Assessment of Aortic Stenosis with Special Reference to Doppler Ultrasound

ASSESSMENT OF AORTIC STENOSIS WITH SPECIAL REFERENCE TO DOPPLER ULTRASOUND

Akademisk avhandling som med vederbörligt tills tå n d av Rektorsäm betet vid Umeå u n iv e rs ite t fö r avläggande av medicine doktorsexamen kommer att offentligen försvaras i Rosa Salen, Tandiäkarhögskolan, Umeå regionsjukhus, fredagen den 19 december 1986, kl 09.15.

DAG TEIEN

UMEÅ 1986 UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New Series No 177 - ISSN 0346-6612

Assessment of aortic stenosis with special reference to Doppler ultrasound

Dag Teien, Department of C lin ic a l Physiology, U niversity of Umeå, S-901 85 Umeå, Sweden

ABSTRACT

The severity of le ft ventricular outflow obstruction, the le ft ven tricular function and the occurrence of concomitant coronary artery disease are three important issues in the evaluation of patients with suspected aortic stenosis. The systolic peak pressure gradient obtained by continuous wave Doppler echocardiography was compared with the peak to peak pres sure gradient measured at cardiac catheterization in 58 consecutive patients. The two measurements agreed favourably; r = 0.85, SEE = 14 mm Hg. The Doppler peak gradient was compared with the peak gradient mea sured at catheterization by simultaneous recordings in the left ventricle and ascending aorta in 26 patients. A close correlation was obtained; r = 0.94, SEE = 9.1 mm Hg. Estimation of the mean systolic pressure drop across the valve from Doppler tracings of the same patients, using a simple formula APmean = 0*64 x Appeak, was compared with the invasive mean pressure gradient obtained by planimetry. The mean gradient was accurately estimated, r = 0.91, SEE = 7.6 mm Hg. A o rtic valve area measurements were obtained in 22 patients both by application of a modified Gorlin formula and by application of the continuity equation using the Doppler derived systolic pressure gradient and velocity integral as well as the stroke volume ob tained by radionuclide angiography. The noninvasi ve valve area was compared with that obtained by invasive registrations and an excel lent relationship was found; r = 0.95, SEE = 0.13 cm2 by the Gorlin formula and r = 0.94, SEE = 0.14 cm2 by the integration method. Exercise ECG was used to estimate the occurrence of coexistent coronary heart disease in 35 patients. Patients with significant coronary lesions documented by angiography showed a lower mean working capacity, larger ST depressions and higher effort angina score during the exercise test than patients with no or only insig nificant coronary obstructions. Application of simultaneous ECG, phonocardiography and Doppler echocardiography were used to determine the interval between the QRS complex and the mitral valve opening (Q-MC) and the interval between the aortic valve closure and mitral valve opening (A2-MO). The ra tio Q-MC/A 2-MO correlated closely with the pul monary capillary wedge pressure measured at catheterization; r = 0.94, SEE = 2.9 mm Hg, n = 20. A reduction of invasive investigations in the evaluation of pa tients with aortic stenosis is suggested.

Key words: Aortic stenosis, aortic valve area, transvaivular pres sure drop, coronary heart disease, pulmonary c a p illa ry wedge pres sure, Doppler echocardiography, radionuclide angiography UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New Series No 177 - ISSN 0346-6612

From The Department of C lin ic a l Physiology U niversity of Umeå, Umeå, Sweden

ASSESSMENT OF AORTIC STENOSIS WITH SPECIAL REFERENCE TO DOPPLER ULTRASOUND

DAG TEI EN

UMEÅ 1986

To Anne-Britt Mona, Eva, K ris tin

CONTENTS

ABBREVIATIONS ...... 6

ABSTRACT ...... 7

ORIGINAL PAPERS ...... , ...... 8

INTRODUCTION ...... 9

Evaluation of patients with suspected aortic stenosis ...... 9 The Doppler principle ...... 10 The Bernoulli equation ...... 11

AIMS OF THE STUDY ...... 12

PATIENTS (Studies I-V) ...... 13

METHODS ...... 14

Cardiac catheterization ...... 14 Cineangiography ...... 14 A ortic valve area ...... 15 Coronary angiography ...... 15 Estimation of pulmonary c a p illa ry wedge pressure ...... 17 Radionuclide angiography ...... 17 Noninvasive calculation of aortic valve area ...... 18

RESULTS AND DISCUSSION ...... 19

Quantification of the transvalvular pressure differences (Studies I and II) ...... 19 The mean gradient (Study II) ...... 20 Determination of a o rtic valve area (Study I I I ) ...... 21 The coexistence of coronary heart disease (CHD) (Study IV) .... 23 Estimation of pulmonary c a p illa ry wedge pressure in a o rtic stenosis (Study V) ...... 23

FUTURE DEVELOPMENT ...... 25

GENERAL SUMMARY AND CONCLUSIONS ...... 27

ACKNOWLEDGEMENTS ...... 28

REFERENCES ...... 29

STUDY I - V

5 ABBREVIATIONS

CHD = Coronary heart disease ET = Ejection time PCWP = Pulmonary c a p illa ry wedge pressure SV = Stroke volume SVI = Systolic velocity integral Vmax = Maximum ve lo city Wmax = Maximum working capacity WmaxN = Predicted normal Wmax Wmax% = Wmax • 100/WmaxN

6 ABSTRACT

The severity of le ft ventricular outflow obstruction, the le ft ventric ular function and the occurrence of concomitant coronary artery disease are three important issues in the evaluation of patients with suspected aortic stenosis. The systolic peak pressure gradient obtained by continuous wave Dop pler echocardiography was compared with the peak to peak pressure gra dient measured at cardiac catheterization in 58 consecutive patients. The two measurements agreed favourably; r = 0.85, SEE = 14 mm Hg. The Doppler peak gradient was compared with the peak gradient measured at catheterization by simultaneous recordings in the le ft ventricle and ascending aorta in 26 patients. A close correlation was obtained; r = 0.94, SEE = 9.1 mm Hg. Estimation of the mean systolic pressure drop across the valve from Doppler tracings of the same patients, using a simple formula APmean = 0*64 x Appeak» was compared with the invasive mean pressure gradient obtained by planimetry. The mean gradient was accur ately estimated, r = 0.91, SEE = 7.6 mm Hg. A o rtic valve area measurements were obtained in 22 patients both by application of a modified Gorlin formula and by application of the con tinuity equation using the Doppler derived systolic pressure gradient and velocity integral as well as the stroke volume obtained by radio nuclide angiography. The noninvasi ve valve area was compared with that obtained by invasive registrations and an excellent relationship was found; r = 0.95, SEE = 0.13 cm2 by the Gorlin formula and r = 0.94, SEE = 0.14 cm2 by the integration method. Exercise ECG was used to estimate the occurrence of coexistent coro nary heart disease in 35 patients. Patients with significant coronary lesions documented by angiography showed a lower mean working capacity, larger ST depressions and higher effort angina score during the exer cise test than patients with no or only insignificant coronary obstruc tio n s. Application of simultaneous ECG, phonocardiography and Doppler echocar diography were used to determine the interval between the QRS complex and the mitral valve opening (Q-MC) and the interval between the aortic valve closure and mitral valve opening (A 2-MO). The ra tio Q-MC/A 2- MO correlated closely with the pulmonary capillary wedge pressure mea sured at catheterization; r = 0.94, SEE = 2.9 mm Hg, n = 20. A reduction of invasive investigations in the evaluation of patients with aortic stenosis is suggested.

Key words: Aortic stenosis, aortic valve area, transvaivular pressure drop, coronary heart disease, pulmonary c a p illa ry wedge pressure, Doppler echocardiography, radionuclide angiography

7 ORIGINAL PAPERS

The present study is based on the following papers, which w ill be re ferred to by their Roman numerals.

I Teien D, Eriksson P. Quantification of transvaivular pressure d if ferences in aortic stenosis by Doppler ultrasound. Int J Cardiol 1985;7:121-6.

II Teien D, Karp K, Eriksson P. Noninvasi ve estimation of the mean pressure difference in aortic stenosis by Doppler ultrasound. Br Heart J 1986 (in press).

I I I Teien D, Karp K, Eriksson P, B jerle P, Österman G. Noninvasi ve de termination of the valvular area in aortic valve disease by Doppler echocardiography and radionuclide angiography. Manuscript

IV Linderholm H, Österman G, Teien D. Detection of coronary artery disease by means of exercise ECG in patients with aortic stenosis. Acta Med Scand 1985;218:181-8.

V Teien D, Karp K, Eriksson P. Estimation of pulmonary c a p illa ry wedge pressure in aortic stenosis by combined ECG, phonocardio graphy and Doppler echocardiography. J Cardiovasc U1trasonography (in press).

8 INTRODUCTION

Evaluation of patients with suspected aortic stenosis Several factors have to be taken into consideration in the evaluation of patients with aortic stenosis, particularly the severity of outflow obstruction, le ft ventricular performance and concomitant coronary heart disease. An accurate assessment is crucial since surgery may be life-saving.

The systolic pressure gradient across the valve is a measure of ob struction to flow . The s y s to lic pressure gradient is , however, also dependent on le f t ve n tricu la r stroke volume, i.e . the s y s to lic flow across the valve. Patients with a compromised le ft ventricular function may show a low gradient despite a small valve area. The a o rtic valve area is thus an important entity and it should be calculated in all patients in whom there is any doubt about the severity of the lesion. L e ft ve n tricu la r performance may be assessed by measuring the ejection fra c tio n , stroke volume or cardiac output, and the pulmonary c a p illa ry wedge pressure.

The occurrence of concomitant significant coronary heart disease is an other major issue. There are conflicting reports regarding the preva lence of coronary heart disease in patients with aortic stenosis and the significance of coronary atheromatosis in this setting is disputed (1-11). Angina pectoris is a frequent complaint in patients with aortic stenosis both with and without angiographic evidence of obstructive coronary artery lesions. On the other hand, a negative history of ef fort related chest pain does not rule out significant coronary obstruc tions even if this combination is rare (1-11).

A noninvasi ve procedure to screen patients with suspected aortic valve lesions in respect to the severity of obstruction, left ventricular function and concomitant coronary disease is thus desirable. In addi tion, the accuracy of the noninvasive procedures should allow the use of invasive methods to be minimized. The advent of Doppler ultrasound is an important step in this direction (12).

9 The Doppler principle The Doppler effect was described in the firs t half of the 19th century by C hristian Johann Doppler (1803-1853). I t applies to a ll wave-forms including sound. Several illustrations have been used to exemplify this phenomenon. One example is a row boat; when the boat is stationary a certain frequency of the sea waves is fe lt in the boat. If the boat is rowed against the waves a sensation of increased wave frequency is ex perienced, while the opposite is true when the boat is rowed in the other direction. The difference in frequency fe lt when the boat is moving and not moving is the Doppler s h ift or the Doppler frequency. The Doppler frequency can be used to calculate the velocity of the boat provided that the frequency of the waves is known.

In Doppler echocardiography a transducer emits ultrasound with a known frequency (usually 1-10 MHz). The moving blood cells act as reflectors and the velocity of the blood corpuscules can be calculated from the Doppler equation,

c • A f v = ------2 • f • cosa

v = velocity of the blood c = velocity of sound in tissues A f = the frequency shift of reflected ultrasound f = ultrasound frequency c*= angle of incidence between the ultrasonic beam and velocity vectors of the moving blood

A small angle of incidence between the blood flow and the ultrasonic beam is critical to obtain accurate estimation of the flow velocity. In aortic stenosis as well as in other valvular lesions, it is usually possible to obtain a transducer position with an angle less than 15°, which gives a cosine of more than 0.95. I t is , however, of the utmost importance to try many different transducer positions to minimize the angle.

10 The Bernoulli equation This equation is essential in Doppler echocardiography because it de notes the relation between the difference in blood velocity and the pressure difference between the two sides of an obstructive lesion (13).

p - p2 = 1/2 p (V 22 - v i2) +p2 \ i l d r+ R (v) dt flow ac- viscious cel erati on fri cti on

S u ffix 1 denotes the position p rio r to the obstruction and s u ffix 2 post the obstruction.

Pl = pressure prior to obstruction

P2 = pressure post to obstruction P = mass density of blood per unit volume vi = velocity prior to obstruction

V2 = velocity in the jet

The second term on the right side of the equation represents the pres sure drop caused by changes of the velocity with time. The last term represent viscous losses. It has been shown (12-13) that the two last paragraphs can be omitted when the equation is applied to the pressures and dimensions found in stenotic valve disease. Since the velocity prior to a stenosis is small compared with the velocity in the jet it can also be neglected in most cases with obstruction localized to the valvular plane. The equation can then be written,

Ap = 4 (Vmax)2

Ap = pressure drop across the stenosis Vmax = maximal ve lo city in the je t

The validity of the simplified equation has been tested by several authors (12-17) and Hatle and Angelsen (18) have presented a detailed discussion on this subject in their excellent textbook.

11 AIMS OF THE STUDY

To evaluate the applicability of Doppler ultrasound for quantification of pressure difference across the valve in patients with suspected aortic stenosis.

To evaluate a simple method to estimate the mean pressure difference in aortic stenosis from Doppler tracings.

To noninvasively estimate the aortic valve area by combining Doppler echocardiography and radionuclide angiography in patients with aortic valve disease.

To evaluate the usefulness of the standard exercise test to identify coexistant coronary heart disease in patients with aortic stenosis.

To noninvasi vely estimate the pulmonary c a p illa ry wedge pressure in pa tients with aortic stenosis.

12 PATIENTS

Study I In this study 58 consecutive patients were included. They were referred for evaluation of suspected aortic stenosis. Nineteen patients had sig nificant aortic regurgitation and three of them also had mitral regur g ita tio n according to angiography. No patient was excluded on grounds of imperfect Doppler recordings.

Studies II, III Twenty-six consecutive patients examined for suspected aortic stenosis were included. There were 15 patients with pure a o rtic stenosis or only minimal a o rtic re g u rg ita tio n , 8 had stenosis with moderate or moderate ly severe regurgitation and 3 had severe reg u rg ita tio n . Of the 26 pa tie n ts , 14 had m itral re g u rg ita tio n , which was moderate in 2 and m ini mal in 12. No patient included in Studies II and III was recruited from Study I . Four patients were excluded in Study I I I as they did not fu lly complete the invasive protocol of the study.

Study IV In this study, 35 patients were included, mainly those participating in Study I. However, patients with aortic or mitral regurgitation were ex- cl uded.

Study V Twenty consecutive patients were included, 18 of them from Studies II and I I I .

13 METHODS

Cardiac cat heterization In all studies the pressure curves were recorded by liquid fille d cath eters connected to mechanoelectrical transducers interfaced with an UY recorder and calibrated against a hydrostatic standard. Cardiac output was measured by the d ire c t Fick method. No premedication was given. The following systolic pressure differences were calculated: a) the peak to peak gradient which is the difference between the peak systolic pressure in the le ft ventricle and the peak systolic pressure in the aorta, b) the peak pressure difference which is the largest instantaneous sys tolic pressure difference, c) the mean gradient which is the difference between the mean systolic pressure in the le ft ventricle and the mean systolic pressure in the ascending aorta. It was calculated by integrating the area between the le ft ventricle and the aortic pressure recordings (19).

A single catheter was used in Study I and this catheter was positioned in the le ft ventricle and subsequently a pull-back procedure was per formed. Two catheters were used in Studies II and III, one catheter was positioned with the tip in the left ventricle and the other in the ascending aorta.

In Study V, the wedge position of the catheter was confirmed by typical phasic wave form and fluoroscopic determination of the site of the catheter tip . The mean pulmonary capillary wedge pressure was obtained by electronic damping of the signal.

Cineangiography Left ventriculography and aortography were performed in the 45° right and 45° left anterior oblique projections. Left ventricular end-dia- s to lic and end-systolic volumes were calculated using a computer system with a biplane e llip s o id method, or in cases with a deformed ventricu la r ca vity, a biplane integration method. The stroke volume was calcu lated as the difference between them (20). The severity of aortic and

14 m itral regurgitatio n was assessed as described by Dexter and Grossman (19). Left ventricular motion score (LVMS) was calculated visually according to the motion of different wall segments (21). The le ft ven tricle was thereby divided into seven segments according to the Ameri can Heart Association (22). The motion o f each segment was graded as follows: normal motion = 0, hypokinesia = 1, akinesia = 2, dyskinesia = 3 points. By adding these points LVMS was obtained.

A o rtic valve area The aortic valve area was calculated according to Gorlin and Gorlin (23):

A V A = ^ ______. ET • 44.5 VApmean

AVA = a o rtic valve area SV = stroke volume ET = ejection time Ap mean = mean systolic gradient

Both the Fick (or net) and the cineangiographic (or total) stroke volume were used in the calculations of the valve area.

Coronary angiography Coronary angiography was carried out according to Judkins' technique (24). Obstruction of a coronary vessel re su ltin g in a decrease o f the transsectional area by more than 75% was regarded as a significant stenosis. Calculation of myocardial coronary obstruction score is a method employed to estimate the degree o f a rte ria l obstruction, and i t includes an evaluation of the myocardial mass subserved by the stenotic vessel (21). The maximal score is 15 which indicates occlusion of all coronary arteries.

Routine ECG. A 12-lead ECG (aVL, I , -aVR, I I aVF, I I I , and V ^g) was recorded using a d ire c t w ritin g ECG-recorder (Mingograph, Siemens Elema).

15 Phonocardiography. Phonocardiography was recorded together with the ECG and Doppler recording. A 500-1000 Hz f i l t e r was used.

Exercise te st. A standard exercise ECG test was performed (25). Maximum working capacity (Wmax) according to Strandell (26) was calculated and compared with the predicted normal maximum working capacity (WmaxN) (Linderholm, unpublished observations). Wmax as a percentage of WmaxN has been used as an index of physical performance (Wmax%). ST depres sion during exercise was measured as the difference between ST levels during maximum work and at re st. The ra tio 100 x STdepression/ Wmax% was used as an index of coronary insufficiency (CI-index). Effort angina during the exercise test was recorded and an effort angina score was defined. 0 = No angina during or after the exercise test, 1 = Typi cal or atypical angina during exercise but no angina after exercise, 2 = Typical or atypical angina during exercise which disappears within 5 minutes after exercise, 3 = Typical or atypical angina during exer cise which persists for more than 5 minutes after exercise.

Doppi er echocardi ography. Continuous wave Doppler examinations were performed using a Pedof system with a 2.0 MHz nonimaging transducer (Vingmed A/S). Flow velocities up to 6 m/s corresponding to a pressure difference of approximately 140 mm Hg can be measured with this equip ment. Spectral frequency analysis was not available for these studies.

In Studies I and IV the investigations were performed prior to the in vasive investigation, usually one day before. In Studies II, III and V the examinations were carried out w ith in one hour of the subsequent in vasive procedure. Some patients have been examined simultaneously by noninvasi ve and invasive methods.

The maximal flow velocity was recognized by listening to an audio signal of alternating pitch from the loudspeaker in the apparatus (the higher the pitch, the higher the velocity) and observing recorded vel ocity profiles.

Doppler gradients. The simplified Bernoulli formula was used to calcu late the peak pressure gradient from the recorded peak velocity in Stu dies I-Y (13). 16 Estimation of pulmonary c a p illa ry wedge pressure Combined ECG, phonocardiogram and Doppler amplitude recordings were

applied to record the Q-MC and A 2-MO in te rv a ls . The Q-MC was measured as the interval between the onset of the QRS complex and the top of the

amplitude signal of the mitral valve closure. The A 2-MO interval was measured as the in te rva l between the onset of the second heart sound and the beginning of the amplitude signal of the mitral valve opening.

These intervals as well as the quotient Q-MC/A 2-MO were compared with the mean capillary wedge pressure measured at catheterization.

S ystolic velo city integral (SVI). The area under the maximum ve lo city recording was manually integrated to obtain the systolic velocity in tegral. This measure represents the distance covered by the stroke volume per beat. Analysis of flow in aortic stenosis indicates that the velocity profile is fla t (18). The aortic valve area can therefore also be calculated as the relation between stroke volume and systolic velocity integral (SV/ SVI).

Radionuclide angiography A non-geometric attenuation-corrected technique was used to determine absolute le ft ventricular end-diastolic and end-systolic volumes from ECG-gated blood pool studies (27). Red blood cell labelling in vivo was accomplished by intravenous ad m in istra tio n of stannous pyrophosphate followed a fte r 20 minutes by administration of 740 MBq technetium-99m pertechnetate. Left ventricu la r depth (d) in the le f t a n te rio r oblique 45° projection was measured, and used according to a previously described technique for attenuation correction (27). In accordance with previous reports (28) and our own experience an attenuation coefficient (fi) of 0.13 cm“ 1 was used. Each cardiac cycle was divided in to 24 frames w ith ECG-gating and data were collected using a Portacamera lie (General E le c tric ) with a high resolution parallel hole collimator interfaced with a Gamma 11 computer system (D ig ita l Equipment Corporation) fo r 10 minutes in the LAO 45° projection with the patients in the supine position. Approximately 500 to 1000 cardiac cycles were collected depending on the actual heart rate. No arrhythmia rejection was used.

17 Regions of in te re s t were manually drawn around the le f t ve n tricle in the end-diastolic and the end-systolic frames, with a manually drawn postero-lateral region in the end-systolic frame fo r background correc tio n .

Immediately before data collection in the LAO 45° view a 10 ml venous blood sample was drawn from the patients arm contralateral to the in jection site. The blood sample was contained in a Petri dish and the a c tiv ity was counted with the same gammacamera equipment fo r 3 minutes to obtain the sp e cific a c tiv ity in the blood, corrected fo r background in the field and also corrected for decay to midtime of the sampling from the le ft anterior oblique view. Left ventricular stroke volume (SY) was thereafter calculated according to the formula:

5y - 1 (end-diastolic counts - end-systolic counts) x 1.8 x 106 e“ ^ counts in blood sample x cycles collected x frame time

The procedure was previously validated against stroke volume determina tio n by cineangiography and there was close agreement between the two methods in line with previous reports (29, 30).

Noninvasive calculations of aortic valve area Calculations of the aortic valve area from noninvasive data was accomp lished by two different approaches. A modified Gorlin equation derived by results from Study II was used,

AVA = SV/(ET x 71.2 x Ymax)

Furthermore, according to the continuity equation (31-33):

AVA = SV/SVI

AVA = a o rtic valve area SV = stroke volume ET = left ventricular ejection time SVI = systolic velocity integral

18 RESULTS AND DISCUSSION

Quantification of the transvalvular pressure difference(Studies I , I I ) Study I was carried out to examine the applicability of systolic pres sure difference estimation by Doppler ultrasound in a consecutive series of patients with aortic stenosis. The peak gradient calculated from the Doppler recordings was compared with the peak to peak gradient measured at catheterization. Noninvasi ve and invasive pressure gradi ents showed an acceptable c o rre la tio n , r = 0.85, SEE = + 14 mm Hg. The Doppler peak gradient exceeded the invasive peak to peak gradient in some patients. Although it was realized that the peak to peak gradient and the Doppler peak gradient were conceptually different the possibil ity of a large difference between the two pressure differences was not fu lly appreciated. On the other hand, there was an obvious underestima tion of the pressure difference in several patients with severe steno s is . The patients were examined in 1982 and 1983. The number of reports in the literature at that time was limited and only involved small series of patients and children (who are more easily examined). In addition, patients had often been excluded because of unsatisfactory Doppler recordings (34, 35).

In 1985, when Studies II and III were performed it was apparent that a more refined invasive procedure had to be adopted. A dual catheter technique was used and the Doppler gradient was compared with the in stantaneous peak gradient measured at catheterization. A correlation r = 0.94, SEE = + 9.1 mm Hg was obtained in Study I I . A difference of up to 48 mm Hg between the peak and the peak to peak gradient was also observed. Differences of this magnitude have also been reported by Krafchek (36). Since Hatle's (12) original report numerous studies have been reported, and today Doppler ultrasound is a generally accepted method to estimate the peak gradient in aortic stenosis (37-40). In most studies comparisons have been made with gradients measured by fluid fille d catheters. However, the accuracy of the reference method must also be considered. The fidelity of a fluid fille d catheter pres sure measurement systems and the presence of catheter induced artefacts are thus important considerations. Oscillations may lead to errors in estimation of peak and mean pressure gradients. When a single catheter

19 study is performed, i t is necessary to superimpose the a o rtic tracing onto the recording from the left ventricle and small errors in the timing may result in large differences in the gradients. In addition, measurement of the gradient is based on nonsimultaneous beats. Id e a lly, simultaneous registrations from the le ft ventricle and the ascending aorta using micromanometertipped catheters should be undertaken. Simul taneous invasive and Doppler recordings would seem ideal, but the Doppler investigation may be hindered by difficulties in getting the patient in an optimal position in the catheterization laboratory. Smith et al (40) even reported closer correlations between noninvasi ve and invasive data when the noninvasi ve tests were carried out at a separate time in the noninvasi ve laboratory. Of course, this only holds true if the patient is in a stable hemodynamic condition. Some of the discrep ancies between Doppler and invasive measurements reported may thus be explained by inaccuracies in the reference method. A perfect Doppler recording might even render a better measurement of the pressure d if ference than that achieved by a pull-back procedure using a fluid fille d catheter.

The mean gradient (Study I I ) The mean gradient describes the degree of pressure drop throughout sys tole, and it is usually included in hydraulic orifice formulas. The area between the le ft ventricular and aortic pressure curves is inte grated to determine the mean gradient. The corresponding Doppler gradi ent can be calculated by transferring the velocity curve to a pressure drop curve by applying the Bernoulli equation. This is a burdensome procedure involving numerous measurements of the instantaneous veloci ties at short intervals throughout systole. The instantaneos velocities then have to be squared and the re su ltin g pressure drop curve plotted. The mean pressure difference can then be obtained by planimetry. Good correlation has been reported (36 41) between invasive and noninvasi ve measurements of the mean pressure gradient. The complicated procedure involved has, however, stimulated an in te re st in fin d in g a simpler method of assessment. Zhang et al (41) reported a formula in which the peak and mean ve lo city were included: A pm = 8 V2m [V p/(V p + Ym)] where APm is the mean7 gradient, vp is the peak systolic velocity, and Vm is the mean systolic velocity. We observed that in

20 most patients the plotted pressure drop curve had a configuration very similar to a sine wave. Applied to a pressure drop curve this means that the mean pressure gradient is equivalent to 2/7T multiplied by the peak gradient according to basic trigonometric principles. Others (33, 41) have considered the velocity curve to be symmetrical for practical purposes. Our results showed that there was no significant difference between the mean gradient obtained by planimetry at catheterization and that obtained by multiplying the peak gradient by 2/7T.

Of course the curves cannot invariably be regarded as a sine wave. In theory it may be triangular at one extreme and rectangular at the oth er, resulting in a mean pressure gradient ranging between 1/3 and 1 of the peak gradient. These hypothetical situations were not encountered in our material and do not seem to represent a major problem. Moreover, a simplification should indeed be simple.

A possibility to modify the Gorlin equation is also given by the simple re la tio n between mean and peak pressure gradient. When 0.64 x 4(Vmax)2 is substituted for Ap mean, the equation is transformed to: AVA = SV/(ET x 71.2 x Ymax), a formula very sim ila r to th a t suggested by Ohlsson et al (42).

Determination of aortic valve area (Study III) The Gorlin equation was used to calculate the aortic valve area from invasive data. Stroke volumes measured by le ft ventriculography were used since aortic regurgitation was a frequent finding in our pa tients. In the absence of significant mitral regurgitation this method gives the actual flow across the a o rtic valve. In the noninvasi ve stucjy both the modified Gorlin equation and the integration method were used. In both stroke volume was obtained by radionuclide angiography, which lik e cineangiography gives the to ta l stroke volume. Close correlations with the invasive data were obtained, r = 0.95, SEE = + 0.13 cm2 by the modified Gorlin equation and r = 0.94, SEE = +_ 0.14 cm2 by the integra tion method. The two noninvasive tests gave almost ide n tica l re s u lts , r = 0.98, SEE = +_ 0.09 cm2. More important, all patients with an aortic valve area of less than 1 cm2 by the invasive study were identified by both the noninvasive procedures.

21 There are some previous reports o f valve area calculations in which Doppler ultrasound was used to measure the mean gradient or the SVI (or both). The principal difference between these studies is the method chosen to estimate the stroke volume. Skjärpe et al (31), Zhang et al (32) and Zoghbi et al (33) used a combined echo and Doppler ultrasound method to calculate the stroke volume. The stroke volume can be estima ted from the flow through pulmonary annulus, m itral annulus and through the a o rtic annulus d ire c tly below the obstruction. When stroke volume is calculated using flow through the mitral valve, patients with sig nificant aortic and/or mitral regurgitation has to be excluded, since the actual aortic flow is needed.

Ohlson et al (42) used a CO 2 rebreathing technique, which gives the e ffe c tiv e (or net) stroke volume. This technique has the same lim ita tions, the aortic flow being undervalued in patients with aortic regur gitation. It seems attractive to use the stroke volume through the aor tic annulus because no patient needs to be excluded because of a o rtic or mitral regurgitation. Skjärpe et al (31), however, stress that this approach is d if f ic u lt to handle and subject to s ig n ific a n t errors.

Estimation of le ft ventricular stroke volume by radionuclide angio graphy has the same limitation in respect to mitral regurgitation. In our study there was a close correlation between stroke volume estimated by the isotope technique and le ft ventriculography, r = 0.91, SEE + 9.9 ml. In addition, the ejection fraction is obtained and regional ven tric u la r function can be assessed.

The present study concerned patients with aortic stenosis. Doppler examinations have shown that the so called isolated aortic stenosis is a rarity; most patients show some degree of concomitant regurgitation. A high percentage of leaking valves has been reported also in healthy individuals. There were only 4 out of the 26 patients who did not show either aortic or mitral regurgitation in Studies II and III. This has important implications for calculations of the aortic valve area as discussed previously. The diagnosis aortic regurgitation may be d iffi cult and we believe that Doppler is the most sensitive method. However, quantification of valvular regurgitation is often difficult. Angio

22 graphy has been regarded as the gold standard in many studies but when the severity of regurgitation is divided into four grades there is a considerable overlap (Croft et al, 43). Grading of aortic regurgitation with pulsed wave Doppler technique has been described (44-45) but the v a lid ity of these approaches is doubtful (46, 47). In our experience discrepancies between angiographic gradings and that of pulsed Doppler are often observed. Continuous wave Doppler has also been used for this purpose (48). There is s t i l l no optimal method fo r accurate q u a n tifica tion of regurgitant lesions.

The coexistence of coronary heart disease (CHD) (Study IV) The reported prevalence of coronary heart disease in aortic stenosis shows considerable variation with figures ranging between 20-60% (49- 55). In Study IV, 43% of the patients had significant CHD. This is in agreement with a larger study recently published from our hospital (49), where 37% of 221 patients with aortic valve disease had sig n ifi cant CHD.

It is controversial whether coronary angiography should be carried out in all patients despite the absence of angina pectoris (49-55). A screening procedure to identify patients with clinically significant CHD would be desirable. A standard ECG exercise te s t fo r th is purpose was examined. The coronary in su fficie n cy-in d e x was able to separate pa tients with and without CHD fairly accurate. When an index 3 and an index < 3 was used to identify those patients with and without CHD, 11 out of 15 patients with CHD and 18 out of 19 without CHD could be iden t if ie d . When the presence of angina at the exercise te s t was included the s e n s itiv ity of the te s t was somewhat enhanced. About 50% of the angiographies may be omitted when these criteria are used. On the other hand, approximately 15% of the patients with CHD would be missed which is not acceptable and our policy is to perform coronary angiography in all patients aged 40 years or more regardless of a history of angina.

Estimation of pulmonary c a p illa ry pressure in a o rtic stenosis (Study V) An estimate of left ventricular performance is valuable and often achieved by combined evaluation of two-dimensional echocardiography,

23 radionuclide angiography, and le f t ventriculography. The pulmonary capillary wedge pressure provides an estimate of the le ft ventricular end-diastolic pressure. An abnormally elevated pulmonary c a p illa ry wedge pressure in the presence of other signs o f depressed ve n tricu la r function, i.e . reduced cardiac output, ejection fra c tio n , or stroke work indicates that contractility is probably impaired (56) However, a reduced ve n tricu la r compliance caused by e.g. hypertrophy may elevate the f i l l i n g pressure while the end-diastolic volume remains normal (56). Despite these d iffic u ltie s the pulmonary c a p illa ry wedge pressure might add information, particularly in patients with coexisting valvu lar lesions or ischemic myocardial injuries.

The A 2-MO in te rva l decreased and the Q-MC in te rva l increased in pa tients with mitral stenosis when the degree of obstruction increased (57). Further studies showed that these intervals were useful for esti mation of the pulmonary c a p illa ry pressure in valvular diseases other than mitral stenosis even if there are some conflicting results re ported (58-63). In those studies, the intervals have been recorded using M-mode echocardiography. In Study V, the Doppler amplitude signal of the mitral valve was used, since in our experience the measurement points were more easily identified by Doppler than by M-mode echocar diography. The intervals were combined to a quotient

A2 - MO which compared favourable with the mean pulmonary capillary wedge pres sure at catheterization, r = 0.94, SEE = +_ 2.9 mm Hg. The method accu ra te ly separated those patients with a normal or near normal pulmonary capillary wedge pressure from those with an elevated wedge pressure.

We believe that the measurement points of the amplitude signals cor responds to the start and end of the flow through the mitral valve as indicated by the mean velocity signal. The exact relation to the pres sure crossover of the le ft ventricular pressure curve and le ft atrium/ pulmonary c a p illa ry wedge pressure curves remains to be c la rifie d .

24 FUTURE DEVELOPMENT

Accurate determination of the aortic systolic pressure drop can be achieved by Doppler ultrasound in most patients. Sofisti cation of the transducers/receivers and s till better signal analysis can be expected. This may optimize the recordings in patients in whom difficulties in obtaining sufficient signals are encountered today. The colour Doppler system might make the exact area of in te re s t simpler to id e n tify and w ill speed up the examination. On the other hand, the single nonimaging transducers s till w ill have its definite place in experienced hands. In our experience, it is often quicker to obtain good Doppler recordings when a single transducer is used, at least when compared to duplex transducers. According to wellknown psychological mechanisms (64), it is d ifficult to concentrate on two tasks simultaneously, i.e. both a 2D-picture and a high quality sound signal (Doppler).

Regarding stroke volume determination more refined methods for separate le ft and right ventricular output w ill be developed. Tickner et al (65) have developed a so called microbubble technique, which seems promis ing. This technique may also be applied for estimation of rightsided cardiac pressures.

Identification of patients with coronary heart disease in conjunction with aortic stenosis can be done with some degree of certainty applying the described work test procedure. It w ill be interesting whether the results reported in Study V w ill be validated in a larger group of pa tients with and without concomitant aortic regurgitation. Such a study is in progress.

Kemper et al (66) have injected H 2O2 in the aortic root in dogs and have been able to evaluate regional wall perfusion. H 2O2 or similar echogenic substances may be developed in the future fo r th is purpose. Even echogenic "bubbles" injected on the venous side and passing through the pulmonary c irc u la tio n to the le f t side of the heart have been described (67). Probably we w ill see fu rth e r development in the field of echo "contrast".

25 The application of exercise nuclear angiography and myocardial scinti graphy to detect coronary heart disease in patients with a o rtic steno sis have been reported (68, 69) and fu rth e r advancement can be expected with this technique. The applicability of digital subtraction angio graphy of coronary lesions is another in te re stin g issue.

The use of Doppler in the evaluation of coronary circulation w ill be an expanding fie ld . Gramiak et al (70) have recently published a study of Doppler examination of le ft coronary arterial blood flow. The introduc tion of colour Doppler system may even make diagnosis of significant coronary heart disease possible. At least central major obstruction may be id e n tifie d . The combination of echo dimension measurements and Doppler velocity recordings to obtain coronary volume flow in analogy with stroke volume estimation may lead to further progress in the noninvasive estimation of coronary heart disease. However, coronary angio graphy is not likely to be fully replaced by noninvasive procedures.

When it comes to the evaluation of aortic regurgitation it is the colour Doppler system which w ill be most in te re s tin g . I t is important, however, to re a lize that the same evaluation is available with conven tional Doppler but the colour system may make it easier to detect a re gurgitation and w ill probably speed up the investigation. Even if it is well known it is worth mentioning that the red "stream" coming towards you from the aortic valves (apical view) is not a volume of blood but blood velocity.

26 GENERAL SUMMARY AND CONCLUSIONS

The present study reports on noninvasi ve evaluation of patients with suspected aortic stenosis with emphasis on Doppler ultrasound. The de gree of outflow obstruction can be satisfactorily evaluated as far as the peak gradient and the mean gradient are concerned. The determina tion of aortic valve area can be accomplished by combining radioisotope technique with Doppler tracings.

Noninvasi ve estimation of the pulmonary c a p illa ry wedge pressure seems to be a feasible procedure.

Today there is no generally accepted method to reduce the number of coronary angiographies. The method described may serve as a screening procedure but we feel that coronary angiography s till have to be recom mended as a part of the standard evaluation program.

It is important to realize that not all patients can be satisfactorily evaluated by noninvasi ve methods and therefore invasive procedures should be carried out whenever there is any remaining doubt about the severity of the valvular lesion.

It is suggested that the preliminary evaluation of patients with sus pected aortic stenosis should include a standard ECG, a bicycle ergo- meter test, Doppler and two-dimensional echocardiography. These exam inations make it possible to accurately estimate the pressure gradient and to make a fa ir assessment of the le f t ve n tricu la r function. Assessment of the valve area is also possible by Doppler and tw o-di mensional echocardiography alone, but radionuclide angiography is a useful approach. Before surgery, coronary angiography should probably be undertaken in patients aged 40 years or more.

27 ACKNOWLEDGEMENTS

I want to express my sincere gratitude to:

My wife Anne-Britt, who is very much my wife.

Acting Professor Per Bjerle, for his wholehearted support, enthusiasm and positive criticism ,

Professor Emeritus Håkan Linderholm, whose extensive knowledge of clinical physiology has been to a great help,

Dr. Kjell Karp and Dr. Peter Eriksson, who have been my closest co workers and friends and our combined efforts have made this study pos s ib le ,

Dr. Göran Österman, for his ability to deliver the gold standard of radiology,

Dr. Harold Lake, Newcastle upon Tyne, whom I learned to know during my Yemenitic adventure. He has corrected the language,

Inger G. Andersson and Loi ornai Örnehult, who typed and retyped the manuscript in an excellent way,

The laboratory technicians and other s ta ff at the Department of C lin i cal Physiology and the section of Nuclear Medicine for their kindest cooperation throughout the study.

This study was supported by grants from the Faculty of Medicine, Uni versity of Umeå, and the National Association for Heart and Lung Pa tie n ts .

28 REFERENCES

1. Bonchek L I, Anderson RP, Rösch J. Should coronary arteriography be performed ro u tin e ly before valve replacement? Am J Cardiol 1973;31: 462-6. 2. Basta LL, Raines D, N aijar S, Kioschos JM. C lin ic a l haemodynamic and coronary angiographic correlates of angina pectoris in patients with severe aortic valve disease. Br Heart J 1975;37:150-7. 3. Fallen EL, E lliot WC, Gorlin R. Mechanisms of angina in aortic stenosis. Circulation 1967;36:480-8. 4. Graboys TB, Cohn PF. The prevalence of angina pectoris and abnormal coronary arteriograms in severe aortic valvular disease. Am Heart J 1977;93:683-6. 5. Thompson RH, Ahmed MS, M itchell AG, Towers MK, Yacoub MH. Angina, aortic stenosis and coronary heart disease. Clin Cardiol 1979;2:26- 32. 6. Bermudez GA, Abdelnur R, Mi del1 A, DeMeester T. Coronary artery disease in aortic stenosis: Importance of coronary arteriography and surgical implications. Angiology 1983;34:591-6. 7. Sos TA, Baltaxe HA. The importance of coronary angiography in the evaluation of patients with aortic valvular disease. Am J Roent genol Radium Ther Nucl Med 1975;122:793-9. 8. Nakib A, L ille h e i CW, Eduards JE. The degree of coronary athero sclerosis in aortic valve disease. Arch Pathol 1965;80:517-20. 9. Linhart JW, De La Torre A, Ramsey HW, Wheat J r MW. The significance of coronary artery disease in aortic valve replacement. J Thoracic Cardiovasc Surg 1968;55:811-9. 10. Vincent WR, Buckberg GD, Hoffman JIE . L e ft ve n tricu la r subendocar dial ischemia in severe valvar and supravalvar aortic stenosis. A common mechanism. C irculation 1974;49:326-33. 11. Bell house B, Bell house F. F lu id mechanics of model normal and stenosed aortic valves. Circ Res 1969;25:693-704. 12. Hatle L, Angelsen BA, Tromsdal A. Non-invasive assessment of a o rtic stenosis by Doppler ultrasound. Br Heart J 1980;43:284-92. 13. Hatle L, Brubakk A, Tromsdal A, Angelsen B. Noninvasive assessment of pressure drop in mitral stenosis by Doppler ultrasound. Br Heart J 1978;40:131-40.

29 14. Holen J, Aaslid R, Landmark K, Simonsen S: Determination of pres sure gradient in mitral stenosis with a noninvasive ultrasound Doppler technique. Acta Med Scand 1976;199:455-60. 15. Holen J, Aaslid R, Landmark K, Simonsen S, Östrem T. Determination of effective orifice area in mitral stenosis from non-invasive ultrasound Doppler data and mitral flow rate. Acta Med Scand 1977; 201:83-8. 16. Teirstein P, Yock P, Popp R. The accuracy of Doppler ultrasound measurements of pressure gradients across irregular, dual and tun nellike obstructions to blood flow. Circulation 1985;72:577-84. 17. Requarth J, Goldberg S, Yasko S, Allen H. In v itro v e rific a tio n of Doppler prediction of transvalve pressure gradient and orifice area in stenosis. Am J Cardiol 1984;53:1369-73. 18. Hatle L, Angel sen B. Physics of blood flow. In: Doppler ultrasound in cardiology. Lea & Febiger, 1985. 19. Grossman W, Dexter L. Profiles in valvular heart disease. In: Grossman W, ed. Cardiac catheterization and angiography. 2nd ed. Philadelphia: Lea £ Febiger, 1980:305-324. 20. Lähde S. Cineangiographic determination of le f t ve n tricu la r volume. Accuracy o f method. Acta Radiol 1976:Suppl 348. 21. Brandt PW, Partridge JB, Wattie WJ. Coronary arteriography: Method of presentation of the arteriogram report and scoring system. Clin Radiol 1977;28:361-5. 22. Austen WG, Edwards JE, Frye RL et a l. A reporting system on pa tients evaluated for coronary artery disease. Circulation 1975; 51:4. 23. Gorlin R, Gorlin SG. Hydraulic formula for calculation of the area of the stenotic mitral valve, other cardiac valves, and central circulatory shunts. I. Am Heart J 1951;41:1-29. 24. Judkins MP. Percutaneous transfemoral selective coronary arterio graphy. Radiol Clin North Am 1968;6:467-92. 25. Sjöstrand T. Functional capacity and exercise tolerance in patients with impaired cardiovascular function. Clinical cardiopulmonary physiology. New York: Grune and Stratton, 1960:201. 26. Strandell T. Heart rate and work load at maximal working in te n s ity in old men. Acta Med Scand 1964;176:301-18.

30 27. Links JM, Becker LC, Shindledecker JG et al. Measurement of absol ute le ft ventricular volume from gated blood pool studies. Circula tion 1982;65:82-91. 28. Schneider RM, Jaszczak RJ, Coleman RE, Cobb FR. Disproportionate effects of regional hypokinesis on radionuclide ejection fraction: compensation using attenuation-corrected v e n tricu la r volumes. J Nucl Med 1984;25:747-54. 29. Dehmer GJ, Firth BG, Lewis SE, Willerson JT, H ill is LD. Direct mea surement of cardiac output by gated q u iiibriurn blood pool s c in ti graphy: V alidation of scintigraphic volume measurements by a non geometric method. Am J Cardiol 1981,47:293-300. 30. Verani MS, Gaeta J, LeBlanc AD et a l. V alidation o f le f t ventricu la r stroke volume measurements by radionuclide angiography. J Nucl Med 1985;26:1394-1401. 31. Skjaerpe T, Hegrenaes L, Hatle L. Noninvasive estimation of valve area in patients with aortic stenosis by Doppler ultrasound and two-dimensional echocardiography.Circulation 1985;72:810-8. 32. Zhang Y, Myhre E, Nitter-Hauge S. Noninvasive quantification of the aortic valve area in aortic stenosis by Doppler echocardiography. Eur Heart J 1985;6:992-8. 33. Zoghbi WA, Farmer KL, Soto JG, Nelson JG, Quinones MA. Accurate noninvasive quantification of stenotic aortic valve area by Doppler echocardiography. Circulation 1986;73:452-9. 34. Oliveira Lima C, Sahn DJ, Yaldes-Cruz LM et al. Prediction of the severity of le ft ventricular outflow tract obstruction by quantita tive two-dimensional echocardiographic Doppler studies. Circulation 1983;68:348-54. 35. Stamm RB, Martin RP. Quantification of pressure gradients across stenotic valves by Doppler ultrasound. J Am Coll Cardiol 1983;2: 707-18. 36. Krafchek J, Robertson JH, Radford M, Adams D, K isslo J. A recon sideration of Doppler assessed gradients in suspected a o rtic steno sis. Am Heart J 1985;110:765-73. 37. Teien D, Eriksson P. Quantification of transvaivular pressure d if ferences in aortic stenosis by Doppler ultrasound. Int J Cardiol 1985;7:121-6.

31 38. Currie PJ, Seward JB, Reeder GS et a l. Continuous-wave Doppler echocardiographic assessment o f severity o f c a lc ific a o rtic steno sis: a simultaneous Doppier-catheter correlative study in 100 adult patients. Circulation 1985;71:1162-9. 39. Hegrenaes L, Hatle L. Aortic stenosis in adults. Non-invasi ve esti mation of pressure differences by continuous wave Doppler echocar diography. Br Heart J 1985;54:396-404. 40. Smith MD, Dawson PL, Elion JL et a l. Systematic co rrelation of con tinuous-wave Doppler and hemodynamic measurements in patients with aortic stenosis. Am Heart J 1986;111:245-52. 41. Zhang Y, Nitter-Hauge S. Determination of the mean pressure grad ient in aortic stenosis by Doppler echocardiography. Eur Heart J 1985;6:999-1005. 42. Ohlsson J, Wranne B. Noninvasi ve assessment of valve area in pa tients with aortic stenosis. J Am Coll Cardiol 1985;7:501-8. 43. Croft C, Lipscomb K et al. Limitations of qualitative angiographic grading in aortic or mitral regurgitation. Am J Cardiol 1984;53: 1593-8. 44. Ciobanu M, Abbasi AS, Allen M, Hermer A, Spell berg P. Pulsed Doppler echocardography in the diagnosis and estimation of severity of aortic insufficiency. Am J Cardiol 1982;49:339-43. 45. Bommer WJ, Mapes R, M ille r L, Mason DT, DeMaria AN. Q uantification of aortic regurgitation with two-dimensional Doppler echocardio graphy (A bstract). Am J Cardiol 1981;47:412. 46. Hatle L, Angelsen B. Doppler ultrasound in cardiology. Physical principles and clinical applications. 2nd ed. Philadelphia: Lea & Febiger, 1985:154-162. 47. Wranne B, Ask P, Loyd D. Quantification of heart valve regurgita tion: a critical analysis from a theoretical and experimental point of view. Clin Physiol 1985;5:81-8. 48. Teien D, Karp K, Eriksson P, Österman G. Q uantification of a o rtic regurgitation by continuous wave Doppler ultrasound and diastolic blood pressure. J Cardiovasc Ultrasonogr 1986;5:107-10. 49. Olofsson B-0, B jerle P, Åberg T, Österman G, Jacobsson K-A. Preval ence of coronary artery disease in patients with valvular heart disease. Acta Med Scand 1985;218:365-71.

32 50. Baxter RH, Reid JM, McGuiness JB, Stevenson JG. Relation of angina to coronary artery disease in mitral and in aortic valve disease. Br Heart J 1978;40:918-22. 51. Morrison GW, Thomas RD, Grimmer SFM, S ilverton PN, Smith DR. In c i dence of coronary artery disease in patients with valvular heart disease. Br Heart J 1980;44:630-7. 52. Ramsdale KM, Bennett DH, Bray CL, Ward C, Beton CD, Faragher EB. Coronary arteriography prior to valve replacement. Eur Heart J 1981;2:83-6. 53. Exdactylos N, Sugrue DD, Oakley CM. Prevalence of coronary artery disease in patients with isolated aortic valve stenosis. Br Heart J 1984;51:121-4. 54. Mandai AB, Gray IR. Significance of angina pectoris in aortic valve stenosis. Br Heart J 1976;38:811-5. 55. Hancock EW. Aortic stenosis, angina pectoris, and coronary artery disease. Am Heart J 1977;93:382-93. 56. Braunwald E. Assessment of cardiac function. In: Braunwald E, ed. Heart Disease. A textbook of cardiovascular medicine. New York: Saunders, 1984:467-487. 57. Bayer 0, Loogan F, Wolter HH. The m itral opening snap in the quan tita tive diagnosis of mitral stenosis. Am Heart J 1956;51:234-45. 58. Abdulla AM, Kavouras T, Rivas F, Stefadouros MA. Determination of mean pulmonary c a p illa ry pressure by a noninvasi ve technique. JAMA 1980;243:1539-42. 59. Paiamo AR, Quinones MA, Waggoner AD, Kumpuris AG, M ille r RR. Echo- phonocardiographic determination of le ft atrial and le ft ventricu lar fillin g pressures with and without mitral stenosis. Circulation 1980;61:1043-7. 60. Askenazi J, Koenigsberg DI, Ziegler JH, Lesch M. Echocardiographic estimates of pulmonary artery wedge pressure. N Engl J Med 1981; 305:1566-8. 61. Askenazi J, Koenigsberg DI, Ribner HS, Plucinski D, Silverman IM, Lesch M. Prospective study comparing different echocardiographic measurements of pulmonary capillary wedge pressure in patients with organic heart disease other than mitral stenosis. J Am Coll Cardiol 1983;2:919-25.

33 62. Rahko PS, Shaver JA, Salerni R, Gamble WH, Reddy PS. Echophonocar- diographic estimates of pulmonary artery wedge pressure in mitral stenosis. Am J Cardiol 1985;55:462-9. 63. Gamble WH, Salerni R, Shaver JA. The noninvasive assessment of pul monary capillary wedge pressure in mitral regurgitation. Am Heart J 1984;107:950-8. 64. Moray N. A ttention . Selective processes in vision and hearing. London: Hutchinson Educational Ltd, 1969. 65. Tickner E. Precision microbubbles for right side intracardiac pres sure and flow measurements. In: Meitzer RS, Roelandt J, eds. Con trast echocardiography. Martinus Nijhoff Publishers. 1982:313-324. 66. Kemper AJ, O'Boyle JE et al. Hydrogen peroxide contrast-enhanced two-dimensional echocardiography: real-time in vivo delineation of regional myocardial perfusion. Circulation 1983;68:603-11. 67. Transmission of echocardiographic contrast through the lungs. In: Meitzer RS, Roelandt J, eds. Contrast echocardiography. Martinus Nijhoff Publishers. 1982:304-312. 68. Mi lanes JC, Palei i J et al. Detection of coronary artery disease in aortic stenosis by exercise gated nuclear angiography. Am J Cardiol 1984;54:787-91. 69. Huikuri H, Korhonen U, Heikkilä J, Takkunen J. Detection of coro nary artery disease by thallium scintigraphy in patients with valv ular heart disease. Br Heart J 1986;56:146-51. 70. Gramiak R, Holen J, Moss AJ et a l. Left coronary a rte ria blood flow: noninvasive detection by Doppler US. Radiology 1986;159:657- 62.