- Email: [email protected]
Doppler ultrasound evaluation of valvar pulmonary stenosis from multiple transducer positions in children requiring pulmonary valvuloplasty
CONGENITAL HEART DISEASE
Doppler Ultrasound Evaluationof Valvar Pulmonary Stenosis from Multiple Transducer Positions in Children Requiring PulmonaryValvuloplasty ELMAN G. FRANTZ,
MD, and NORMAN H. SILVERMAN,
The modified Bernoulli equation has been used to determine the peak pressure difference across stenotic pulmonary valves. The direction of the poststenotic jet may be eccentric in relation to the axis of the pulmonary artery. Consequently, the maximal velocity obtained from the parastemal transducer position may inaccurately estimate peak pressure difference. Thus, Doppler-derived estimates of pressure difference from the parasternal, subcostal, apical, and suprasternal notch transducer positions were compared with peak-to-peak pulmonary artery to right ventricle catheter withdrawal pressure differences in 24 patients admitted for pulmonary valvuloplasty. Suprastemal, subcostal or apical transducer positions produced higher maximal velocities than the parastemal transducer position in 12 of 24 patients when the studies were performed before
T
he application of the modified Bernoulli equation to estimate pressure differences across stenotic orifices requires axial alignment of the jet and the Doppler ultrasound beam.l-3 This technique has become increasingly important for determining which children require valvuloplasty. Positioning the transducer in the left parasternal area and directing the ultrasound beam through the main axis of the pulmonary artery is the accepted approach to measure normal blood flow From the Department of Pediatrics, Division of Pediatric Cardiology, and the Cardiovascular Research Institute, University of California, San Francisco, California. This study was supported in part by training grant HL07544 from the National Institutes of Health, Bethesda, Maryland. Manuscript received October 5, 1987; revised manuscript received and accepted December 10, 1987. Address for reprints: Norman H. Silverman, MD, M342A, University of California-San Francisco, Third and Parnassus Avenue, San Francisco, California 94143.
MD
cardiac catheterization and in 8 of 12 patients when performed during cardiac catheterization. The Doppler-derived estimates using the highest maximal velocity predicted catheterization pressure difference accurately when the measurements were not performed simultaneously (y = 1.05x - 3.3, r = 0.88, standard error of the estimate f 18.7 mm Hg) and the correlation was closer when the measurements were performed simultaneously (y = 1.09x - 2.7, r = 0.97, standard error of the estimate f 9.4 mm Hg). The transducer position that yielded the highest maximal velocity in an individual patient was the same before and after valvuloplasty. In all groups, the correlation with pressure at cardiac catheterization was improved by using the highest maximal velocity rather than the parastemal maximal velocity. (Am J Cardiol 1988;81:844-849)
velocity in the main pulmonary artery4-6 and to measure poststenotic maximal velocity in valvar pulmonic stenosis.7,8In valvar pulmonic stenosis, the direction of the poststenotic jet may deviate significantly from the ultrasound axis obtained from the parasternal transducer position, thereby adversely affecting the accuracy of the Doppler-derived estimate of pressure difference. The use of an alternative transducer position such as the subcostal, suprasternal notch or apical approach might result in a closer ultrasound beam alignment with the jet, allowing a more accurate estimate of pressure difference. With the advent of percutaneous balloon valvuloplasty for valvar pulmonic stenosis, appropriate timing of valvuloplasty can be aided by accurate Doppler-derived estimates of severity. The objective of this study was to determine whether the use of multiple transducer positions would improve the accuracy of estimation of pressure difference in children considered for balloon pulmonary valvuloplasty both before and after the procedure. 644
April I,1988
Methods Patients: We identified 74 cases of isolated valvar pulmonic stenosis from January 1984 to October 1986. The patients’ agesranged from 1 day to 16 years. Fortyeight patients had undergone catheterization and continuous wave ultrasound studies. In the 24 patients who form the basis of this report, pulmonary artery velocity was recorded by Doppler ultrasound from more than 1 transducer position. Their ages ranged from 10 days to 13 years (mean 34 months f 43 months standard deviation) and the median age at the time of study was 20 months. Continuous wave Doppler echocardiographic studies were obtained on an IREX 3B system using a Pedoff ~-MHZ continuous wave stand-alone transducer or on an ATL UltraMark 8 system, which uses a similar transducer. These small stand-alone transducers can be acutely angled in directions that larger transducers cannot achieve. The recordings were performed with the patient resting quietly in the left lateral decubitus position for the parasternal and apical transducer positions or supine for the subcostal transducer position. For the suprasternal notch transducer position the patient was supine with the neck extended and the head turned toward the right shoulder in order to obtain anterior alignment of the ultrasound beam. From each transducer position the transducer was angled to obtain the signal of highest velocity. The left parasternal area was evaluated with the transducer positioned in the second through fifth intercostal spaces and the highest velocity obtained was taken to represent the parasternal transducer position for analysis. Most studies were performed within the 18 hours before cardiac catheterization. The continuous wave Doppler signals were recorded on silver iodide paper at a speed of 50 mm/s, on one-half inch videotape for subsequent analysis or on both. The continuous wave Doppler records were reviewed by one of us (EGF) without knowledge of catheterization results or prior echocardiographic interpretations. The Doppler signal was considered to be
THE AMERICAN
JOURNAL
Volume 61
045
satisfactory when there was a well-defined outline of the envelope.3 Maximal velocity was determined to the nearest 0.25 m/s. Doppler-estimated pressure differences were calculated by the simplified Bernoulli equation (AP = 4Vz) and compared by linear regression analysis with measurements of peak-to-peak pressure difference obtained by pulmonary artery to right ventricle withdrawal during cardiac catheterization. We studied 15 consecutive patients referred for pulmonary valvuloplasty prospectively. Continuous wave Doppler studies were performed during cardiac catheterization immediately after pulmonary artery to right ventricle catheter withdrawal before and after valvuloplasty. All Doppler recordings were obtained on an IREX 3B system without Z-dimensional guidance. The transducer position that produced the highest maximal velocity was assumed to approximate the axis of blood flow most closely. The Doppler signals were recorded and analyzed as already described without knowledge of cardiac catheterization findings. Adequate Doppler data from at least 2 transducer positions were obtained in 12 patients before valvuloplasty and in 10 patients after valvuloplasty. Doppler signals were not obtained from the apical transducer position in these patients. Peak-to-peak catheter pressure differences and Doppler-derived pressure differences from various transducer positions were compared by linear regression analysis. To evaluate inter- and intraobserver variability, the maximal velocity of 50 of the Doppler recordings was interpreted by a second observer (NHS) and by the first observer (EGF] at a second reading without knowledge of prior interpretations. Interobserver variability was calculated as the difference between the 2 observers’ readings divided by the average of their readings and expressed as a positive percentage. Intraobserver variability was calculated as the difference between the first observer’s 2 readings divided by the average of the z: I seadings and expressed as a positive percentage.g
Swrosternal
FIGURE 1. Left, a lateral right ventricular angiogram showing an “eccentric” poststenotic jet in a patient with valvar pulmonic stenosis. Right, line drawing corresponding to angiogram showing the possible range of directions of a poststenotic jet (shaded cone above pulmonary valve) and the axis of ultrasound from multiple transducer positions.
OF CARDIOLOGY
Chest
notch
wall-
Parasternal
Subcortol-
Diaphragm Right
ventricle
848
MULTIPLE DOPPLER TRANSDUCER
POSITIONS
IN PULMONARY
STENOSIS
Results Of the 48 patients who had nonsimultaneous catheterization and Doppler determinations of pressure difference, 24 had Doppler signals from parasternal and suprasternal, subcostal or apical transducer positions, 21 had parasternal signals alone and 3 had suprasternal signals alone. Of the 24 patients studied from multiple transducer positions, the maximal velocity from the parasternal transducer position was equal to or greater than the maximal velocity of the highest other transducer position in only 12 (50%]. In the other 12 patients, a higher maximal velocity was obtained from the suprasternal, subcostal or apical transducer position. In a right ventricular angiogram obtained from 1 of the patients an anterosuperiorly directed poststenotic jet is demonstrated (Figure 11. Continuous-wave Doppler’signals in this patient confirmed that suprasternal, subcostal and apical transducer positions produced higher maximal velocities that were more predictive of catheterization findings than the parasternal maximal velocity (Figure 2). Linear regression analysis comparing nonsimultaneous Doppler-derived pressure differences using the parasternal transducer position versus catheterization pressure differences showed suboptimal strength of correlation (r = 0.78) with considerable variability (standard error of the estimate f 20.6 mm Hg). More importantly, the least squares line of best fit (y = 1.36x - 12.5) deviated significantly from the line of identity reflecting underestimation of the more severe obstructions (Figure 3, top]. When Doppler-derived pressure difference using the signal of highest maximal velocity [regardless of transducer position] was compared with catheterization pressure difference, the correlation was stronger (r = 0.86) with similar variability (standard error of the estimate f 18.7 mm Hg) but the regression line (y = 1.05x - 3.3) very closely approximated the line of identity (Figure 3, bottom].
Of the 12 patients who underwent simultaneous catheterization and Doppler measurements before valvuloplasty, the parasternal maximal velocity was equal to or greater than the highest maximal velocity from a suprasternal or subcostal transducer position in only 4 (33%). A suprasternal or subcostal transducer position resulted in a higher maximal velocity than the parasternal transducer position in 8 patients (67%, Figure 4). Despite the fact that the measurements in these 12 patients were made simultaneously, linear regression analysis resulted in a weak correlation (r = 0.59) with greater variability (standard error of the estimate f 32.1 mm Hg) and poor approximation of the line of
Parasternal y-
Transducer
1.36x-12.5
R-0.78
Position
SEEk20.6
nr45
180
/
/
100 -
60 4020 . ol!, 0
.
/ 20
,
.
40
, 60
.
,
,
80
,
.
100
, 120
.
,
,
140
, 160
.
. 180
AP-Doppler
Highest Velocity Transducer y-
1.05x
- 3.3
R -0.66
SEE t18.7
Position n-24
1801
k
160 -1 14O120 100 80 60 -t
, . , . , , , . , . , . , .
ok/I,. 0
20
40
60
80
100
120
140
160
1 0
AP-Doppler FIGURE 2. Continuous-wave Doppler signals from multiple transducer positions in the patient shown in Figure 1. Catheterization pressure difference was 154 mm Hg. The parasternal (PARA) maximal velocity is lower and less predictive than the subcostal (SUBC), apical (APEX) or suprasternal notch (SSN) maximal velocities (arrowheads). Note the inverted zero line and scale for the SSN velocities.
FIGURE 3. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Cath), both in mm Hg. At the fop, AP Doppler was obtained using the nonsimultaneous parasternal maximal velocity in 45 patients. Boffom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position in 24 patients. The solid line is the line of best fit. The dashed line is the line of identity.
April 1, 1988
identity [y = 1.36x + 5.1) when the parasternal signal was applied (Figure 5, top). However, the Doppler signals from the transducer position of highest maximal velocity correlated very strongly (r = 0.97) with catheterization measurements. The variability was reduced [standard error of the estimate f 9.4 mm Hg) and the line of best fit (y = 1.09x - 2.7) closely approximated the line of identity (Figure 5, bottom). After valvuloplasty, simultaneous Doppler and catheterization data were obtained in 10 patients. The parasternal signal (obtained in only 9 patients] was equal to or greater than the highest signal from the other position in only 4 patients (44%). The suprasternal or subcostal transducer position produced a higher maximal velocity in 5 patients (56%). An improved predictive accuracy using Doppler signals from the transducer position yielding the highest velocity is confirmed by linear regression analysis (Figure 6). The transducer position that yielded the highest velocity in an individual patient was the same before and after valvuloplasty. Interobserver variability was 3.3 f 4.3% [mean f standard deviation] and intraobserver variability was 3.1 f 3.9%. Linear regression analysis provides an interobserver correlation of y = 1.01x + 0.03, r = 0.97, standard error of the estimate f 0.23 m/s and an intraobserver correlation of y = 0.95x + 0.09, r = 0.98, standard error of the estimate f 0.17 m/s. In no patient was more than a l/4 m/s difference found between the 2 observers’ readings in any view.
THE AMERICAN
JOURNAL
Parasternal Dprpf+prnp+
Transducer Transducer
y = 1.36x + 5.1
R =0.59 = 0.59
Position Position
SEE k32.1 r 32.1
n = 12
I
c 5 Ii
100 80 -
a
,
op. 0
.
20
,
.
40
,
.
60
,
.
80
,
,
100
,
.
120
,
.
140
1 0
AP-Doppler
Highest Velocity y-
Transducer
847
140-
The parasternal transducer position has been the standard portal for the evaluation of valvar pulmonic stenosis by Doppler echocardiography. Oliveira Lima et al7 reported a very high accuracy of prediction of vs. Parasternal
Volume 61
nonsimultaneous catheterization findings by Dopplerestimated pressure difference in 16 patients using only the parasternal transducer position. Kosturakis et a1,8 also using only the parasternal approach, achieved acceptable predictive accuracy in 11 patients although some severe obstructions were underestimated. We were unable to achieve this degree of accuracy using parasternal measurements alone. In contrast, we identified patients in whom the parasternal approach alone had seriously underestimated true pressure difference at subsequent cardiac catheterization and, thus, led to this investigation. Multiple Doppler transducer positions have been used in adults for the evaluation of aortic stenosislO
Discussion
Highest
OF CARDIOLOGY
1.09x-2.7
Transducer R-O.97
Position
SEEA9.4
n-12
160
Position
140
160 -. 140 J
120 -
S d Ii. a
100 -
80 60
80 60 40'L
,
0
20
.
,
40
.,
.
60
I
80
.,
.,
100
.,
120
.
140
160
AP-Doppler
Parasternal
Cath
Highest
FIGURE 4. Patient-by-patient comparison of pressure difference (AP) predicted by the parasternal maximal velocity (leff), the highest maximal velocity regardless of transducer position (right) or measured at catheterization (Cath, middle). Individual patient values are connected by lines.
FIGURE 5. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Calh), both in mm Hg, before valvuloplasty. Top, AP Doppler was obtained using the simultaneous parasternal maximal velocity. Bottom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position. The solid line is the line of best fit. The dashed line is the line of identity.
a48
MULTIPLE DOPPLER TRANSDUCER
POSITIONS
IN PULMONARY
and nonstenotic pulmonaryll and aorti$ blood flow. Williams et allo reported improved assessment of aortic stenosis using 3 transducer positions, although the highest of 3 signals was only marginally more accurate than the apical signal alone. Stevenson and KawaborP3 used both parasternal and subcostal approaches in applying modified pulsed Doppler systems to high velocity poststenotic flow in children with pulmonic stenosis but did not comment on the relative accuracy of the 2 transducer positions. Johnson et alI4 measured maximal velocities from the parasternal and subcostal approaches in 12 children with valvar pulmonic stenosis, but the subcostal signal substantially improved the prediction of catheterization pressure difference in Parasternal
Transducer
y= 2.95x-43.6
.‘,
ok 0
. 20
R=0.81
,
Position
SEEi14.5
,
.
.
60
40
n=9
,
,
1
100
80
AP-Doppler Highest
Velocity
y = 0.91x + 0.6
0
20
40
Transducer
R = 0.67
Position
SEE + 13.8
60
80
n = 10
100
AP-Doppler FIGURE 6. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Cath), both in mm Hg, after vaivuiopiasty. At fop, AP Doppler was obtained using the simultaneous parasternai maximal velocity. Bottom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position. The so/id line is the line of best fit. The dashed line is the line of identity.
STENOSIS
only 1 patient. Murphy et alI5 also have supported the use of subcostal imaging but have suggested that the suprasternal position does not provide adequate alignment of the ultrasound beam. Based on these studies, the value of multiple transducer positions in valvar pulmonic stenosis had not previously been conclusively demonstrated. Our results, however, demonstrate superior prediction of catheterization pressure differences by applying the Bernoulli equation to the highest velocity signal obtained from multiple transducer positions. The direction of the poststenotic jet in valvar pulmonic stenosis is unknown and varies from patient to patient; thus, the transducer position that will produce the closest alignment with this jet and the highest velocity cannot be anticipated. In our 10 patients in whom echocardiographic studies were performed before and after valvuloplasty, the transducer position producing the highest velocity in an individual was not altered. However, because our sample size is small, we cannot generalize to the population at large. The direction of the poststenotic jet might be altered by valvuloplasty in some patients. For these reasons, multiple transducer positions should be used in each patient before and after valvuloplasty. Several other findings of our study are notable. The highest Doppler velocities recorded at cardiac catheterization (Figure 5, bottom) more accurately predicted catheterization pressure difference than the highest velocities not recorded simultaneously (Figure 3, bottom]. This may be due to the level of sedation and transvalvar flow rate, which were nearly identical in the former group and highly variable in the latter. A poor correlation was found using the parasternal velocities obtained at cardiac catheterization. [Figure 5, top), which might be related to technical limitations of performing Doppler echocardiography in the catheterization laboratory where optimal patient position and time constraints may have had an adverse impact on the Doppler study. In 7 of 24 patients (29%] with nonsimultaneous measurements (Figure 3), Doppler-derived pressure difference overestimated catheterization findings by at least 10 mm Hg. Level of arousal and, thus, cardiac output are likely to be greater in unsedated outpatients than in sedated patients undergoing catheterization. Doppler-derived measurements are known to overestimate simultaneous peak-to-peak catheterization measurements and to correlate more closely with peak instantaneous pressure differences.16r17Nonetheless, we and others have found excellent predictive accuracy with peak-to-peak measurements without overestimation.7~8~13~14 Accurate Doppler-derived estimates of severity especially in the range of values where treatment may be contemplated are valuable because they influence timing of balloon valvuloplasty and aid in the accurate follow-up thereafter.
References 1. Hatle L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea ~ and Febiger, 1986:24. 2. Valdes-Cruz LM, Horowitz S, Sahn DJ, Larson D, Olive& Lima C, Mesel E. Validation of a Doppler echocardiographic method for calculating severity of discrete stenotic obstructions in a canine preparation with a pulmonary arte-
April 1, 1988
rial bond. Circulation 1984;69:li77-1181. 3. Ha& L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea and Febiger, 1985:108-110. 4. Gardin JM, Burn CS, Childs WJ, Henry WL. Evaluation of blood flow velocity in the ascending aorta and main pulmonary artery of normal subjects by Doppler echocardiography. Am Heart r 1984;107:310-319. 5. Wilson N, Goldberg S], Dickinson DF, Scott 0. Normal intracardiac and great artery blood velocity measurements by pulsed Doppler echocardiography. Br Heart / X985$3:451-458, 6. Grenadier E, Oliveira Lima C, Allen HD, Sahn DJ, Barron JV, Valdes-Cruz LM, Goldberg SJ. Normal intracardiac and great vessel Doppler flow velocities in infants and children. [ACC 1984;4:343-350. 7. Oliveira Lima C, Sahn DJ, Valdes-Cruz LM, Goldberg SJ,Barron JV, Allen HD, Grenadier E. Noninvasive prediction of transvalvular pressure gradient in patients with pulmonary stenosis by quantitative two-dimensional echocardiographic Doppler studies. Circulation 1983;67:866-871. 8. Kosturakis D, Allen HD, Goldberg SJ,Sahn DJ, Valdes-Cruz LM. Noninvasive quantification ofstenotic semilunar valve areas by Doppler echocardiography. JACC 1984;3:1256-1262. 9. Gardin JM, Dabestani A, Matin K, Allfie A, Russell D, Henry WL. Reproducibility of Doppler aortic blood flow measurements: studies on intraobserver, interobserver and day-to-day variability in normal subjects. Am J Cardiol 1984;54:1092-1098, 10. Williams GA, Labovitz AJ, Nelson JG, Kennedy HL. Value of multiple
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume 61
049
echocardiographic views in the evaluation of oortic stenosis in adults by continuous-wave Doppler. Am J Cardiol 1985;55:445-449, 11. Lighty GW, Gargiulo A, Kronzon I, Politzer F. Comparison of multiple views for the evaluation of pulmonary arterial blood flow Doppler echocordiography. Circulation 1986;74:1002-1006, 12. Vijayaraghavan G, Singham KT, Tei C, Wong AL, Shah PM. Aortic flow velocity in older subjects using continuous-wave Doppler echocardiography. Am
Heart
T 1984:107:1275.
13. Stevenson JG, Kawahori I. Noninvasive determination of pressure gradients in children: two methods employing pulsed Doppler echocardiography. IACC
1984;3:179-192.
14. Johnson GL, Kwan OL, Handshoe S, Noonan I, DeMaria AN. Accuracy of combined two-dimensional echocardiography and continuous wave Doppler recordings in the estimation of pressure gradient in right ventricular outlet obstruction. JACC 1984;3:1013-1018. 15. Murphy DJ, Ludomirsky A, Danford AD, Huhta ]C. Doppler echocardiography in pulmonary stenosis. Echocardiography 1987;4:187-202. 16. Currie PJ,Seward JB, Reeder GS, Vlietstra RE, Bresnahan DR, Bresnahan JF, Smith HC, Hagler DJ, Tajik AJ. Continuous-wave Doppler echocardiographic assessmentof severity of calcific aortic stenosis: a simultaneous Doppler-catheter correlative study in 100 adult patients. Circulation 1985;71:11621169. 17. Hatle L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea and Febiger, 1985:130-132.
Doppler Ultrasound Evaluationof Valvar Pulmonary Stenosis from Multiple Transducer Positions in Children Requiring PulmonaryValvuloplasty ELMAN G. FRANTZ,
MD, and NORMAN H. SILVERMAN,
The modified Bernoulli equation has been used to determine the peak pressure difference across stenotic pulmonary valves. The direction of the poststenotic jet may be eccentric in relation to the axis of the pulmonary artery. Consequently, the maximal velocity obtained from the parastemal transducer position may inaccurately estimate peak pressure difference. Thus, Doppler-derived estimates of pressure difference from the parasternal, subcostal, apical, and suprasternal notch transducer positions were compared with peak-to-peak pulmonary artery to right ventricle catheter withdrawal pressure differences in 24 patients admitted for pulmonary valvuloplasty. Suprastemal, subcostal or apical transducer positions produced higher maximal velocities than the parastemal transducer position in 12 of 24 patients when the studies were performed before
T
he application of the modified Bernoulli equation to estimate pressure differences across stenotic orifices requires axial alignment of the jet and the Doppler ultrasound beam.l-3 This technique has become increasingly important for determining which children require valvuloplasty. Positioning the transducer in the left parasternal area and directing the ultrasound beam through the main axis of the pulmonary artery is the accepted approach to measure normal blood flow From the Department of Pediatrics, Division of Pediatric Cardiology, and the Cardiovascular Research Institute, University of California, San Francisco, California. This study was supported in part by training grant HL07544 from the National Institutes of Health, Bethesda, Maryland. Manuscript received October 5, 1987; revised manuscript received and accepted December 10, 1987. Address for reprints: Norman H. Silverman, MD, M342A, University of California-San Francisco, Third and Parnassus Avenue, San Francisco, California 94143.
MD
cardiac catheterization and in 8 of 12 patients when performed during cardiac catheterization. The Doppler-derived estimates using the highest maximal velocity predicted catheterization pressure difference accurately when the measurements were not performed simultaneously (y = 1.05x - 3.3, r = 0.88, standard error of the estimate f 18.7 mm Hg) and the correlation was closer when the measurements were performed simultaneously (y = 1.09x - 2.7, r = 0.97, standard error of the estimate f 9.4 mm Hg). The transducer position that yielded the highest maximal velocity in an individual patient was the same before and after valvuloplasty. In all groups, the correlation with pressure at cardiac catheterization was improved by using the highest maximal velocity rather than the parastemal maximal velocity. (Am J Cardiol 1988;81:844-849)
velocity in the main pulmonary artery4-6 and to measure poststenotic maximal velocity in valvar pulmonic stenosis.7,8In valvar pulmonic stenosis, the direction of the poststenotic jet may deviate significantly from the ultrasound axis obtained from the parasternal transducer position, thereby adversely affecting the accuracy of the Doppler-derived estimate of pressure difference. The use of an alternative transducer position such as the subcostal, suprasternal notch or apical approach might result in a closer ultrasound beam alignment with the jet, allowing a more accurate estimate of pressure difference. With the advent of percutaneous balloon valvuloplasty for valvar pulmonic stenosis, appropriate timing of valvuloplasty can be aided by accurate Doppler-derived estimates of severity. The objective of this study was to determine whether the use of multiple transducer positions would improve the accuracy of estimation of pressure difference in children considered for balloon pulmonary valvuloplasty both before and after the procedure. 644
April I,1988
Methods Patients: We identified 74 cases of isolated valvar pulmonic stenosis from January 1984 to October 1986. The patients’ agesranged from 1 day to 16 years. Fortyeight patients had undergone catheterization and continuous wave ultrasound studies. In the 24 patients who form the basis of this report, pulmonary artery velocity was recorded by Doppler ultrasound from more than 1 transducer position. Their ages ranged from 10 days to 13 years (mean 34 months f 43 months standard deviation) and the median age at the time of study was 20 months. Continuous wave Doppler echocardiographic studies were obtained on an IREX 3B system using a Pedoff ~-MHZ continuous wave stand-alone transducer or on an ATL UltraMark 8 system, which uses a similar transducer. These small stand-alone transducers can be acutely angled in directions that larger transducers cannot achieve. The recordings were performed with the patient resting quietly in the left lateral decubitus position for the parasternal and apical transducer positions or supine for the subcostal transducer position. For the suprasternal notch transducer position the patient was supine with the neck extended and the head turned toward the right shoulder in order to obtain anterior alignment of the ultrasound beam. From each transducer position the transducer was angled to obtain the signal of highest velocity. The left parasternal area was evaluated with the transducer positioned in the second through fifth intercostal spaces and the highest velocity obtained was taken to represent the parasternal transducer position for analysis. Most studies were performed within the 18 hours before cardiac catheterization. The continuous wave Doppler signals were recorded on silver iodide paper at a speed of 50 mm/s, on one-half inch videotape for subsequent analysis or on both. The continuous wave Doppler records were reviewed by one of us (EGF) without knowledge of catheterization results or prior echocardiographic interpretations. The Doppler signal was considered to be
THE AMERICAN
JOURNAL
Volume 61
045
satisfactory when there was a well-defined outline of the envelope.3 Maximal velocity was determined to the nearest 0.25 m/s. Doppler-estimated pressure differences were calculated by the simplified Bernoulli equation (AP = 4Vz) and compared by linear regression analysis with measurements of peak-to-peak pressure difference obtained by pulmonary artery to right ventricle withdrawal during cardiac catheterization. We studied 15 consecutive patients referred for pulmonary valvuloplasty prospectively. Continuous wave Doppler studies were performed during cardiac catheterization immediately after pulmonary artery to right ventricle catheter withdrawal before and after valvuloplasty. All Doppler recordings were obtained on an IREX 3B system without Z-dimensional guidance. The transducer position that produced the highest maximal velocity was assumed to approximate the axis of blood flow most closely. The Doppler signals were recorded and analyzed as already described without knowledge of cardiac catheterization findings. Adequate Doppler data from at least 2 transducer positions were obtained in 12 patients before valvuloplasty and in 10 patients after valvuloplasty. Doppler signals were not obtained from the apical transducer position in these patients. Peak-to-peak catheter pressure differences and Doppler-derived pressure differences from various transducer positions were compared by linear regression analysis. To evaluate inter- and intraobserver variability, the maximal velocity of 50 of the Doppler recordings was interpreted by a second observer (NHS) and by the first observer (EGF] at a second reading without knowledge of prior interpretations. Interobserver variability was calculated as the difference between the 2 observers’ readings divided by the average of their readings and expressed as a positive percentage. Intraobserver variability was calculated as the difference between the first observer’s 2 readings divided by the average of the z: I seadings and expressed as a positive percentage.g
Swrosternal
FIGURE 1. Left, a lateral right ventricular angiogram showing an “eccentric” poststenotic jet in a patient with valvar pulmonic stenosis. Right, line drawing corresponding to angiogram showing the possible range of directions of a poststenotic jet (shaded cone above pulmonary valve) and the axis of ultrasound from multiple transducer positions.
OF CARDIOLOGY
Chest
notch
wall-
Parasternal
Subcortol-
Diaphragm Right
ventricle
848
MULTIPLE DOPPLER TRANSDUCER
POSITIONS
IN PULMONARY
STENOSIS
Results Of the 48 patients who had nonsimultaneous catheterization and Doppler determinations of pressure difference, 24 had Doppler signals from parasternal and suprasternal, subcostal or apical transducer positions, 21 had parasternal signals alone and 3 had suprasternal signals alone. Of the 24 patients studied from multiple transducer positions, the maximal velocity from the parasternal transducer position was equal to or greater than the maximal velocity of the highest other transducer position in only 12 (50%]. In the other 12 patients, a higher maximal velocity was obtained from the suprasternal, subcostal or apical transducer position. In a right ventricular angiogram obtained from 1 of the patients an anterosuperiorly directed poststenotic jet is demonstrated (Figure 11. Continuous-wave Doppler’signals in this patient confirmed that suprasternal, subcostal and apical transducer positions produced higher maximal velocities that were more predictive of catheterization findings than the parasternal maximal velocity (Figure 2). Linear regression analysis comparing nonsimultaneous Doppler-derived pressure differences using the parasternal transducer position versus catheterization pressure differences showed suboptimal strength of correlation (r = 0.78) with considerable variability (standard error of the estimate f 20.6 mm Hg). More importantly, the least squares line of best fit (y = 1.36x - 12.5) deviated significantly from the line of identity reflecting underestimation of the more severe obstructions (Figure 3, top]. When Doppler-derived pressure difference using the signal of highest maximal velocity [regardless of transducer position] was compared with catheterization pressure difference, the correlation was stronger (r = 0.86) with similar variability (standard error of the estimate f 18.7 mm Hg) but the regression line (y = 1.05x - 3.3) very closely approximated the line of identity (Figure 3, bottom].
Of the 12 patients who underwent simultaneous catheterization and Doppler measurements before valvuloplasty, the parasternal maximal velocity was equal to or greater than the highest maximal velocity from a suprasternal or subcostal transducer position in only 4 (33%). A suprasternal or subcostal transducer position resulted in a higher maximal velocity than the parasternal transducer position in 8 patients (67%, Figure 4). Despite the fact that the measurements in these 12 patients were made simultaneously, linear regression analysis resulted in a weak correlation (r = 0.59) with greater variability (standard error of the estimate f 32.1 mm Hg) and poor approximation of the line of
Parasternal y-
Transducer
1.36x-12.5
R-0.78
Position
SEEk20.6
nr45
180
/
/
100 -
60 4020 . ol!, 0
.
/ 20
,
.
40
, 60
.
,
,
80
,
.
100
, 120
.
,
,
140
, 160
.
. 180
AP-Doppler
Highest Velocity Transducer y-
1.05x
- 3.3
R -0.66
SEE t18.7
Position n-24
1801
k
160 -1 14O120 100 80 60 -t
, . , . , , , . , . , . , .
ok/I,. 0
20
40
60
80
100
120
140
160
1 0
AP-Doppler FIGURE 2. Continuous-wave Doppler signals from multiple transducer positions in the patient shown in Figure 1. Catheterization pressure difference was 154 mm Hg. The parasternal (PARA) maximal velocity is lower and less predictive than the subcostal (SUBC), apical (APEX) or suprasternal notch (SSN) maximal velocities (arrowheads). Note the inverted zero line and scale for the SSN velocities.
FIGURE 3. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Cath), both in mm Hg. At the fop, AP Doppler was obtained using the nonsimultaneous parasternal maximal velocity in 45 patients. Boffom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position in 24 patients. The solid line is the line of best fit. The dashed line is the line of identity.
April 1, 1988
identity [y = 1.36x + 5.1) when the parasternal signal was applied (Figure 5, top). However, the Doppler signals from the transducer position of highest maximal velocity correlated very strongly (r = 0.97) with catheterization measurements. The variability was reduced [standard error of the estimate f 9.4 mm Hg) and the line of best fit (y = 1.09x - 2.7) closely approximated the line of identity (Figure 5, bottom). After valvuloplasty, simultaneous Doppler and catheterization data were obtained in 10 patients. The parasternal signal (obtained in only 9 patients] was equal to or greater than the highest signal from the other position in only 4 patients (44%). The suprasternal or subcostal transducer position produced a higher maximal velocity in 5 patients (56%). An improved predictive accuracy using Doppler signals from the transducer position yielding the highest velocity is confirmed by linear regression analysis (Figure 6). The transducer position that yielded the highest velocity in an individual patient was the same before and after valvuloplasty. Interobserver variability was 3.3 f 4.3% [mean f standard deviation] and intraobserver variability was 3.1 f 3.9%. Linear regression analysis provides an interobserver correlation of y = 1.01x + 0.03, r = 0.97, standard error of the estimate f 0.23 m/s and an intraobserver correlation of y = 0.95x + 0.09, r = 0.98, standard error of the estimate f 0.17 m/s. In no patient was more than a l/4 m/s difference found between the 2 observers’ readings in any view.
THE AMERICAN
JOURNAL
Parasternal Dprpf+prnp+
Transducer Transducer
y = 1.36x + 5.1
R =0.59 = 0.59
Position Position
SEE k32.1 r 32.1
n = 12
I
c 5 Ii
100 80 -
a
,
op. 0
.
20
,
.
40
,
.
60
,
.
80
,
,
100
,
.
120
,
.
140
1 0
AP-Doppler
Highest Velocity y-
Transducer
847
140-
The parasternal transducer position has been the standard portal for the evaluation of valvar pulmonic stenosis by Doppler echocardiography. Oliveira Lima et al7 reported a very high accuracy of prediction of vs. Parasternal
Volume 61
nonsimultaneous catheterization findings by Dopplerestimated pressure difference in 16 patients using only the parasternal transducer position. Kosturakis et a1,8 also using only the parasternal approach, achieved acceptable predictive accuracy in 11 patients although some severe obstructions were underestimated. We were unable to achieve this degree of accuracy using parasternal measurements alone. In contrast, we identified patients in whom the parasternal approach alone had seriously underestimated true pressure difference at subsequent cardiac catheterization and, thus, led to this investigation. Multiple Doppler transducer positions have been used in adults for the evaluation of aortic stenosislO
Discussion
Highest
OF CARDIOLOGY
1.09x-2.7
Transducer R-O.97
Position
SEEA9.4
n-12
160
Position
140
160 -. 140 J
120 -
S d Ii. a
100 -
80 60
80 60 40'L
,
0
20
.
,
40
.,
.
60
I
80
.,
.,
100
.,
120
.
140
160
AP-Doppler
Parasternal
Cath
Highest
FIGURE 4. Patient-by-patient comparison of pressure difference (AP) predicted by the parasternal maximal velocity (leff), the highest maximal velocity regardless of transducer position (right) or measured at catheterization (Cath, middle). Individual patient values are connected by lines.
FIGURE 5. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Calh), both in mm Hg, before valvuloplasty. Top, AP Doppler was obtained using the simultaneous parasternal maximal velocity. Bottom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position. The solid line is the line of best fit. The dashed line is the line of identity.
a48
MULTIPLE DOPPLER TRANSDUCER
POSITIONS
IN PULMONARY
and nonstenotic pulmonaryll and aorti$ blood flow. Williams et allo reported improved assessment of aortic stenosis using 3 transducer positions, although the highest of 3 signals was only marginally more accurate than the apical signal alone. Stevenson and KawaborP3 used both parasternal and subcostal approaches in applying modified pulsed Doppler systems to high velocity poststenotic flow in children with pulmonic stenosis but did not comment on the relative accuracy of the 2 transducer positions. Johnson et alI4 measured maximal velocities from the parasternal and subcostal approaches in 12 children with valvar pulmonic stenosis, but the subcostal signal substantially improved the prediction of catheterization pressure difference in Parasternal
Transducer
y= 2.95x-43.6
.‘,
ok 0
. 20
R=0.81
,
Position
SEEi14.5
,
.
.
60
40
n=9
,
,
1
100
80
AP-Doppler Highest
Velocity
y = 0.91x + 0.6
0
20
40
Transducer
R = 0.67
Position
SEE + 13.8
60
80
n = 10
100
AP-Doppler FIGURE 6. Linear regression plots comparing Doppler-derived pressure difference (AP Doppler) with catheterization pressure difference (AP Cath), both in mm Hg, after vaivuiopiasty. At fop, AP Doppler was obtained using the simultaneous parasternai maximal velocity. Bottom, AP Doppler was obtained using the highest maximal velocity regardless of transducer position. The so/id line is the line of best fit. The dashed line is the line of identity.
STENOSIS
only 1 patient. Murphy et alI5 also have supported the use of subcostal imaging but have suggested that the suprasternal position does not provide adequate alignment of the ultrasound beam. Based on these studies, the value of multiple transducer positions in valvar pulmonic stenosis had not previously been conclusively demonstrated. Our results, however, demonstrate superior prediction of catheterization pressure differences by applying the Bernoulli equation to the highest velocity signal obtained from multiple transducer positions. The direction of the poststenotic jet in valvar pulmonic stenosis is unknown and varies from patient to patient; thus, the transducer position that will produce the closest alignment with this jet and the highest velocity cannot be anticipated. In our 10 patients in whom echocardiographic studies were performed before and after valvuloplasty, the transducer position producing the highest velocity in an individual was not altered. However, because our sample size is small, we cannot generalize to the population at large. The direction of the poststenotic jet might be altered by valvuloplasty in some patients. For these reasons, multiple transducer positions should be used in each patient before and after valvuloplasty. Several other findings of our study are notable. The highest Doppler velocities recorded at cardiac catheterization (Figure 5, bottom) more accurately predicted catheterization pressure difference than the highest velocities not recorded simultaneously (Figure 3, bottom]. This may be due to the level of sedation and transvalvar flow rate, which were nearly identical in the former group and highly variable in the latter. A poor correlation was found using the parasternal velocities obtained at cardiac catheterization. [Figure 5, top), which might be related to technical limitations of performing Doppler echocardiography in the catheterization laboratory where optimal patient position and time constraints may have had an adverse impact on the Doppler study. In 7 of 24 patients (29%] with nonsimultaneous measurements (Figure 3), Doppler-derived pressure difference overestimated catheterization findings by at least 10 mm Hg. Level of arousal and, thus, cardiac output are likely to be greater in unsedated outpatients than in sedated patients undergoing catheterization. Doppler-derived measurements are known to overestimate simultaneous peak-to-peak catheterization measurements and to correlate more closely with peak instantaneous pressure differences.16r17Nonetheless, we and others have found excellent predictive accuracy with peak-to-peak measurements without overestimation.7~8~13~14 Accurate Doppler-derived estimates of severity especially in the range of values where treatment may be contemplated are valuable because they influence timing of balloon valvuloplasty and aid in the accurate follow-up thereafter.
References 1. Hatle L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea ~ and Febiger, 1986:24. 2. Valdes-Cruz LM, Horowitz S, Sahn DJ, Larson D, Olive& Lima C, Mesel E. Validation of a Doppler echocardiographic method for calculating severity of discrete stenotic obstructions in a canine preparation with a pulmonary arte-
April 1, 1988
rial bond. Circulation 1984;69:li77-1181. 3. Ha& L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea and Febiger, 1985:108-110. 4. Gardin JM, Burn CS, Childs WJ, Henry WL. Evaluation of blood flow velocity in the ascending aorta and main pulmonary artery of normal subjects by Doppler echocardiography. Am Heart r 1984;107:310-319. 5. Wilson N, Goldberg S], Dickinson DF, Scott 0. Normal intracardiac and great artery blood velocity measurements by pulsed Doppler echocardiography. Br Heart / X985$3:451-458, 6. Grenadier E, Oliveira Lima C, Allen HD, Sahn DJ, Barron JV, Valdes-Cruz LM, Goldberg SJ. Normal intracardiac and great vessel Doppler flow velocities in infants and children. [ACC 1984;4:343-350. 7. Oliveira Lima C, Sahn DJ, Valdes-Cruz LM, Goldberg SJ,Barron JV, Allen HD, Grenadier E. Noninvasive prediction of transvalvular pressure gradient in patients with pulmonary stenosis by quantitative two-dimensional echocardiographic Doppler studies. Circulation 1983;67:866-871. 8. Kosturakis D, Allen HD, Goldberg SJ,Sahn DJ, Valdes-Cruz LM. Noninvasive quantification ofstenotic semilunar valve areas by Doppler echocardiography. JACC 1984;3:1256-1262. 9. Gardin JM, Dabestani A, Matin K, Allfie A, Russell D, Henry WL. Reproducibility of Doppler aortic blood flow measurements: studies on intraobserver, interobserver and day-to-day variability in normal subjects. Am J Cardiol 1984;54:1092-1098, 10. Williams GA, Labovitz AJ, Nelson JG, Kennedy HL. Value of multiple
THE AMERICAN
JOURNAL
OF CARDIOLOGY
Volume 61
049
echocardiographic views in the evaluation of oortic stenosis in adults by continuous-wave Doppler. Am J Cardiol 1985;55:445-449, 11. Lighty GW, Gargiulo A, Kronzon I, Politzer F. Comparison of multiple views for the evaluation of pulmonary arterial blood flow Doppler echocordiography. Circulation 1986;74:1002-1006, 12. Vijayaraghavan G, Singham KT, Tei C, Wong AL, Shah PM. Aortic flow velocity in older subjects using continuous-wave Doppler echocardiography. Am
Heart
T 1984:107:1275.
13. Stevenson JG, Kawahori I. Noninvasive determination of pressure gradients in children: two methods employing pulsed Doppler echocardiography. IACC
1984;3:179-192.
14. Johnson GL, Kwan OL, Handshoe S, Noonan I, DeMaria AN. Accuracy of combined two-dimensional echocardiography and continuous wave Doppler recordings in the estimation of pressure gradient in right ventricular outlet obstruction. JACC 1984;3:1013-1018. 15. Murphy DJ, Ludomirsky A, Danford AD, Huhta ]C. Doppler echocardiography in pulmonary stenosis. Echocardiography 1987;4:187-202. 16. Currie PJ,Seward JB, Reeder GS, Vlietstra RE, Bresnahan DR, Bresnahan JF, Smith HC, Hagler DJ, Tajik AJ. Continuous-wave Doppler echocardiographic assessmentof severity of calcific aortic stenosis: a simultaneous Doppler-catheter correlative study in 100 adult patients. Circulation 1985;71:11621169. 17. Hatle L, Angelsen B. Doppler Ultrasound in Cardiology. Philadelphia: Lea and Febiger, 1985:130-132.