- Email: [email protected]
The Effect of Changes in Afterload on Doppler Echocardiographic Indexes of Left Ventricular Performance
The Effect of Changes in Afterload on Doppler Echocardiographic Indexes of Left Ventricular Performance Kenneth Wallmeyer, MD, L. Samuel Wann, MD, Kiran B. Sagar, MD, Peter Czakanski, MD, John Kalbfleisch, PhD, and H. Sidney Klopfenstein, MD, PhD, Milwaukee, Wis.
To investigate the influence of changes in afterload on Doppler echocardiographic determination of peak aortic blood velocity, mean acceleration, and systolic velocity integral, eight dogs with their chests opened were studied in four inotropic states at varying levels of heart rate and mean aortic blood pressure. Data were collected in the control state, at two different levels of dobutamine administration (5 and 10 ,_..g/kg/min intravenously), and after administration of propranolol (0.5 mg/kg intravenously). In each inotropic state, phenylephrine was infused intravenously to produce at least two successive steady state increases of 10 mm Hg or more in mean aortic blood pressure. Within a given animal, peak velocity emerged as the Doppler index most closely correlated with changes in Q...., dQ/dt, and dP/dt (r = 0.94, 0.91, and 0.89, respectively). Mean acceleration also correlated closely with the invasive indexes (r 0.87, 0.89, and 0.89). The effect of changes in mean aortic blood pressure on Doppler index measurements was not statistically significant in any of the inotropic states and did not affect the closeness of their correlation with the invasive indexes. We conclude that Doppler echocirdiographic measurements of aortic blood peak velocity and mean acceleration remained as sensitive to changes in the inotropic state under conditions of varying increases in afterload as did the conventional invasive indexes tested. (JAM Soc EcHo 1988;1:135-40.)
=
Many pl,lblished reports lend support to the use of Doppler echocardiographic measurements of aortic blood velocity to derive indexes of left ventricular performance. The low cost, noninvasive nature, and the ability to make measurements during exercise make this technique particularly attractive. Clinical studies have demonstrated the sensitivity of Doppler aortic blood velocimetry to exercise-induced left ven tricular dysfunction in patients with coronary artery disease, l-s and the ease of making serial determina-
From the Cardiology Division, Department ofMedicine, Medical College of Wisconsin; and Zablocki Veterans Administration Medical Center. Supported in part by grants from the National Heart, Lung, and Blood Institute, National Institutes of Health, the Veterans Ad ministration, the American Heart Association, and the Heath Foundation. Reprint requests: L. Samuel Wann, MD, Cardiology Division, Medical College of Wisconsin, 8700 W. Wisconsin Avenue, Mil waukee, WI 53226.
tions suggests possible applications in monitoring the effectiveness of therapeutic interventions aimed at improving left ventricular function. Changes in myocardial contractile function have always been difficult to isolate from changes in load ing conditions.- Therefore, as with conventional in dexes, the clinical application of Doppler echocar diographic indexes of performance must be guided by an appreciation of the degree to which they are influenced by changes in preload and afterload. Re cent studies in our laboratory demonstrated that un der conditions of varying preload, Doppler deter mination of peak aortic blood velocity and mean ac celeration was as sensitive to changes in inotropic states as the conventional invasive measurement of peak aortic blood flow, maximum rate of change in aortic blood flow (dQ/dt), and peak left ventricular dP/dt. 6 This study was done to evaluate the influence of changes in afterload on Doppler echocardiogra phic indexes of ventricular performance. 135
Journal of the American Society of Echocardiography
136 Wallmeyer et al.
. ... '· '·
Doppler Afterload r1.0 msec
CONTROL
At..·ji
......
135
122
12.0 msec
LOW DOSE DOBUTAMINE
HIGH DOSE DOBUTAMINE
96
I 1.0 msec PROPIRANOLOL
....... !·'-. "-·"'.
....
86
o~M
~
130
117
I•• · 112 ·-
,. l.j,
I
. . . . .,
124
.i.:
M..
156
139
I 2.0 msec
-' ~·~ ~~ 184
Ja.. A.. •.
,,L....
116
......l..~~
163
172
1.1~ j.,,,"' 148
j
J ol
1111 ol
A• I"'" o
150
Figure 1 Doppler recordings obtained from single animal. Horizontal rows represent four inotropic states. Note change in velocity scale during inotropic stimulation. Numbers below recordings are simultaneous measurements of mean aortic blood pressure (millimeters of mercury).
METHODS
Eight mongrel dogs weighing 24.6 to 33.1 kg were given no food overnight, anesthetized by an intravenous administration of alpha chloralose and urethane, intubated, and ventilated with a volume respirator (Harvard Apparatus Co., Inc., S. Natick, Massachusetts) with air enriched with oxy gen (6 L/min). Polyvinyl catheters (Tygon Micro bore Tubing, 0.05 inch, Norton Performance Plas tics, Wayne, New Jersey) were introduced into both femoral veins for infusion of drugs, and a third cath eter was advanced via a femoral artery to the proximal aorta. A micromanometer-tipped catheter (Millar In struments, Inc., Houston) was inserted into a femoral artery and advanced to the proximal aorta, allowing measurement of mean aortic blood pressure as an indicator of changing afterload. The electrocardio gram was monitored throughout, and arterial blood gas levels were measured at 1-hour intervals and were maintained in the normal range. Thoracotomy was performed in the left fifth in tercostal space, and the position of the aortic cath eters was verified by palpation. A fluid-filled catheter was placed into the left atrium and secured by purse string suture. Aortic and left atrial catheters were connected to fluid-filled Statham P23Db pressure transducers (Gould Inc. Cardiovascular Products, Oxnard, California), with the zero pressure reference point at the middle right atrial level. An electromag netic flow probe (Howell Instrument Co., Camarillo,
California) was placed around the ascending aorta, and blood flow was measured with a N arcomatic flowmeter (model RT-500, Narco Bio-Systems, Houston). A precalibrated high fidelity pressure transducer (Konigsberg Instruments, Inc., Pasadena, California) was inserted into the left ventricle through an apical stab wound, secured with a purse string suture, and referenced to the high fidelity prox imal aortic pressure waveform during systole and the left atrial catheter during diastole. Two sonomicro meter crystals (PZT5A [flat and cylinder], Valpey Fisher Med. Div., Hopkinton, Massachusetts) were inserted into the anterior left ventricular free wall at the midmyocardial level, perpendicular to the long axis of the ventricle, and the distance between them was measured (Hartley Multi-channel Flow Dimension System, Instrumentation Development Lab, Baylor College of Medicine, Houston) as an index of preload. Finally, pacing wires were sutured to the left atrium, and a snare placed around the inferior vena cava was used to attenuate any increases in end-diastolic segment length that occurred because ofincreases produced in mean arterial blood pressure. All hemodynamic measurements were made under steady state conditions with mechanical ventilation briefly suspended to eliminate respiratory variation. Hemodynamic data were recorded on an analog FM tape recorder (A.R. Vetter Co., Rebersburg, Penn sylvania) and on an eight-channel strip chart recorder (model 2800, Gould, Inc., Cleveland). Ten-second data files were subsequently transferred to a digital
Volume l Number 2 March-Aprill988
computer (DEC LSI ll/23) and sequential inter active programs used to calculate peak volumetric blood flow (Qmax), maximum dQ/dt, and maximum left ventricular dP/dt. Simultaneous with the collec tion of hemodynamic data, Doppler measurements of blood velocity were obtained in the continuous mode with a dedicated nonimaging Doppler trans ducer (Pedof, New Brunswick, New Jersey) with a 2 MHz carrier frequency (Irex III B, Johnson & Johnson, New Brunswick, New Jersey) and recorded on a strip-chart recorder with a paper speed of 50 mm/sec. The transducer was handheld and applied to the aortic arch with the investigator's hand stea died against the thoracic cage. The ultrasonic beam was oriented parallel to ascending aortic blood flow, guided by audio and graphic signals to obtain max imal Doppler shift, with velocity envelopes having distinct borders. Doppler measurements of peak blood velocity and mean acceleration were made for six consecutive beats and averaged. The systolic ve locity integral was determined by manual planimetry ofa representative Doppler envelope, applying Simp son's rule. Measurements were obtained initially in the base line contractile state, followed sequentially by measurements in three additional inotropic states produced by low-dosage infusion of dobutamine (5 J.Lglkg/min intravenously), high-dosage infu sion of dobutarnine (10 J.Lg/kg/min intravenously), and finally after administration of propranolol (0.5 mg/kg intravenously, given more than 10 minutes after discontinuation of dobutarnine). In each ino tropic state data were collected at a baseline mean aortic blood pressure, followed by measurements during infusion of phenylephrine at rates sufficient to produce at least two successive steady state in creases of at least 10 mm Hg in mean aortic blood pressure. Whenever the heart rate in sinus rhythm was less than 150 beats/min, data were also collected during rapid atrial pacing (150 beats/min) at the same level of infusion of phenylephrine. Premature beats and the cycles after them were excluded from analysis. Data samples with gross irregularity of rhythm or with pulsus alternans were also excluded. For each animal studied the values for each index of contractility obtained in each inotropic state were averaged. These within-animal means were then an alyzed with a two-way analysis of variance followed by the paired t test to compare means of determi nations in adjacent inotropic states. The correlation of Doppler indexes with the simultaneous invasive index determinations was analyzed by linear regres sion, with the pooled within-animal correlation co-
Effect of changes in afterload 137
3.0
~
C3
o~
-' w
0
(J)
p
< .05
p < .01
p
< .02
< .01
p
< .01
2.0
>~
~.s 1.0
< w a..
z
0
100
~
80 60
a:~
w ~ -' (/)
wo() ~
p < .05
p
40
<"E
(/)
20
z~
< w
~
>
1o~
g.§.. W-'
>< () a: _
200
p
< .05
NS
NS
150 100
-'W
0 I1-Z
oo > CJ)
50 PROPRANOLOL CONTROL VS. DOBUTAMINE (0.5 mg/kg) VS. CONTROL (5 meg/kg/min)
DOBUTAMINE 5 vs. 10 (meg/kg/min)
Doppler index measurements for each inotropic state at all levels of mean aortic blood pressure and sig nificance ofdifferences in mean values obtained in adjacent inotropic states. (Mean ± SD; Student's t test.)
Figure 2
efficients computed with Fisher's z-transformation? Within each inotropic state partial correlation coef ficients were calculated to determine the independent influences of heart rate, changes in end-diastolic seg ment length, and changes in mean aortic blood pres sure on Doppler index measurements. RESULTS
Satisfactory data for all inotropic states were obtained in six animals. Data were not obtained after admin istration of propranolol in one animal, and in a sec ond animal measurements could not be obtained dur ing high-dosage infusion of dobutamine or after ad ministration of propranolol. The ranges of mean aortic blood pressure over which data were collected were 99 to 200 mm Hg for the control condition, 92 to 183 mm Hg for low-dosage infusion of do butarnine, 88 to 166 mm Hg for high-dosage do-· butarnine, and 81 to 160 mm Hg after the admin istration of propranolol. Doppler envelopes obtained
Journal of the American Society of Echocardiography
138 W allmeyer et al.
Table 1 Correlation between Doppler (horiwntal rows) and invasively derived (vertical columns) indexes of left ventricular function for all eight animals studied Peak velocity
r
Mean acceleration
r
r(HR) r(HR)
Systolic velocity integral
r (HR)
Q,...
dQ/dt
dP/dt
0.93 0.97 0.89 0.89 0.73 0.79
0.91 0.92 0.89 0.88 0.61 0.70
0.89 0.91 0.89 0.89 0.55 0.66
r (HR), Partial correlation coefficient, adjusted for heart rate.
Table 2 Doppler correlations between mean aortic blood pressure and measurements of Doppler indexes of left ventricular performances (vertical column) for each inotropic state (horiwntal row)*
Propranolol Control Low-dosage dobutamine High-dosage dobutamine
Systolic velocity integral
Peak velocity
Mean acceleration
0.07 -0.14 -0.11 -0.19 0.11
-0.03 -0.25 -0.13 0.21 0.02
0.17 -0.15 -0.15
r (SL) r
0.11 0.08
0.07 -0.01
-0.04 0.23
r (SL)
-0.12
0.00
0.09
r (SL) r (SL)
0.15
r (SL), Partial correlation cooefficient, adjusted for segment length. *No coefficients were statistically significant, and they did not change ap preciably when adjusted for simultaneous changes in end-diastolic segment length.
from a single animal at different levels ofmean aortic blood pressure in each of the four inotropic states (horizontal rows) are illustrated in Figure I. Very little change in the height (peak velocity) and contour of the Doppler envelopes was seen with changes in the level of mean aortic blood pressure. By contrast, differences in the envelopes between inotropic states are readily apparent. (The scale is compressed during administration of dobutamine.) The mean ± standard deviation for each Doppler index measurement in the four inotropic states is illustrated in Figure 2. Noted also is the level of significance of differences between inotropic states in mean values obtained for each index. Ofthe Dopp ler indexes, peak velocity and mean acceleration both showed significant differences between adjacent ino tropic states. Measurements of the systolic velocity integral obtained under differing inotropic condi tions showed considerable overlap. Differences in
values of the systolic velocity integral were statisti cally significant only for the propranolol versus con trol values, suggesting that it is a less sensitive in dicator of changes in left ventricular performance than is peak velocity or mean acceleration. The correlation coefficients of Doppler index mea surements and simultaneous measurements of con ventional invasive indexes are presented in Table I. Peak velocity and mean acceleration both correlated closely with each of the invasive indexes under con ditions of widely varying increases in the level of afterload. For example, the relationships of peak ve locity and mean acceleration to maximum dP/dt and maximum aortic blood flow (Cb.ax) for all four ino tropic states in two ofthe eight animals are illustrated in Figure 3. In each correlation regression lines for the other six animals showed similar correlation val ues, with slopes intermediate between the two ani mals shown. The composite correlation coefficients based on results in all eight animals were between 0.89 and 0.93. The systolic velocity integral corre lated less well with the invasively derived indexes studied. Also shown in Table l are correlation co efficients adjusted for the effect of heart rate, dem onstrating that changes in heart rate did not appre ciably affect the correlation of the Doppler indexes with the invasive indexes. To more precisely examine the effect of changing afterload on the Doppler indexes of performance, regression analysis of index determinations in each inotropic state against the level of mean aortic blood pressure was performed (Table 2). The correlation coefficients were not statistically significant (different from zero) for any of the Doppler indexes in any of the inotropic states. Table 2 also gives the correlation coefficients of index measurements with mean aortic blood pressure (statistically) adjusted for simulta neous changes in end-diastolic segment length. Re moving the influence of changes in preload did not reveal any material effect of changing afterload on the Doppler index measurements.
Volume 1 Number 2 March-Aprill988
3.0
u
80 60
2.S
...
c
Gl
._
.s
~
1.S
a;
>
r
a.
40
"'.S c
30
Gl
Gl
...
II
o.s
10 Composite r = .89
0.0
0 u •
0
1000
2000
3000
4000
sooo
6000
0
7000
max dP/dt
>
~ 0 a;
>
~
... a.
®
0
1000
2000
3000
4000
sooo
=
6000
.89 7000
._ ....
2.0
60
iii
so
-... u Gl • u u._
.92
"'.S c
1.S
II
1.0
Gl
:1!
II
70
c
.2
o.s
40
r
10
0
20
30
20
0 40
Qmax
0
.88
=
30
10 Composite r = .93
CD
Composite r
80
2.S
0.0
= .91
max dP/dt
3.0
Gl
!
r
20
:1!
II Gl
:i=b u • u._
= .92
1.0
~
so
.2 iii
2.0
u0
139
Effect of changes in afterload
Composite r 0
200
400
600
800
= .89 1000
max dQ/dt
Figure 3 Linear regression analysis of Doppler measurements vs invasive measurements of ventricular performance. A, Peak velocity vs dP/dt. B, Mean acceleration vs dP/dt. C, Peak velocity vs maximum aortic blood flow. D, Mean acceleration vs maximum aortic blood flow. Regression lines shown represent two animals with slopes at extremes. Composites of corre lation· coefficients obtained in each of eight animals are indicated at lower right in each panel.
DISCUSSION
The results of this study confirm the sensitivity of Doppler echocardiographic measurements ofascend ing aortic blood velocity to changes in left ventricular systolic performance. Even with the wide variations in afterload produced in this study, peak velocity and mean acceleration both remained sensitive to changes in inotropic state and correlated closely with con ventional invasive indexes of performance. As in our prior study in which preload was varied, peak velocity appears marginally superior to mean acceleration as an index of performance, with the systolic velocity integral far less sensitive. 6 Data collected in this study again indicate that although Doppler measurements of peak aortic blood velocity and mean acceleration correlated closely with invasively derived indexes within a given animal, there was substantial vari ability in the slopes of the regression lines in different animals. This precludes use of Doppler measure
ments for comparisons of ventricular performance between animals but demonstrates that the Doppler indexes are as effective as the invasive indexes tested in detecting changes in performance within a given animal. One would expect that if myocardial contractility remained constant, increases in afterload would de press Doppler echocardiographic measurements of peak aortic blood velocity and mean acceleration. Several factors may have blunted the potential influ ence of changes in afterload in our experiments. One such possibility is that the infusion of phenylephrine used to affect increases in afterload also provided some degree of inotropic stimulation. It has been shown, however, that the inotropic effect of phen ylephrine is weak in comparison with its pressor ef fects. 8 More important in the interpretation of our data is the inability in the intact animal to affect alterations in preload and afterload, completely in dependent of each other. Partial manual caval occlu sion blunted but would not prevent some increase
Joumal of the American Society of Echocardiography
140 Wallmeyer et al.
in preload in response to increases in afterload in this study. This opposing influence of simultaneous pre load changes also affects conventional indexes of per formance, including maximum aortic blood flow and dP/dt. Our data do suggest, however, that Doppler indexes are no more influenced by afterload than are the invasive indexes studied and that their correlation with invasive indexes remains close, even under con ditions of substantial increases in afterload. Further studies are required to define the effect of decreases in afterload and to assess the sensitivity of Doppler echocardiographic measurements to more subtle al terations in left ventricular performance. Peak velocity emerged in our experiments as the Doppler index most sensitive to changes in inotropic state and most closely correlated with the traditional invasively derived indexes. This may have resulted from our determination of acceleration as the mean change in velocity with respect to time over the pe riod from onset of ejection to attainment of peak velocity. In our prior study acceleration was subject to an interobserver variability of 16.8%, compared with interobserver variability for peak velocity ofonly 3.7%. 6 This variability is in close agreement with reports of interobserver and intraobserver variability of Doppler measurements in humans. 9 Although the recording of Doppler measurements in the contin uous mode avoided potential errors in the placement of the sample volume in our experiments, the use of pulsed mode may allow more accurate measurement of aortic blood velocities. Previous work 10- 13 has demonstrated that the sys tolic velocity integral correlates closely with stroke volume and that it is useful in the noninvasive esti mation of cardiac output. Our finding that it is less sensitive than peak velocity or mean acceleration as an index ofleft ventricular performance suggests that the systolic velocity integral, like stroke volume, is strongly influenced by heart rate and loading con ditions. Given the minimal beat-to-beat variability observed in our study, it is unlikely that measurement of the systolic velocity integral for a series of beats or by automated methods would substantially affect these results. The significant decline in systolic ve locity integral from control levels after the adminis tration of propranolol in this study may point to its value in conditions under which myocardial function is depressed or the therapeutic end point has in creased cardiac output rather than improved con tractile function. In conclusion, Doppler echocardiographic mea surements of ascending aortic peak blood velocity and mean acceleration correlate closely with invasive
indexes of left ventricular function over a wide range of loading conditions in this short-term canine prep aration. The ability of these indexes of systolic per formance to reflect changes in contractile state, even when loading conditions are changing, lends further support to the study of their clinical applications. We gratefully acknowledge the technical assistance ofJames Roberts and Donna Siegesmund, and the secretarial assis tance of Jessica Provine.
REFERENCES l. Daley PJ, Sagar KB, Collier BD, Kalbfleisch J, Wann LS. Detection of exercise induced changes in left ventricular sys tolic performance using Doppler echocardiography. Br Heart J [In press]. 2. Bryg RJ, Labovitz AJ, Mehdirad AA, Williams GA, Chairman BR. Effect of coronary artery disease on Doppler-derived parameters of aortic flow during upright exercise. Am J Car diol1986;58:l4 -9. 3. Teague SM, Mark DB, Radford M, eta!. Doppler velocity profiles reveal ischemic exercise responses [Abstract]. Circu lation 1984;70: 185II. 4. Sabbah HN, Khaja S, Brymer JS, et a!. Noninvasive evalu ation of left ventricular performance based on aortic blood acceleration measured with a continuous-wave Doppler ve locity meter. Circulation 1986;74:323-9. 5. Mehta N, Bennett D, Dawkins K, Ward D, Mannering D. Doppler monitor hemodynamic response and early ischemia in exercise stress testing predict triple vessel disease [Abstract]. JAm Coil Cardiol1986;7:16A. 6. Wallmeyer K, Wann LS, Sagar KB, Kalbfleisch J, Klopfen stein HS. The influence of preload and heart rate on Doppler echocardiographic indices of left ventricular performance: comparison with invasive indices in an experimental model. Circulation 1986;74:181-6. 7. Snedecor GW, Cochran WGL. Statistical methods. Ames, Iowa: Iowa State University Press, 1969:185. 8. Goldberg Ll, Cotten MV, Darby TD, Howell EV. Com parative heart contractile force effects of equipressor doses of several sympathomimetic amines. J Pharmacal Exp Ther 1953;108:177-85. 9. Gardin JM, Dabestani A, Marin K, Allfie A, Russell D, Henry WL. Reproducibility of Doppler aortic blood flow measure ments: studies on intraobserver, interobserver and day-to-day variability in normal subjects. Am J Cardiol 1984;54: 1092-8. 10. Steingart RM, Meller J, Barovick J, Patterson R, Herman M, Teichholz LE. Pulsed Doppler echocardiographic measure ment of beat-to-beat changes in stroke volume in dogs. Cir culation 1980;62:542-8. 11. Colocousis JS, Huntsman LL, Curreri PW. Estimation of stroke volume changes by ultrasonic Doppler. Circulation 1977;56:914-7. 12. Chandraratna PAN, Nanna M, McKay C, eta!. Determina tion of cardiac output by transcutaneous continuous-wave ultrasonic Doppler computer. Am J Cardiol1984;53:234-7. 13. Magnin PA, Stewart JA, Myers S, von Ramm 0, Kisslo JA. Combined Doppler and phased array echocardiographic es timation of cardiac output. Circulation 1981;63:388-92.
To investigate the influence of changes in afterload on Doppler echocardiographic determination of peak aortic blood velocity, mean acceleration, and systolic velocity integral, eight dogs with their chests opened were studied in four inotropic states at varying levels of heart rate and mean aortic blood pressure. Data were collected in the control state, at two different levels of dobutamine administration (5 and 10 ,_..g/kg/min intravenously), and after administration of propranolol (0.5 mg/kg intravenously). In each inotropic state, phenylephrine was infused intravenously to produce at least two successive steady state increases of 10 mm Hg or more in mean aortic blood pressure. Within a given animal, peak velocity emerged as the Doppler index most closely correlated with changes in Q...., dQ/dt, and dP/dt (r = 0.94, 0.91, and 0.89, respectively). Mean acceleration also correlated closely with the invasive indexes (r 0.87, 0.89, and 0.89). The effect of changes in mean aortic blood pressure on Doppler index measurements was not statistically significant in any of the inotropic states and did not affect the closeness of their correlation with the invasive indexes. We conclude that Doppler echocirdiographic measurements of aortic blood peak velocity and mean acceleration remained as sensitive to changes in the inotropic state under conditions of varying increases in afterload as did the conventional invasive indexes tested. (JAM Soc EcHo 1988;1:135-40.)
=
Many pl,lblished reports lend support to the use of Doppler echocardiographic measurements of aortic blood velocity to derive indexes of left ventricular performance. The low cost, noninvasive nature, and the ability to make measurements during exercise make this technique particularly attractive. Clinical studies have demonstrated the sensitivity of Doppler aortic blood velocimetry to exercise-induced left ven tricular dysfunction in patients with coronary artery disease, l-s and the ease of making serial determina-
From the Cardiology Division, Department ofMedicine, Medical College of Wisconsin; and Zablocki Veterans Administration Medical Center. Supported in part by grants from the National Heart, Lung, and Blood Institute, National Institutes of Health, the Veterans Ad ministration, the American Heart Association, and the Heath Foundation. Reprint requests: L. Samuel Wann, MD, Cardiology Division, Medical College of Wisconsin, 8700 W. Wisconsin Avenue, Mil waukee, WI 53226.
tions suggests possible applications in monitoring the effectiveness of therapeutic interventions aimed at improving left ventricular function. Changes in myocardial contractile function have always been difficult to isolate from changes in load ing conditions.- Therefore, as with conventional in dexes, the clinical application of Doppler echocar diographic indexes of performance must be guided by an appreciation of the degree to which they are influenced by changes in preload and afterload. Re cent studies in our laboratory demonstrated that un der conditions of varying preload, Doppler deter mination of peak aortic blood velocity and mean ac celeration was as sensitive to changes in inotropic states as the conventional invasive measurement of peak aortic blood flow, maximum rate of change in aortic blood flow (dQ/dt), and peak left ventricular dP/dt. 6 This study was done to evaluate the influence of changes in afterload on Doppler echocardiogra phic indexes of ventricular performance. 135
Journal of the American Society of Echocardiography
136 Wallmeyer et al.
. ... '· '·
Doppler Afterload r1.0 msec
CONTROL
At..·ji
......
135
122
12.0 msec
LOW DOSE DOBUTAMINE
HIGH DOSE DOBUTAMINE
96
I 1.0 msec PROPIRANOLOL
....... !·'-. "-·"'.
....
86
o~M
~
130
117
I•• · 112 ·-
,. l.j,
I
. . . . .,
124
.i.:
M..
156
139
I 2.0 msec
-' ~·~ ~~ 184
Ja.. A.. •.
,,L....
116
......l..~~
163
172
1.1~ j.,,,"' 148
j
J ol
1111 ol
A• I"'" o
150
Figure 1 Doppler recordings obtained from single animal. Horizontal rows represent four inotropic states. Note change in velocity scale during inotropic stimulation. Numbers below recordings are simultaneous measurements of mean aortic blood pressure (millimeters of mercury).
METHODS
Eight mongrel dogs weighing 24.6 to 33.1 kg were given no food overnight, anesthetized by an intravenous administration of alpha chloralose and urethane, intubated, and ventilated with a volume respirator (Harvard Apparatus Co., Inc., S. Natick, Massachusetts) with air enriched with oxy gen (6 L/min). Polyvinyl catheters (Tygon Micro bore Tubing, 0.05 inch, Norton Performance Plas tics, Wayne, New Jersey) were introduced into both femoral veins for infusion of drugs, and a third cath eter was advanced via a femoral artery to the proximal aorta. A micromanometer-tipped catheter (Millar In struments, Inc., Houston) was inserted into a femoral artery and advanced to the proximal aorta, allowing measurement of mean aortic blood pressure as an indicator of changing afterload. The electrocardio gram was monitored throughout, and arterial blood gas levels were measured at 1-hour intervals and were maintained in the normal range. Thoracotomy was performed in the left fifth in tercostal space, and the position of the aortic cath eters was verified by palpation. A fluid-filled catheter was placed into the left atrium and secured by purse string suture. Aortic and left atrial catheters were connected to fluid-filled Statham P23Db pressure transducers (Gould Inc. Cardiovascular Products, Oxnard, California), with the zero pressure reference point at the middle right atrial level. An electromag netic flow probe (Howell Instrument Co., Camarillo,
California) was placed around the ascending aorta, and blood flow was measured with a N arcomatic flowmeter (model RT-500, Narco Bio-Systems, Houston). A precalibrated high fidelity pressure transducer (Konigsberg Instruments, Inc., Pasadena, California) was inserted into the left ventricle through an apical stab wound, secured with a purse string suture, and referenced to the high fidelity prox imal aortic pressure waveform during systole and the left atrial catheter during diastole. Two sonomicro meter crystals (PZT5A [flat and cylinder], Valpey Fisher Med. Div., Hopkinton, Massachusetts) were inserted into the anterior left ventricular free wall at the midmyocardial level, perpendicular to the long axis of the ventricle, and the distance between them was measured (Hartley Multi-channel Flow Dimension System, Instrumentation Development Lab, Baylor College of Medicine, Houston) as an index of preload. Finally, pacing wires were sutured to the left atrium, and a snare placed around the inferior vena cava was used to attenuate any increases in end-diastolic segment length that occurred because ofincreases produced in mean arterial blood pressure. All hemodynamic measurements were made under steady state conditions with mechanical ventilation briefly suspended to eliminate respiratory variation. Hemodynamic data were recorded on an analog FM tape recorder (A.R. Vetter Co., Rebersburg, Penn sylvania) and on an eight-channel strip chart recorder (model 2800, Gould, Inc., Cleveland). Ten-second data files were subsequently transferred to a digital
Volume l Number 2 March-Aprill988
computer (DEC LSI ll/23) and sequential inter active programs used to calculate peak volumetric blood flow (Qmax), maximum dQ/dt, and maximum left ventricular dP/dt. Simultaneous with the collec tion of hemodynamic data, Doppler measurements of blood velocity were obtained in the continuous mode with a dedicated nonimaging Doppler trans ducer (Pedof, New Brunswick, New Jersey) with a 2 MHz carrier frequency (Irex III B, Johnson & Johnson, New Brunswick, New Jersey) and recorded on a strip-chart recorder with a paper speed of 50 mm/sec. The transducer was handheld and applied to the aortic arch with the investigator's hand stea died against the thoracic cage. The ultrasonic beam was oriented parallel to ascending aortic blood flow, guided by audio and graphic signals to obtain max imal Doppler shift, with velocity envelopes having distinct borders. Doppler measurements of peak blood velocity and mean acceleration were made for six consecutive beats and averaged. The systolic ve locity integral was determined by manual planimetry ofa representative Doppler envelope, applying Simp son's rule. Measurements were obtained initially in the base line contractile state, followed sequentially by measurements in three additional inotropic states produced by low-dosage infusion of dobutamine (5 J.Lglkg/min intravenously), high-dosage infu sion of dobutarnine (10 J.Lg/kg/min intravenously), and finally after administration of propranolol (0.5 mg/kg intravenously, given more than 10 minutes after discontinuation of dobutarnine). In each ino tropic state data were collected at a baseline mean aortic blood pressure, followed by measurements during infusion of phenylephrine at rates sufficient to produce at least two successive steady state in creases of at least 10 mm Hg in mean aortic blood pressure. Whenever the heart rate in sinus rhythm was less than 150 beats/min, data were also collected during rapid atrial pacing (150 beats/min) at the same level of infusion of phenylephrine. Premature beats and the cycles after them were excluded from analysis. Data samples with gross irregularity of rhythm or with pulsus alternans were also excluded. For each animal studied the values for each index of contractility obtained in each inotropic state were averaged. These within-animal means were then an alyzed with a two-way analysis of variance followed by the paired t test to compare means of determi nations in adjacent inotropic states. The correlation of Doppler indexes with the simultaneous invasive index determinations was analyzed by linear regres sion, with the pooled within-animal correlation co-
Effect of changes in afterload 137
3.0
~
C3
o~
-' w
0
(J)
p
< .05
p < .01
p
< .02
< .01
p
< .01
2.0
>~
~.s 1.0
< w a..
z
0
100
~
80 60
a:~
w ~ -' (/)
wo() ~
p < .05
p
40
<"E
(/)
20
z~
< w
~
>
1o~
g.§.. W-'
>< () a: _
200
p
< .05
NS
NS
150 100
-'W
0 I1-Z
oo > CJ)
50 PROPRANOLOL CONTROL VS. DOBUTAMINE (0.5 mg/kg) VS. CONTROL (5 meg/kg/min)
DOBUTAMINE 5 vs. 10 (meg/kg/min)
Doppler index measurements for each inotropic state at all levels of mean aortic blood pressure and sig nificance ofdifferences in mean values obtained in adjacent inotropic states. (Mean ± SD; Student's t test.)
Figure 2
efficients computed with Fisher's z-transformation? Within each inotropic state partial correlation coef ficients were calculated to determine the independent influences of heart rate, changes in end-diastolic seg ment length, and changes in mean aortic blood pres sure on Doppler index measurements. RESULTS
Satisfactory data for all inotropic states were obtained in six animals. Data were not obtained after admin istration of propranolol in one animal, and in a sec ond animal measurements could not be obtained dur ing high-dosage infusion of dobutamine or after ad ministration of propranolol. The ranges of mean aortic blood pressure over which data were collected were 99 to 200 mm Hg for the control condition, 92 to 183 mm Hg for low-dosage infusion of do butarnine, 88 to 166 mm Hg for high-dosage do-· butarnine, and 81 to 160 mm Hg after the admin istration of propranolol. Doppler envelopes obtained
Journal of the American Society of Echocardiography
138 W allmeyer et al.
Table 1 Correlation between Doppler (horiwntal rows) and invasively derived (vertical columns) indexes of left ventricular function for all eight animals studied Peak velocity
r
Mean acceleration
r
r(HR) r(HR)
Systolic velocity integral
r (HR)
Q,...
dQ/dt
dP/dt
0.93 0.97 0.89 0.89 0.73 0.79
0.91 0.92 0.89 0.88 0.61 0.70
0.89 0.91 0.89 0.89 0.55 0.66
r (HR), Partial correlation coefficient, adjusted for heart rate.
Table 2 Doppler correlations between mean aortic blood pressure and measurements of Doppler indexes of left ventricular performances (vertical column) for each inotropic state (horiwntal row)*
Propranolol Control Low-dosage dobutamine High-dosage dobutamine
Systolic velocity integral
Peak velocity
Mean acceleration
0.07 -0.14 -0.11 -0.19 0.11
-0.03 -0.25 -0.13 0.21 0.02
0.17 -0.15 -0.15
r (SL) r
0.11 0.08
0.07 -0.01
-0.04 0.23
r (SL)
-0.12
0.00
0.09
r (SL) r (SL)
0.15
r (SL), Partial correlation cooefficient, adjusted for segment length. *No coefficients were statistically significant, and they did not change ap preciably when adjusted for simultaneous changes in end-diastolic segment length.
from a single animal at different levels ofmean aortic blood pressure in each of the four inotropic states (horizontal rows) are illustrated in Figure I. Very little change in the height (peak velocity) and contour of the Doppler envelopes was seen with changes in the level of mean aortic blood pressure. By contrast, differences in the envelopes between inotropic states are readily apparent. (The scale is compressed during administration of dobutamine.) The mean ± standard deviation for each Doppler index measurement in the four inotropic states is illustrated in Figure 2. Noted also is the level of significance of differences between inotropic states in mean values obtained for each index. Ofthe Dopp ler indexes, peak velocity and mean acceleration both showed significant differences between adjacent ino tropic states. Measurements of the systolic velocity integral obtained under differing inotropic condi tions showed considerable overlap. Differences in
values of the systolic velocity integral were statisti cally significant only for the propranolol versus con trol values, suggesting that it is a less sensitive in dicator of changes in left ventricular performance than is peak velocity or mean acceleration. The correlation coefficients of Doppler index mea surements and simultaneous measurements of con ventional invasive indexes are presented in Table I. Peak velocity and mean acceleration both correlated closely with each of the invasive indexes under con ditions of widely varying increases in the level of afterload. For example, the relationships of peak ve locity and mean acceleration to maximum dP/dt and maximum aortic blood flow (Cb.ax) for all four ino tropic states in two ofthe eight animals are illustrated in Figure 3. In each correlation regression lines for the other six animals showed similar correlation val ues, with slopes intermediate between the two ani mals shown. The composite correlation coefficients based on results in all eight animals were between 0.89 and 0.93. The systolic velocity integral corre lated less well with the invasively derived indexes studied. Also shown in Table l are correlation co efficients adjusted for the effect of heart rate, dem onstrating that changes in heart rate did not appre ciably affect the correlation of the Doppler indexes with the invasive indexes. To more precisely examine the effect of changing afterload on the Doppler indexes of performance, regression analysis of index determinations in each inotropic state against the level of mean aortic blood pressure was performed (Table 2). The correlation coefficients were not statistically significant (different from zero) for any of the Doppler indexes in any of the inotropic states. Table 2 also gives the correlation coefficients of index measurements with mean aortic blood pressure (statistically) adjusted for simulta neous changes in end-diastolic segment length. Re moving the influence of changes in preload did not reveal any material effect of changing afterload on the Doppler index measurements.
Volume 1 Number 2 March-Aprill988
3.0
u
80 60
2.S
...
c
Gl
._
.s
~
1.S
a;
>
r
a.
40
"'.S c
30
Gl
Gl
...
II
o.s
10 Composite r = .89
0.0
0 u •
0
1000
2000
3000
4000
sooo
6000
0
7000
max dP/dt
>
~ 0 a;
>
~
... a.
®
0
1000
2000
3000
4000
sooo
=
6000
.89 7000
._ ....
2.0
60
iii
so
-... u Gl • u u._
.92
"'.S c
1.S
II
1.0
Gl
:1!
II
70
c
.2
o.s
40
r
10
0
20
30
20
0 40
Qmax
0
.88
=
30
10 Composite r = .93
CD
Composite r
80
2.S
0.0
= .91
max dP/dt
3.0
Gl
!
r
20
:1!
II Gl
:i=b u • u._
= .92
1.0
~
so
.2 iii
2.0
u0
139
Effect of changes in afterload
Composite r 0
200
400
600
800
= .89 1000
max dQ/dt
Figure 3 Linear regression analysis of Doppler measurements vs invasive measurements of ventricular performance. A, Peak velocity vs dP/dt. B, Mean acceleration vs dP/dt. C, Peak velocity vs maximum aortic blood flow. D, Mean acceleration vs maximum aortic blood flow. Regression lines shown represent two animals with slopes at extremes. Composites of corre lation· coefficients obtained in each of eight animals are indicated at lower right in each panel.
DISCUSSION
The results of this study confirm the sensitivity of Doppler echocardiographic measurements ofascend ing aortic blood velocity to changes in left ventricular systolic performance. Even with the wide variations in afterload produced in this study, peak velocity and mean acceleration both remained sensitive to changes in inotropic state and correlated closely with con ventional invasive indexes of performance. As in our prior study in which preload was varied, peak velocity appears marginally superior to mean acceleration as an index of performance, with the systolic velocity integral far less sensitive. 6 Data collected in this study again indicate that although Doppler measurements of peak aortic blood velocity and mean acceleration correlated closely with invasively derived indexes within a given animal, there was substantial vari ability in the slopes of the regression lines in different animals. This precludes use of Doppler measure
ments for comparisons of ventricular performance between animals but demonstrates that the Doppler indexes are as effective as the invasive indexes tested in detecting changes in performance within a given animal. One would expect that if myocardial contractility remained constant, increases in afterload would de press Doppler echocardiographic measurements of peak aortic blood velocity and mean acceleration. Several factors may have blunted the potential influ ence of changes in afterload in our experiments. One such possibility is that the infusion of phenylephrine used to affect increases in afterload also provided some degree of inotropic stimulation. It has been shown, however, that the inotropic effect of phen ylephrine is weak in comparison with its pressor ef fects. 8 More important in the interpretation of our data is the inability in the intact animal to affect alterations in preload and afterload, completely in dependent of each other. Partial manual caval occlu sion blunted but would not prevent some increase
Joumal of the American Society of Echocardiography
140 Wallmeyer et al.
in preload in response to increases in afterload in this study. This opposing influence of simultaneous pre load changes also affects conventional indexes of per formance, including maximum aortic blood flow and dP/dt. Our data do suggest, however, that Doppler indexes are no more influenced by afterload than are the invasive indexes studied and that their correlation with invasive indexes remains close, even under con ditions of substantial increases in afterload. Further studies are required to define the effect of decreases in afterload and to assess the sensitivity of Doppler echocardiographic measurements to more subtle al terations in left ventricular performance. Peak velocity emerged in our experiments as the Doppler index most sensitive to changes in inotropic state and most closely correlated with the traditional invasively derived indexes. This may have resulted from our determination of acceleration as the mean change in velocity with respect to time over the pe riod from onset of ejection to attainment of peak velocity. In our prior study acceleration was subject to an interobserver variability of 16.8%, compared with interobserver variability for peak velocity ofonly 3.7%. 6 This variability is in close agreement with reports of interobserver and intraobserver variability of Doppler measurements in humans. 9 Although the recording of Doppler measurements in the contin uous mode avoided potential errors in the placement of the sample volume in our experiments, the use of pulsed mode may allow more accurate measurement of aortic blood velocities. Previous work 10- 13 has demonstrated that the sys tolic velocity integral correlates closely with stroke volume and that it is useful in the noninvasive esti mation of cardiac output. Our finding that it is less sensitive than peak velocity or mean acceleration as an index ofleft ventricular performance suggests that the systolic velocity integral, like stroke volume, is strongly influenced by heart rate and loading con ditions. Given the minimal beat-to-beat variability observed in our study, it is unlikely that measurement of the systolic velocity integral for a series of beats or by automated methods would substantially affect these results. The significant decline in systolic ve locity integral from control levels after the adminis tration of propranolol in this study may point to its value in conditions under which myocardial function is depressed or the therapeutic end point has in creased cardiac output rather than improved con tractile function. In conclusion, Doppler echocardiographic mea surements of ascending aortic peak blood velocity and mean acceleration correlate closely with invasive
indexes of left ventricular function over a wide range of loading conditions in this short-term canine prep aration. The ability of these indexes of systolic per formance to reflect changes in contractile state, even when loading conditions are changing, lends further support to the study of their clinical applications. We gratefully acknowledge the technical assistance ofJames Roberts and Donna Siegesmund, and the secretarial assis tance of Jessica Provine.
REFERENCES l. Daley PJ, Sagar KB, Collier BD, Kalbfleisch J, Wann LS. Detection of exercise induced changes in left ventricular sys tolic performance using Doppler echocardiography. Br Heart J [In press]. 2. Bryg RJ, Labovitz AJ, Mehdirad AA, Williams GA, Chairman BR. Effect of coronary artery disease on Doppler-derived parameters of aortic flow during upright exercise. Am J Car diol1986;58:l4 -9. 3. Teague SM, Mark DB, Radford M, eta!. Doppler velocity profiles reveal ischemic exercise responses [Abstract]. Circu lation 1984;70: 185II. 4. Sabbah HN, Khaja S, Brymer JS, et a!. Noninvasive evalu ation of left ventricular performance based on aortic blood acceleration measured with a continuous-wave Doppler ve locity meter. Circulation 1986;74:323-9. 5. Mehta N, Bennett D, Dawkins K, Ward D, Mannering D. Doppler monitor hemodynamic response and early ischemia in exercise stress testing predict triple vessel disease [Abstract]. JAm Coil Cardiol1986;7:16A. 6. Wallmeyer K, Wann LS, Sagar KB, Kalbfleisch J, Klopfen stein HS. The influence of preload and heart rate on Doppler echocardiographic indices of left ventricular performance: comparison with invasive indices in an experimental model. Circulation 1986;74:181-6. 7. Snedecor GW, Cochran WGL. Statistical methods. Ames, Iowa: Iowa State University Press, 1969:185. 8. Goldberg Ll, Cotten MV, Darby TD, Howell EV. Com parative heart contractile force effects of equipressor doses of several sympathomimetic amines. J Pharmacal Exp Ther 1953;108:177-85. 9. Gardin JM, Dabestani A, Marin K, Allfie A, Russell D, Henry WL. Reproducibility of Doppler aortic blood flow measure ments: studies on intraobserver, interobserver and day-to-day variability in normal subjects. Am J Cardiol 1984;54: 1092-8. 10. Steingart RM, Meller J, Barovick J, Patterson R, Herman M, Teichholz LE. Pulsed Doppler echocardiographic measure ment of beat-to-beat changes in stroke volume in dogs. Cir culation 1980;62:542-8. 11. Colocousis JS, Huntsman LL, Curreri PW. Estimation of stroke volume changes by ultrasonic Doppler. Circulation 1977;56:914-7. 12. Chandraratna PAN, Nanna M, McKay C, eta!. Determina tion of cardiac output by transcutaneous continuous-wave ultrasonic Doppler computer. Am J Cardiol1984;53:234-7. 13. Magnin PA, Stewart JA, Myers S, von Ramm 0, Kisslo JA. Combined Doppler and phased array echocardiographic es timation of cardiac output. Circulation 1981;63:388-92.