JM, Baker DW, Rubenstein SA, Johnson SL.Detection of aortic insufficiency by pulse Dopplerechocardiography. JCU 1977; 5: 5-10.

4 Goldberg SJ, Areias JC, Spitaels SEC, de VilleneuveVH. Use of time interval histographic output fromecho-Doppler to detect left-to-right atrial shunts.Circulation 1978; 58: 147-52.

5 Stevenson JG, Kawabori I, Dooley T, GuntherothWG. Diagnosis of ventricular septal defect by pulsedDoppler echocardiography: sensitivity, specificity andlimitations. Circulation 1978; 58: 322-6.

6 Young JB, Quinones MA, Waggoner AD, Miller RR.Diagnosis and quantification of aortic stenosis bypulsed Doppler echocardiography (abstract). Circula-tion 1978; 58, suppl II: 42.

7 Waggoner AD, Quinones MA, Verani MS, MillerRR. Pulsed Doppler echocardiographic detection oftricuspid insufficiency: diagnostic sensitivity andcorrelation with right ventricular hemodynamics(abstract). Circulation 1978; 58, suppl II: 41.

8 Johnson SL, Baker DW, Lute RA, Murrary JA.Detection of mitral regurgitation by Doppler echo-cardiography (abstract). Am Jf Cardiol 1974; 33: 146.

9 Lorch G, Rubenstein S, Baker D, Dooley T, DodgeH. Doppler echocardiography: use of a graphicaldisplay system. Circulation 1977; 56: 576-85.

Requests for reprints to Dr Miguel A Quinones,Section of Cardiology, Department of Medicine,Baylor College of Medicine, The Methodist Hos-pital MS No F1001, 6565 Fannin Street, Houston,Texas 77030, USA.

620