- Email: [email protected]
Doppler echocardiographic estimation of systolic pulmonary artery pressure in patients with aortic-pulmonary shunts
880
lACC Vol 7. No .j Apnl ln6 gX(}-)
Doppler Echocardiographic Estimation of Systolic Pulmonary Artery Pressure in Patients With Aortic-Pulmonary Shunts GERALD R. MARX, MD, FACC, HUGH D. ALLEN, MD, FACC, STANLEYJ.GOLDBERG,MD,FACC Tucson. Arizona
The objective of this study was to determine if the pres• sure drop across various types of aortic-pulmonary shunts could be accurately estimated by Doppler echocardiog• raphy, and if systolic pulmonary pressure could be es• timated by referencing the pressure drop across the aor• tic-pulmonary shunt to systolic systemic arterial pressure measured by cuff sphygmomanometry. This was done in 22 patients and Doppler results were compared with pulmonary artery pressure measured directly by strain gauge manometry. Adequate Doppler waveforms were obtained in 21 of 22 patients; 3 had a Waterston shunt, 10 had a Blalock-Taussig shunt, 1 had a left pulmonary artery-aortic anastomosis, 6 had a patent ductus arte• riosus and 1 had an aortic-pulmonary window. Systolic pulmonary artery pressure estimated by Doppler echocardiography ranged from 12 to 90 mm Hg
Application of the modified Bernoulli equation to jet ve• locities measured by Doppler echocardiography allows mea• surement of pressure drops across stenotic orifices (1-5). This method has been demonstrated to accurately predict pressure drops across stenotic aortic and pulmonary valves (1,6-12) and restrictive ventricular septal defects (1,13). Prior work (l, 14) has shown that referencing pres• sure drops to systolic systemic arterial pressure permits an estimation of systolic pulmonary arterial pressure in patients with an interventricular communication. Doppler detection of jet velocities across aortic-pulmo• nary shunts has been reported (I). The objective of this study was to determine if the pressure drop across various types of aortic-pulmonary shunts could be accurately esti• mated by using Doppler echocardiography, and if systolic pulmonary artery pressure could be estimated by referencing the pressure drop across the aortic-pulmonary shunt to sysFrom the Department of Pediatncs. University of Anzona. Health SCience, Center, Tucson, Anzona. Manuscript received July 16, 1985; revised manuscnpt received No• vember 6. 1985; accepted November 19, 1985 Address for reprints: Gerald R. Marx, MD. Department of Pediatnc;, University of Anzona, Health Sciences Center. Tucson, Arizona 85724 © 1986 by the Amencan College of CardIology
(mean 41.3 ± 21.4 [SO», and measured by strain gauge manometry ranged from 20 to 90 mm Hg (mean 44.7 ± 20.7) (p = NS, r = 0.94, SEE = 7.4 mm Hg; slope = 0.90, y intercept = 7.4 mm Hg). Systolic pulmonary artery to aortic pressure ratios predicted by Doppler recording ranged from 0.1 to 1.0 (mean 0.4 ± 0.2 [SO»; when calculated from direct measurement it ranged from 0.2 to 1.0 (mean 0.4 ± 0.2) (p = NS, r = 0.92; SEE = 0.08, slope = 0.80, y intercept = 0.09). This study demonstrates that Doppler echocardiog• raphy provides an estimation of pressure drop across aortic-pulmonary shunts, and that the data can be used to estimate systolic pulmonary artery pressure by sub• tracting the estimated pressure drop from the systolic systemic arterial pressure. (J Am Coll CardioI1986;7:880-5)
tolic systemic arterial pressure measured by cuff sphyg• momanometry. In this prospective study conducted in pa• tients with aortic-pulmonary shunts, Doppler results were compared with pulmonary artery pressures measured di• rectly by strain gauge manometry.
Methods Study patients. The population consisted of all infants. children and adolescents with an aortic-pulmonary shunt who underwent cardiac catheterization at the University of Arizona from March 1984 to September 1985. This group of 18 patients included 10 patients with a Blalock-Taussig shunt, 3 with a Waterston shunt, 1 with a left pulmonary artery-aortic anastomosis, I with aortic-pulmonary window and 1 with a central Gore-tex shunt. Two patients with patent ductus arteriosus who were catheterized for associated le• sions were also included. The study population also included four other patients with isolated patent ductus arteriosus. Because these patients with isolated patent ductus arteriosus did not undergo catheterization, their pulmonary artery pres• sures were measured at surgery. 0735-1097/86/$3 50
JACC Vol 7, No 4 Apnl 1986 gHO--S
Doppler echocardiographic procedures. Systolic sys• temic arterial pressure was measured with an appropriately sized cuff in the right arm with the patient supine, except for patients who had a right Blalock-Taussig shunt. In these, systemic pressure measurement was obtained in the left arm. Pressure measurement for infants was aided by detecting the radial pulse with a peripheral vascular continuous wave Doppler instrument. Doppler echocardiograms were performed either during cardiac catheterization ( 13 of 18 patients) or within 24 hours prior to cardiac catheterization (5 of 18 patients). Doppler study was done 24 hours before operation in three of the four patients who had surgical closure of a patent ductus alteriosus, and 1 month before surgery in one other patient. Estimation of pulmonary artery pressure by Doppler re• cording was done and data were logged for all patients without knowledge of the direct pressure measurement ob• tained at cathetenzation or surgery. A real-time two-dimensional imaging study was done to assess anatomy in all patients including a standard pulsed Doppler examination (Biosound UltraImager) to assess ve• locities distal to each valve, across a ventricular septal de• fect, and through an aortic-pulmonary shunt. When pulsed Doppler velocities exceeded the Nyquist limit and aliased, a continuous wave system (Irex) was used which allowed measurement of velocities of 7 mls or higher. Pressure drop across aortic-pulmonary shunts was mea• sured by continuous wave Doppler in all patients. The trans• ducer was placed either in the suprasternal notch or supra• clavicular area and directed inferiorly in a nearly vertical direction for those with a Blalock-Taussig shunt. The trans• ducer was placed along the left parasternal border and aimed laterally, posteriorly and to the patient's right side in those with a Waterston shunt. The transducer was aimed toward the patient's left side in the one patient with a left pulmo• nary-aortic anastomosis. Optimal continuous wave Doppler waveforms were obtained from the suprasternal notch by aiming the transducer inferiorly and toward the patient'~ left or from the left parasternal border aiming the Doppler beam posteriorly and leftward for patients with a patent ductus arteriosus. Best waveforms were obtained from the left ster• nal border with the transducer aimed posteriorly and slightly inferiorly for the one patient with an aortic-pulmonary win• dow. More than one transducer location was used to obtain the optimal waveforms in some patients. Computation of Doppler pressure drop. The nonimag• ing transducer was aligned with the jet to the extent possible, and in this situation, Doppler waveforms with the highest peak velocities had a distinct perimeter with minimal fre• quency dispersion. The peak amplitude of the recorded ve• locity profiles was measured by hand calipers and aortic to pulmonary artery pressure drop was estimated according to the modified Bernoulli equation (\): P = 4V 2 , where P = pressure drop in millimeters of mercury and V = velocity
MARX ET AL. DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
881
in meters per second. In the complete equation, P == 4 (V~ - V~), V 2 is velocity distal to the area of obstruction, and V I is velocity proximal to the obstruction, Velocities proximal to the area of obstruction could not be measured in these patients because of the difficulty in accurately po• sitioning the pulsed Doppler sample volume in the exact preobstruction site and thus were not used in the equation. When distal peak velocities were less than 1.5 mis, the pressure drop was considered to be negligible and no pres• sure gradient was calculated. Systolic pulmonary artery pressure was calculated as the difference between systolic systemic arterial pressure measured by cuff sphygmoman• ometry and the calculated pressure drop, Pressure measurements at catheterization. Aortic and pulmonary artery pressures were measured at cardiac cath• eterization using fluid-filled catheters coupled to 23 dB transducers. Pulmonary artery pressure was measured di• rectly at surgery by needle puncture, and simultaneous aortic pressure was measured from the radial artery. Aortic-pul• monary pressure difference was measured at catheterization as the peak to peak pullback gradient, or as the simultaneous peak gradient between the pulmonary artery and radial artery at surgery. Statistics. Systolic pulmonary artery pressure estimated by Doppler echocardiography was compared with systolic pulmonary artery pressure measured at catheterization or during surgery using paired t testing and correlation anal• ysis. Additionally, the systolic pulmonary artery to aortic pressure ratio estimated by Doppler recording was compared by paired t testing and correlation analysis with the ratio measured at catheterization or surgery to correct for differ• ences in pulmonary and systemic arterial pressure that could have been temporally related.
Results Patients. Twenty-two patients were enrolled in the study from March 1984 to September 1985. Adequate aortic-pul• monary artery Doppler waveforms could be obtained for pressure drop estimation in 21 patients ranging in age from 2 weeks to 22 years (mean 4.5 ± 5,6 [SD]) (Table 1). Adequate Doppler waveforms could not be obtained in one 18 year old severely cyanotic girl who had pulmonary atresia and a severely restricted central Gore-tex shunt. Represen• tative waveforms used in this analysis are shown in Figures 1 through 3. Pressure drop comparison. Systolic pulmonary artery pressure estimated by Doppler echocardiography ranged from 12 to 90 mm Hg (mean 41.3 ± 21.4 [SD]); when measured by strain gauge manometry at catheterization or surgery, it ranged from 20 to 90 mm Hg (mean 44.7 ± 20,7) (r = 0.94, SEE = 7.4 mm Hg; slope = 0.90, Y intercept = 7.4 mm Hg) (Fig. 4). Mean pressures were not significantly different. Systolic pulmonary to aortic pressure ratios pre-
00 00 N
01:::
0>• "'»:l
;>Ix tTl tTl
»:l...,
"'>• C:r
r· I:::
Table 1. Patient Data Doppler Study Case 2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17 18 19 20 21 Mean ±SD
Type of Shunt W W W LPA-AO L BT RBT RBT RBT R BT RBT R BT R BT R BT RBT PDA PDA PDA PDA PDA PDA AP window
Age 18 yr 7Y2 yr 22 yr 40 mo 5 yr 13 mo 3Y2 yr 1[12 yr 4 yr 4 mo 3[12 yr 6 yr I yr I Y2 yr 9 rno 4[12 yr 2 yr IY, yr 7 yr 9 mo 2 wk 4.5 yr ±5.6 yr
~
Cardiac Catheterization
BP (mm Hg)
Shunt Vel (rn/s)
Shunt Grad (mm Hg)
PAP (mmHg)
PAPI AOP
PAP (mm Hg)
AOP (mm Hg)
PAPI
AOP
Comments
110 86 90 106 105 94 90 94 90 104 90 90 110 88 80 105 86 96 90 90 70
3.2 4.3 3.7 2.0 3.9 3.4 4.0 4.0 2.0 3.9 4.0 3.7 3.5 4.0 3.0 4.7 3.4 3.9 4.3 4.2 1.4
41 74 55 16 61 46 64 64 16 61 64 55 49 64 36 88 46 61 74 71 0
69 12 35 90 44 48 26 30 74 43 26 35 61 24 44
0.6 0.1 0.4 0.9 0.4 0.5 0.3 0.3 0.8 0.4 0.3 0.4 0.6 0.3 0.6 0.2 0.6 0.4 0.2 0.2 1.0 0.4 ±0.2
80 30 30 90 40 50 28 38 68 40 24 30 68 20 45 30 50 37 36 24 80 44.7 ±20.7
118 100 98 120 110 110 95 115 95 110 105 100 110 100 95 100 110 100 95 80 80
0.7 0.3* 0.3* 0.8 0.4* 0.5* 0.3 0.3* 0.7* 0.4* 0.2 0.3* 0.6* 0.2* 0.5 0.3 0.5* 0.4 0.4 0.3 1.0* 0.4 ±0.2
Obstructed RP A/MPA
17
50 35 16 19 70 41.3 ±21.4
>~ >• »:l
~-<
ill ~
en
c: »:l tTl
I:::
::;en
Absent LPA Obstructed RP A/MPA
c: »:l tTl
I:::
~
All pressures are systolic. *Doppler studies performed during catheterizatIOn. AOP = aortic pressure; AP = aortic-pulmonary; BP = blood pressure; BT = Blalock-Taussig shunt; Grad = gradIent; L = left; LPA = left pulmonary artery; LPA-AO = left pulmonary artery to aortic anastomosis; MPA = main pulmonary artery; PAP = pulmonary artery pressure; PDA = patent ductus arteriosus; R = right; RPA = right pulmonary artery; Vel = velocity; W = Waterston.
;;
i~ _0
'00-
&;z ""
0--.)
?o
v. ..,.
,
,
\
........"._'{"'w.. ""-'.,01'- . .,..-..,...... ,.,
~,;n., -~'l'" ,
CO~y HOf
,
2
Q~ER.I,
ee •
.
,
IREX 2D,
MA« RA«GE
CA\..' 2.0"1$
I
883
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
lACC Vol. 7, No 4 Apnl 1986:880-5
IR~X
D~TEI
DOPPLE~
!8-ee~ee
Figure 1. Patient 8. Doppler tracing across a Blalock-Taussig shunt from the suprasternal notch. The arrow shows where the peak velocity was measured. The calculated pressure drop was 64 mm Hg and the measured gradient was 77 mm Hg. The Doppler• estimated systolic pulmonary artery pressure was 30 mm Hg with a systolic to aortic pressure ratio of 0.3. Systolic pulmonary artery pressure by direct measurement at catheterization was 38 mm Hg with a pressure ratio of 0.3.
dicted by Doppler technique ranged from 0.1 to 1.0 (mean 0.4 ± 0.2); when calculated from catheterization or surgi• cal data, ratios ranged from 0.2 to 1.0 (mean 0.4 ± 0.2) (r = 0.92, SEE = 0.08; slope = 0.80, Y intercept = 0.09) (Fig. 5). Mean ratios were not significantly different.
Discussion This study demonstrates that Doppler echocardiography can estimate pressure drop across aortic-pulmonary shunts and that the data can be used to estimate systolic pulmonary artery pressure by subtracting the estimated pressure drop from systolic systemic arterial pressure. Adequate studies could be obtained in 21 of 22 prospectively evaluated pa• tients. Neither age nor patient size in our group was a lim• iting factor for interrogating the jet velocities. Methodologic considerations. Although echocardiog• raphy did characterize the type and anatomic location of the shunts, optimal Doppler waveforms for measurement could be obtained only with continuous wave Doppler. Depending on shunt type, the continuous Doppler transducer was placed on the chest wall or in the suprasternal notch and then aligned with the jet as recognized by optimal waveforms and audio signals. The small transducer size allowed easy placement, direct contact with the chest and movement in all planes. The wide beam of the continuous wave transducer
probably aided in proper jet alignment. Waveforms were considered optimal when the highest velocities with a clear distinct envelope were obtained during both systole and diastole for a minimum of three beats. The time to obtain the shunt velocities was approximately 10 minutes. Diffi• culty arose in being confident that the peak velocities had been obtained. Optimal waveforms in a discrete obstruction across a semilunar valve have a distinct envelope with con• centration of the peak velocities on the perimeter of the time-velocity curve, and the audio signal has a clear, high pitched tone. Although waveforms across the aortic-pul• monary shunts had a distinct envelope, even when the gray scale of the spectral analysis was optimized, concentration of the peak velocities on the perimeter of the curve could rarely be appreciated. The audio signals had a turbulent continuous quality, unlike the audio signals in discrete semi• lunar valve obstructions. Inability to accurately distinguish and obtain peak velocities in shunt patients may have re• sulted in underestimation of the gradient and therefore over• estimation of the systolic pulmonary artery pressure could have resulted. Preobstruction velocities (V I) could not be obtained by pulsed wave Doppler technique in these patients, and there• fore were not included in the Bernoulli equation. Exclusion of the proximal velocities in patients with a low to moderate gradient may have resulted in gradient overestimation. No gradient was calculated if the peak velocities were less than 150 cmls in this study. At most, this would result Figure 2. Patient 16. Doppler tracings across a patent ductus arteriosus from the left parasternal border. The arrow shows where the peak velocity was measured. The calculated pressure drop was 88 mm Hg, and the measured gradient was 70 mm Hg. The Dop• pler-estimated systolic pulmonary artery pressure was 17 mm Hg, and the systolic pulmonary artery to aortic pressure ratio was 0.2. Systolic pulmonary artery pressure at surgery was 30 mm Hg, with a pressure ratio of 0.3
•
•
•
•
•
•
•
•
•
•
•
•
884
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
JACC Vol 7, No 4 Apnl 1986880-5
a:
8
w
~ to..
oC
..
4
r=O.92 SEE=O.08
2
2
4
6
8
CATH/SURGERY Figure 5. Regression analysis comparing Doppler estimated sys• tolic pulmonary artery to aortic pressure ratio (PA/Ao) with that measured at cardiac catheterization (CA TH) or surgery. d
,"
,
! .,
.t
"·1, ,lRfr, I"
I_EM eD/DOPPlE_
~"
IrT~' ~.-., ...
,
CO"TIMUDUS "Olt
I, lDl~71
,I
,I
Figure 3. Patient 2. Doppler tracings across a Waterston shunt from the left parasternal border aiming the transducer to the pa• tient's right. The arrow shows where the peak velocity was mea• sured. The Doppler-calculated pressure drop was 74 mm Hg. The measured gradient was 70 mm Hg. The Doppler-estimated systolic pulmonary pressure was 12 mm Hg and the systolic pulmonary to aortic pressure ratio was 0.1. Direct measurement at catheterization showed a systolic pulmonary artery pressure of 30 mm Hg with a pressure ratio of 0.3.
in an 8 mm Hg gradient underestimation. If the preobstruc• tion velocities approached 100 cm/s, as might occur in most cases, and the postobstruction velocities were 150 cm/s, then the gradient would calculate to only 4 mm Hg. Although 13 of the patients had Doppler estimation of the systolic pulmonary pressure at catheterization, we did not attempt to match Doppler-estimated maximal instanta• neous gradients to peak to peak gradients measured by pres-
Figure 4. Regression analysis comparing Doppler-estimated sys• tolic pulmonary artery pressure (y axis) with that measured at cardiac catheterization (CATH) or surgery (x axis) in 21 patients. All values are in millimeters of mercury. 100
a:
80
W ..J >;n
a.. -a.. 0 c 40
r=O.94 SEE=7.4
20 0
0
20
40
60
CATH/SURGERY
80
100
sure manometry. Catheterization gradients were measured from records obtained by pullback rather than by dual cath• eterization technique. Thus, instantaneous gradient records were not available for comparative beat to beat measurements. Differences between Doppler estimates and direct measurement of pulmonary artery pressure. The pul• monary artery pressure by Doppler technique was overes• timated by 2 to 6 mm Hg when compared with direct pres• sure measurement in seven patients. Doppler-estimated pulmonary artery pressures were lower than those measured at catheterization or surgery by 2 to 20 mm Hg in 12 patients. The magnitude of underestimation was greater than 10 mm Hg in four patients, two of whom had a patent ductus ar• teriosus. Systolic pulmonary pressure estimated by Doppler technique was 17 and 16 mm Hg, and by direct pressure measurement it was 30 and 36 mm Hg, respectively, in these patients. Since these patients had isolated patent ductus arteriosus, they did not undergo cardiac catheterization; therefore, pulmonary and aortic pressures were measured during surgery and not simultaneously with the Doppler measurements. The two other patients with underestimation of pulmonary pressure greater than 10 mm Hg by Doppler measurement had a Waterston shunt. Systolic pulmonary pressure estimated by Doppler recording was 69 and 12 mm Hg, and when directly measured it was 80 and 30 mm Hg, respectively, in these two patients. The systolic systemic arterial pressure by cuff measurement was 110 mm Hg and by direct measurement was 118 mm Hg in the first patient, and the cuff measurement was 86 mm Hg and direct mea• surement was 100 mm Hg in the second patient. The pres• sure gradient estimated by Doppler recording was 41 and 74 mm Hg, and by direct measurement at catheterization was 38 and 70 mm Hg, respectively. Systolic pulmonary pressure underestimation by the Doppler method was in part related to the low systolic systemic arterial pressure mea• sured by cuff sphygmomanometry in these two patients. The pUlmonary arteries of four patients in whom shunts
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
JACC Vol. 7. No 4 April 1986.880--5
had been attached were severely stenotic or discontinuous with the remainder of the pulmonary artery tree. In these patients, the Doppler method provided good estimation of pulmonary artery pressures distal to the shunt, but did not reflect pressure in the remainder of the pulmonary artery circuit. Conclusion. Our data demonstrate that Doppler echo• cardiography can estimate pressure differences across aortic• pulmonary shunts. The method allows an estimation of sys• tolic pulmonary artery pressure by subtracting the pressure drop from systolic systemic arterial pressure. If proper align• ment with the jet is obtained, pulmonary artery pressure should not be significantly overestimated. If the Doppler velocities are high with a clear waveform envelope. the shunt is restrictive, and therefore pulmonary artery pressure should be low. This method for estimating systolic pul• monary artery pressure has significant clinical utility in the serial evaluation of patients with an aortic-pulmonary shunt. We thank Cheryl Czaplicki for her help In editing and typing the manm,cript and Bill Hanson, Mark Ord and Jeanne Kelter-Marek for thelf help In the cardiac cathetenzation laboratory.
References l. Hatle L, Angelsen B. Doppler Ultrasound In Cardiology. Physical Principles and Clinical Applications. 2nd ed. Philadelphia: Lea & Febiger. 1985:97-152 2. Holen J, Aaslid R, Landmark K. Simonsen S. Detenmnallon of pres• sure gradient In mitral stenosis with a noninvasive ultrasound Doppler technique. Acta Med Scand 1976;199:455-60. 3. Holen 1, Aaslid R, Landmark K, Simonsen S, Ostrem T Determl-
885
nation of effecllve onfice area in mitral stenosIs from noninvasive ultrasound Doppler data and mitral flow rate. Acta Med Scand 1977;201:83-8. 4 Requarth JA. Goldberg SJ, Vasko SD, Allen HD. In vitro verification of Doppler prediction of transvalve pressure gradient and onfice area In steno&is. Am J CardlOl 1984;53:1369-73 5 Vasko SD, Goldberg SJ, Requarth J, Allen HD Factors affecting accuracy of in vitro valvar pressure gradient estimates by Doppler ultrasound Am 1 Cardiol 1984;54:893-6. 6. Halle L. NoninvaSive assessment and differentiation of left ventncular outflow obstruction by Doppler ultrasound. Circulation 1981 ;64: 381-7 7 Hath: L, Angelsen BA, Tromsdal A. NonInvasive assessment of aortic stenosis by Doppler ultrasound. Br Heart J 1980;43:284-92. 8. Lima CO. Sahn DJ, Valdes-Cruz LM, et al. Prediction of the severity of left ventncular outflow tract obstrucllon by quantitative two-di• mem,ional echocardiographic Doppler studies. Circulation 1983;68: 348-54. 9. Berger M, Berdoff RL, GallersteIn PE, Goldberg E. Evaluation of aortIC &tenosis by continuous wave Doppler ultrasound. J Am Coll Cardiol 1984;3:150-5. 10. Stevenson JG. Kawabon 1. Noninvasive determination of pressure gradients in children: two methods emplOYIng pulsed Doppler echo• cardIOgraphy. J Am Coll CardlOl 1984;3: 179-92 II. Goldberg S1, Areias JC, Spitaels SEC, de Villeneuve VH. Echo Dop• pler detectIOn of pulmonary stenosIs by time-Interval histogram anal• YSIS. J Clin Ultrasound 1979:7:183-9.
12 Lima CO. Sahn OJ, Valdes-Cruz LM. et al. Noninvasive predichon of transvalvular pressure gradient in patients With pulmonary stenosis by quantitative two-dimenSIOnal echocardiographlc Doppler studies. Circulation 1983;67:866-71. 13. Hatle L. Rokseth R. Noninvasive diagnosis and assessment of ven• tricular septal defect by Doppler ultrasound Acta Med Scand (Suppi) 1981;645:47-56. 14 Marx GR, Allen HD, Goldberg S1. Doppler echocardiographlc esti• mation of systolic pulmonary artery pressure In pediatric patients with Interventricular communicatIOns. 1 Am Coll Cardiol 1985:6:1132-7.
lACC Vol 7. No .j Apnl ln6 gX(}-)
Doppler Echocardiographic Estimation of Systolic Pulmonary Artery Pressure in Patients With Aortic-Pulmonary Shunts GERALD R. MARX, MD, FACC, HUGH D. ALLEN, MD, FACC, STANLEYJ.GOLDBERG,MD,FACC Tucson. Arizona
The objective of this study was to determine if the pres• sure drop across various types of aortic-pulmonary shunts could be accurately estimated by Doppler echocardiog• raphy, and if systolic pulmonary pressure could be es• timated by referencing the pressure drop across the aor• tic-pulmonary shunt to systolic systemic arterial pressure measured by cuff sphygmomanometry. This was done in 22 patients and Doppler results were compared with pulmonary artery pressure measured directly by strain gauge manometry. Adequate Doppler waveforms were obtained in 21 of 22 patients; 3 had a Waterston shunt, 10 had a Blalock-Taussig shunt, 1 had a left pulmonary artery-aortic anastomosis, 6 had a patent ductus arte• riosus and 1 had an aortic-pulmonary window. Systolic pulmonary artery pressure estimated by Doppler echocardiography ranged from 12 to 90 mm Hg
Application of the modified Bernoulli equation to jet ve• locities measured by Doppler echocardiography allows mea• surement of pressure drops across stenotic orifices (1-5). This method has been demonstrated to accurately predict pressure drops across stenotic aortic and pulmonary valves (1,6-12) and restrictive ventricular septal defects (1,13). Prior work (l, 14) has shown that referencing pres• sure drops to systolic systemic arterial pressure permits an estimation of systolic pulmonary arterial pressure in patients with an interventricular communication. Doppler detection of jet velocities across aortic-pulmo• nary shunts has been reported (I). The objective of this study was to determine if the pressure drop across various types of aortic-pulmonary shunts could be accurately esti• mated by using Doppler echocardiography, and if systolic pulmonary artery pressure could be estimated by referencing the pressure drop across the aortic-pulmonary shunt to sysFrom the Department of Pediatncs. University of Anzona. Health SCience, Center, Tucson, Anzona. Manuscript received July 16, 1985; revised manuscnpt received No• vember 6. 1985; accepted November 19, 1985 Address for reprints: Gerald R. Marx, MD. Department of Pediatnc;, University of Anzona, Health Sciences Center. Tucson, Arizona 85724 © 1986 by the Amencan College of CardIology
(mean 41.3 ± 21.4 [SO», and measured by strain gauge manometry ranged from 20 to 90 mm Hg (mean 44.7 ± 20.7) (p = NS, r = 0.94, SEE = 7.4 mm Hg; slope = 0.90, y intercept = 7.4 mm Hg). Systolic pulmonary artery to aortic pressure ratios predicted by Doppler recording ranged from 0.1 to 1.0 (mean 0.4 ± 0.2 [SO»; when calculated from direct measurement it ranged from 0.2 to 1.0 (mean 0.4 ± 0.2) (p = NS, r = 0.92; SEE = 0.08, slope = 0.80, y intercept = 0.09). This study demonstrates that Doppler echocardiog• raphy provides an estimation of pressure drop across aortic-pulmonary shunts, and that the data can be used to estimate systolic pulmonary artery pressure by sub• tracting the estimated pressure drop from the systolic systemic arterial pressure. (J Am Coll CardioI1986;7:880-5)
tolic systemic arterial pressure measured by cuff sphyg• momanometry. In this prospective study conducted in pa• tients with aortic-pulmonary shunts, Doppler results were compared with pulmonary artery pressures measured di• rectly by strain gauge manometry.
Methods Study patients. The population consisted of all infants. children and adolescents with an aortic-pulmonary shunt who underwent cardiac catheterization at the University of Arizona from March 1984 to September 1985. This group of 18 patients included 10 patients with a Blalock-Taussig shunt, 3 with a Waterston shunt, 1 with a left pulmonary artery-aortic anastomosis, I with aortic-pulmonary window and 1 with a central Gore-tex shunt. Two patients with patent ductus arteriosus who were catheterized for associated le• sions were also included. The study population also included four other patients with isolated patent ductus arteriosus. Because these patients with isolated patent ductus arteriosus did not undergo catheterization, their pulmonary artery pres• sures were measured at surgery. 0735-1097/86/$3 50
JACC Vol 7, No 4 Apnl 1986 gHO--S
Doppler echocardiographic procedures. Systolic sys• temic arterial pressure was measured with an appropriately sized cuff in the right arm with the patient supine, except for patients who had a right Blalock-Taussig shunt. In these, systemic pressure measurement was obtained in the left arm. Pressure measurement for infants was aided by detecting the radial pulse with a peripheral vascular continuous wave Doppler instrument. Doppler echocardiograms were performed either during cardiac catheterization ( 13 of 18 patients) or within 24 hours prior to cardiac catheterization (5 of 18 patients). Doppler study was done 24 hours before operation in three of the four patients who had surgical closure of a patent ductus alteriosus, and 1 month before surgery in one other patient. Estimation of pulmonary artery pressure by Doppler re• cording was done and data were logged for all patients without knowledge of the direct pressure measurement ob• tained at cathetenzation or surgery. A real-time two-dimensional imaging study was done to assess anatomy in all patients including a standard pulsed Doppler examination (Biosound UltraImager) to assess ve• locities distal to each valve, across a ventricular septal de• fect, and through an aortic-pulmonary shunt. When pulsed Doppler velocities exceeded the Nyquist limit and aliased, a continuous wave system (Irex) was used which allowed measurement of velocities of 7 mls or higher. Pressure drop across aortic-pulmonary shunts was mea• sured by continuous wave Doppler in all patients. The trans• ducer was placed either in the suprasternal notch or supra• clavicular area and directed inferiorly in a nearly vertical direction for those with a Blalock-Taussig shunt. The trans• ducer was placed along the left parasternal border and aimed laterally, posteriorly and to the patient's right side in those with a Waterston shunt. The transducer was aimed toward the patient's left side in the one patient with a left pulmo• nary-aortic anastomosis. Optimal continuous wave Doppler waveforms were obtained from the suprasternal notch by aiming the transducer inferiorly and toward the patient'~ left or from the left parasternal border aiming the Doppler beam posteriorly and leftward for patients with a patent ductus arteriosus. Best waveforms were obtained from the left ster• nal border with the transducer aimed posteriorly and slightly inferiorly for the one patient with an aortic-pulmonary win• dow. More than one transducer location was used to obtain the optimal waveforms in some patients. Computation of Doppler pressure drop. The nonimag• ing transducer was aligned with the jet to the extent possible, and in this situation, Doppler waveforms with the highest peak velocities had a distinct perimeter with minimal fre• quency dispersion. The peak amplitude of the recorded ve• locity profiles was measured by hand calipers and aortic to pulmonary artery pressure drop was estimated according to the modified Bernoulli equation (\): P = 4V 2 , where P = pressure drop in millimeters of mercury and V = velocity
MARX ET AL. DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
881
in meters per second. In the complete equation, P == 4 (V~ - V~), V 2 is velocity distal to the area of obstruction, and V I is velocity proximal to the obstruction, Velocities proximal to the area of obstruction could not be measured in these patients because of the difficulty in accurately po• sitioning the pulsed Doppler sample volume in the exact preobstruction site and thus were not used in the equation. When distal peak velocities were less than 1.5 mis, the pressure drop was considered to be negligible and no pres• sure gradient was calculated. Systolic pulmonary artery pressure was calculated as the difference between systolic systemic arterial pressure measured by cuff sphygmoman• ometry and the calculated pressure drop, Pressure measurements at catheterization. Aortic and pulmonary artery pressures were measured at cardiac cath• eterization using fluid-filled catheters coupled to 23 dB transducers. Pulmonary artery pressure was measured di• rectly at surgery by needle puncture, and simultaneous aortic pressure was measured from the radial artery. Aortic-pul• monary pressure difference was measured at catheterization as the peak to peak pullback gradient, or as the simultaneous peak gradient between the pulmonary artery and radial artery at surgery. Statistics. Systolic pulmonary artery pressure estimated by Doppler echocardiography was compared with systolic pulmonary artery pressure measured at catheterization or during surgery using paired t testing and correlation anal• ysis. Additionally, the systolic pulmonary artery to aortic pressure ratio estimated by Doppler recording was compared by paired t testing and correlation analysis with the ratio measured at catheterization or surgery to correct for differ• ences in pulmonary and systemic arterial pressure that could have been temporally related.
Results Patients. Twenty-two patients were enrolled in the study from March 1984 to September 1985. Adequate aortic-pul• monary artery Doppler waveforms could be obtained for pressure drop estimation in 21 patients ranging in age from 2 weeks to 22 years (mean 4.5 ± 5,6 [SD]) (Table 1). Adequate Doppler waveforms could not be obtained in one 18 year old severely cyanotic girl who had pulmonary atresia and a severely restricted central Gore-tex shunt. Represen• tative waveforms used in this analysis are shown in Figures 1 through 3. Pressure drop comparison. Systolic pulmonary artery pressure estimated by Doppler echocardiography ranged from 12 to 90 mm Hg (mean 41.3 ± 21.4 [SD]); when measured by strain gauge manometry at catheterization or surgery, it ranged from 20 to 90 mm Hg (mean 44.7 ± 20,7) (r = 0.94, SEE = 7.4 mm Hg; slope = 0.90, Y intercept = 7.4 mm Hg) (Fig. 4). Mean pressures were not significantly different. Systolic pulmonary to aortic pressure ratios pre-
00 00 N
01:::
0>• "'»:l
;>Ix tTl tTl
»:l...,
"'>• C:r
r· I:::
Table 1. Patient Data Doppler Study Case 2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17 18 19 20 21 Mean ±SD
Type of Shunt W W W LPA-AO L BT RBT RBT RBT R BT RBT R BT R BT R BT RBT PDA PDA PDA PDA PDA PDA AP window
Age 18 yr 7Y2 yr 22 yr 40 mo 5 yr 13 mo 3Y2 yr 1[12 yr 4 yr 4 mo 3[12 yr 6 yr I yr I Y2 yr 9 rno 4[12 yr 2 yr IY, yr 7 yr 9 mo 2 wk 4.5 yr ±5.6 yr
~
Cardiac Catheterization
BP (mm Hg)
Shunt Vel (rn/s)
Shunt Grad (mm Hg)
PAP (mmHg)
PAPI AOP
PAP (mm Hg)
AOP (mm Hg)
PAPI
AOP
Comments
110 86 90 106 105 94 90 94 90 104 90 90 110 88 80 105 86 96 90 90 70
3.2 4.3 3.7 2.0 3.9 3.4 4.0 4.0 2.0 3.9 4.0 3.7 3.5 4.0 3.0 4.7 3.4 3.9 4.3 4.2 1.4
41 74 55 16 61 46 64 64 16 61 64 55 49 64 36 88 46 61 74 71 0
69 12 35 90 44 48 26 30 74 43 26 35 61 24 44
0.6 0.1 0.4 0.9 0.4 0.5 0.3 0.3 0.8 0.4 0.3 0.4 0.6 0.3 0.6 0.2 0.6 0.4 0.2 0.2 1.0 0.4 ±0.2
80 30 30 90 40 50 28 38 68 40 24 30 68 20 45 30 50 37 36 24 80 44.7 ±20.7
118 100 98 120 110 110 95 115 95 110 105 100 110 100 95 100 110 100 95 80 80
0.7 0.3* 0.3* 0.8 0.4* 0.5* 0.3 0.3* 0.7* 0.4* 0.2 0.3* 0.6* 0.2* 0.5 0.3 0.5* 0.4 0.4 0.3 1.0* 0.4 ±0.2
Obstructed RP A/MPA
17
50 35 16 19 70 41.3 ±21.4
>~ >• »:l
~-<
ill ~
en
c: »:l tTl
I:::
::;en
Absent LPA Obstructed RP A/MPA
c: »:l tTl
I:::
~
All pressures are systolic. *Doppler studies performed during catheterizatIOn. AOP = aortic pressure; AP = aortic-pulmonary; BP = blood pressure; BT = Blalock-Taussig shunt; Grad = gradIent; L = left; LPA = left pulmonary artery; LPA-AO = left pulmonary artery to aortic anastomosis; MPA = main pulmonary artery; PAP = pulmonary artery pressure; PDA = patent ductus arteriosus; R = right; RPA = right pulmonary artery; Vel = velocity; W = Waterston.
;;
i~ _0
'00-
&;z ""
0--.)
?o
v. ..,.
,
,
\
........"._'{"'w.. ""-'.,01'- . .,..-..,...... ,.,
~,;n., -~'l'" ,
CO~y HOf
,
2
Q~ER.I,
ee •
.
,
IREX 2D,
MA« RA«GE
CA\..' 2.0"1$
I
883
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
lACC Vol. 7, No 4 Apnl 1986:880-5
IR~X
D~TEI
DOPPLE~
!8-ee~ee
Figure 1. Patient 8. Doppler tracing across a Blalock-Taussig shunt from the suprasternal notch. The arrow shows where the peak velocity was measured. The calculated pressure drop was 64 mm Hg and the measured gradient was 77 mm Hg. The Doppler• estimated systolic pulmonary artery pressure was 30 mm Hg with a systolic to aortic pressure ratio of 0.3. Systolic pulmonary artery pressure by direct measurement at catheterization was 38 mm Hg with a pressure ratio of 0.3.
dicted by Doppler technique ranged from 0.1 to 1.0 (mean 0.4 ± 0.2); when calculated from catheterization or surgi• cal data, ratios ranged from 0.2 to 1.0 (mean 0.4 ± 0.2) (r = 0.92, SEE = 0.08; slope = 0.80, Y intercept = 0.09) (Fig. 5). Mean ratios were not significantly different.
Discussion This study demonstrates that Doppler echocardiography can estimate pressure drop across aortic-pulmonary shunts and that the data can be used to estimate systolic pulmonary artery pressure by subtracting the estimated pressure drop from systolic systemic arterial pressure. Adequate studies could be obtained in 21 of 22 prospectively evaluated pa• tients. Neither age nor patient size in our group was a lim• iting factor for interrogating the jet velocities. Methodologic considerations. Although echocardiog• raphy did characterize the type and anatomic location of the shunts, optimal Doppler waveforms for measurement could be obtained only with continuous wave Doppler. Depending on shunt type, the continuous Doppler transducer was placed on the chest wall or in the suprasternal notch and then aligned with the jet as recognized by optimal waveforms and audio signals. The small transducer size allowed easy placement, direct contact with the chest and movement in all planes. The wide beam of the continuous wave transducer
probably aided in proper jet alignment. Waveforms were considered optimal when the highest velocities with a clear distinct envelope were obtained during both systole and diastole for a minimum of three beats. The time to obtain the shunt velocities was approximately 10 minutes. Diffi• culty arose in being confident that the peak velocities had been obtained. Optimal waveforms in a discrete obstruction across a semilunar valve have a distinct envelope with con• centration of the peak velocities on the perimeter of the time-velocity curve, and the audio signal has a clear, high pitched tone. Although waveforms across the aortic-pul• monary shunts had a distinct envelope, even when the gray scale of the spectral analysis was optimized, concentration of the peak velocities on the perimeter of the curve could rarely be appreciated. The audio signals had a turbulent continuous quality, unlike the audio signals in discrete semi• lunar valve obstructions. Inability to accurately distinguish and obtain peak velocities in shunt patients may have re• sulted in underestimation of the gradient and therefore over• estimation of the systolic pulmonary artery pressure could have resulted. Preobstruction velocities (V I) could not be obtained by pulsed wave Doppler technique in these patients, and there• fore were not included in the Bernoulli equation. Exclusion of the proximal velocities in patients with a low to moderate gradient may have resulted in gradient overestimation. No gradient was calculated if the peak velocities were less than 150 cmls in this study. At most, this would result Figure 2. Patient 16. Doppler tracings across a patent ductus arteriosus from the left parasternal border. The arrow shows where the peak velocity was measured. The calculated pressure drop was 88 mm Hg, and the measured gradient was 70 mm Hg. The Dop• pler-estimated systolic pulmonary artery pressure was 17 mm Hg, and the systolic pulmonary artery to aortic pressure ratio was 0.2. Systolic pulmonary artery pressure at surgery was 30 mm Hg, with a pressure ratio of 0.3
•
•
•
•
•
•
•
•
•
•
•
•
884
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
JACC Vol 7, No 4 Apnl 1986880-5
a:
8
w
~ to..
oC
..
4
r=O.92 SEE=O.08
2
2
4
6
8
CATH/SURGERY Figure 5. Regression analysis comparing Doppler estimated sys• tolic pulmonary artery to aortic pressure ratio (PA/Ao) with that measured at cardiac catheterization (CA TH) or surgery. d
,"
,
! .,
.t
"·1, ,lRfr, I"
I_EM eD/DOPPlE_
~"
IrT~' ~.-., ...
,
CO"TIMUDUS "Olt
I, lDl~71
,I
,I
Figure 3. Patient 2. Doppler tracings across a Waterston shunt from the left parasternal border aiming the transducer to the pa• tient's right. The arrow shows where the peak velocity was mea• sured. The Doppler-calculated pressure drop was 74 mm Hg. The measured gradient was 70 mm Hg. The Doppler-estimated systolic pulmonary pressure was 12 mm Hg and the systolic pulmonary to aortic pressure ratio was 0.1. Direct measurement at catheterization showed a systolic pulmonary artery pressure of 30 mm Hg with a pressure ratio of 0.3.
in an 8 mm Hg gradient underestimation. If the preobstruc• tion velocities approached 100 cm/s, as might occur in most cases, and the postobstruction velocities were 150 cm/s, then the gradient would calculate to only 4 mm Hg. Although 13 of the patients had Doppler estimation of the systolic pulmonary pressure at catheterization, we did not attempt to match Doppler-estimated maximal instanta• neous gradients to peak to peak gradients measured by pres-
Figure 4. Regression analysis comparing Doppler-estimated sys• tolic pulmonary artery pressure (y axis) with that measured at cardiac catheterization (CATH) or surgery (x axis) in 21 patients. All values are in millimeters of mercury. 100
a:
80
W ..J >;n
a.. -a.. 0 c 40
r=O.94 SEE=7.4
20 0
0
20
40
60
CATH/SURGERY
80
100
sure manometry. Catheterization gradients were measured from records obtained by pullback rather than by dual cath• eterization technique. Thus, instantaneous gradient records were not available for comparative beat to beat measurements. Differences between Doppler estimates and direct measurement of pulmonary artery pressure. The pul• monary artery pressure by Doppler technique was overes• timated by 2 to 6 mm Hg when compared with direct pres• sure measurement in seven patients. Doppler-estimated pulmonary artery pressures were lower than those measured at catheterization or surgery by 2 to 20 mm Hg in 12 patients. The magnitude of underestimation was greater than 10 mm Hg in four patients, two of whom had a patent ductus ar• teriosus. Systolic pulmonary pressure estimated by Doppler technique was 17 and 16 mm Hg, and by direct pressure measurement it was 30 and 36 mm Hg, respectively, in these patients. Since these patients had isolated patent ductus arteriosus, they did not undergo cardiac catheterization; therefore, pulmonary and aortic pressures were measured during surgery and not simultaneously with the Doppler measurements. The two other patients with underestimation of pulmonary pressure greater than 10 mm Hg by Doppler measurement had a Waterston shunt. Systolic pulmonary pressure estimated by Doppler recording was 69 and 12 mm Hg, and when directly measured it was 80 and 30 mm Hg, respectively, in these two patients. The systolic systemic arterial pressure by cuff measurement was 110 mm Hg and by direct measurement was 118 mm Hg in the first patient, and the cuff measurement was 86 mm Hg and direct mea• surement was 100 mm Hg in the second patient. The pres• sure gradient estimated by Doppler recording was 41 and 74 mm Hg, and by direct measurement at catheterization was 38 and 70 mm Hg, respectively. Systolic pulmonary pressure underestimation by the Doppler method was in part related to the low systolic systemic arterial pressure mea• sured by cuff sphygmomanometry in these two patients. The pUlmonary arteries of four patients in whom shunts
MARX ET AL DOPPLER PULMONARY ARTERY PRESSURE MEASUREMENT
JACC Vol. 7. No 4 April 1986.880--5
had been attached were severely stenotic or discontinuous with the remainder of the pulmonary artery tree. In these patients, the Doppler method provided good estimation of pulmonary artery pressures distal to the shunt, but did not reflect pressure in the remainder of the pulmonary artery circuit. Conclusion. Our data demonstrate that Doppler echo• cardiography can estimate pressure differences across aortic• pulmonary shunts. The method allows an estimation of sys• tolic pulmonary artery pressure by subtracting the pressure drop from systolic systemic arterial pressure. If proper align• ment with the jet is obtained, pulmonary artery pressure should not be significantly overestimated. If the Doppler velocities are high with a clear waveform envelope. the shunt is restrictive, and therefore pulmonary artery pressure should be low. This method for estimating systolic pul• monary artery pressure has significant clinical utility in the serial evaluation of patients with an aortic-pulmonary shunt. We thank Cheryl Czaplicki for her help In editing and typing the manm,cript and Bill Hanson, Mark Ord and Jeanne Kelter-Marek for thelf help In the cardiac cathetenzation laboratory.
References l. Hatle L, Angelsen B. Doppler Ultrasound In Cardiology. Physical Principles and Clinical Applications. 2nd ed. Philadelphia: Lea & Febiger. 1985:97-152 2. Holen J, Aaslid R, Landmark K. Simonsen S. Detenmnallon of pres• sure gradient In mitral stenosis with a noninvasive ultrasound Doppler technique. Acta Med Scand 1976;199:455-60. 3. Holen 1, Aaslid R, Landmark K, Simonsen S, Ostrem T Determl-
885
nation of effecllve onfice area in mitral stenosIs from noninvasive ultrasound Doppler data and mitral flow rate. Acta Med Scand 1977;201:83-8. 4 Requarth JA. Goldberg SJ, Vasko SD, Allen HD. In vitro verification of Doppler prediction of transvalve pressure gradient and onfice area In steno&is. Am J CardlOl 1984;53:1369-73 5 Vasko SD, Goldberg SJ, Requarth J, Allen HD Factors affecting accuracy of in vitro valvar pressure gradient estimates by Doppler ultrasound Am 1 Cardiol 1984;54:893-6. 6. Halle L. NoninvaSive assessment and differentiation of left ventncular outflow obstruction by Doppler ultrasound. Circulation 1981 ;64: 381-7 7 Hath: L, Angelsen BA, Tromsdal A. NonInvasive assessment of aortic stenosis by Doppler ultrasound. Br Heart J 1980;43:284-92. 8. Lima CO. Sahn DJ, Valdes-Cruz LM, et al. Prediction of the severity of left ventncular outflow tract obstrucllon by quantitative two-di• mem,ional echocardiographic Doppler studies. Circulation 1983;68: 348-54. 9. Berger M, Berdoff RL, GallersteIn PE, Goldberg E. Evaluation of aortIC &tenosis by continuous wave Doppler ultrasound. J Am Coll Cardiol 1984;3:150-5. 10. Stevenson JG. Kawabon 1. Noninvasive determination of pressure gradients in children: two methods emplOYIng pulsed Doppler echo• cardIOgraphy. J Am Coll CardlOl 1984;3: 179-92 II. Goldberg S1, Areias JC, Spitaels SEC, de Villeneuve VH. Echo Dop• pler detectIOn of pulmonary stenosIs by time-Interval histogram anal• YSIS. J Clin Ultrasound 1979:7:183-9.
12 Lima CO. Sahn OJ, Valdes-Cruz LM. et al. Noninvasive predichon of transvalvular pressure gradient in patients With pulmonary stenosis by quantitative two-dimenSIOnal echocardiographlc Doppler studies. Circulation 1983;67:866-71. 13. Hatle L. Rokseth R. Noninvasive diagnosis and assessment of ven• tricular septal defect by Doppler ultrasound Acta Med Scand (Suppi) 1981;645:47-56. 14 Marx GR, Allen HD, Goldberg S1. Doppler echocardiographlc esti• mation of systolic pulmonary artery pressure In pediatric patients with Interventricular communicatIOns. 1 Am Coll Cardiol 1985:6:1132-7.