- Email: [email protected]
Constant postoperative monitoring of cardiac output after correction of congenital heart defects
J
THoRAc CARDIOVASC SURG
1987;93:658-64
Constant postoperative monitoring of cardiac output after correction of congenital heart defects A new method has been developed that permits constant postoperative monitoring of mean and phasic cardiac output in patients after correction of congenital heart defects. A miniature ultrasound probe is attached to the adventitia of the ascending aorta at the conclusion of the operative procedure. This is connected to the monitoring equipment by means of polyurethane-covered wires that exit the chest wall through a small stab wound. The probe can easily be removed by gentle traction when the patient's condition is stable. The technique was developed, validated, and refined in extensive animal studies, and this report describes the first series of 20 consecutive human implants, performed between August 1984 and September 1985, in which the absolute cardiac output determination obtained with the ultrasound probe at the time of its application was correlated with cardiac output as measured with a standard electromagnetic flow probe. Fourteen male and six female patients (mean age 5.5 years) were studied. Operations performed included eight atrial septal defect repairs, four procedures for tetralogy of Fallot, three ventricular septal defect repairs, three stenotic valve corrections, and two Senning operations. One operative death occurred, but no complications were related to probe application or removal. The average cardiac output in the 20 patients as measured with the ultrasound probe was 2.2 ± 1.1 L/min (range 0.67 to 5.27 L/min). This is nearly identical to the results noted with the electromagnetic flow probe, where the mean cardiac output was 2.3 ± 1.2 L/min (range 0.7 to 6 L/min). Regression analysis revealed a high linear correlation (r = 0.9) between the two techniques. A monitor can display the cardiac output trend with 1 minute updates, which greatly enhance management of intravenous drug therapy and volume administration. In conclusion, this new extraluminal removable probe allows virtually continuous monitoring of the postoperative cardiac output after correction of congenital heart defects and should become a standard technique in the postoperative care of these patients.
Blair A. Keagy, M.D. (by invitation), Benson R. Wilcox, M.D., Carol L. Lucas, Ph.D. (by invitation), Henry S. Hsiao, Ph.D. (by invitation), G. William Henry, M.D. (by invitation), Michael Baudino, B.S. (by invitation), and Gene Bornzin, B.S. (by invitation), Chapel Hill. N. C; and Minneapolis, Minn.
A
knowledge of cardiac output is important in making appropriate patient care decisions after cardiac operations. A number of invasive and noninvasive techniques have been developed to measure this parameter.':" The standard against which most of these methods have been compared is the electromagnetic flow probe. Although the electromagnetic flow probe is acceptable for use in the experimental laboratory and in From the Department of Surgery, University of North Carolina, Chapel Hill, N. C., and Medtronic, Inc., Minneapolis, Minn. Read at the Sixty-sixth Annual Meeting of The American Association for Thoracic Surgery, New York, N. Y., April 28-30, 1986. Address for reprints: Blair A. Keagy, M.D., Divisionof Cardiothoracic Surgery, 108 Burnett-Womack Bldg., 229H, Chapel Hill, N. C. 27514.
658
the operating room, the size of the probe, the necessity for encircling the aorta with a rigid cuff, and the inability to remove the device outside of the operating room preclude its use in postoperative monitoring. Thermodilution is another means of monitoring cardiac output, but this technique fails to account for cardiac output fluctuations based on the respiratory cycle and provides only periodic information. In addition, some investigators have questioned the accuracy of this method. 1I • 13 More recently, ultrasound has been used to study cardiovascular hemodynamics. When an ultrasound beam is directed into the lumen of a blood vessel, its backscatter-or Doppler frequency shift by a moving column of red blood cells-is proportional to the velocity of blood flow. This relationship may be quantified by the
Volume 93
Postoperative monitoring of cardiac output
Number 5 May 1987
Dopplerequation, which states that the velocity of blood flow is directly proportional to the frequency of this reflected signal and inversely proportional to the cosine of the angle at which the ultrasound beam insonates the longitudinal axis of blood flOW. I4- 17 v=
cM 2focosO
where v = velocity of blood flow, c = average speed of sound in tissue (1.55 x 105 em/sec), M = frequency of the Doppler shifted signal, fo = frequency of the Doppler used, and () = angle of intersection between the Doppler beam and the longitudinal axis of the vessel. Two issues had to be addressed before an ultrasound device could be assumed to provide accurate cardiac output measurements. First, if velocity was sampled in only a small area of the ascending aorta and this value was used to extrapolate the velocity across the entire vessel, the cross-sectional velocity profile in the aorta had to be established. Second, since cardiac output is a product of velocity and cross-sectional area, it was necessary to establish a method of measuring accurately the area of the ascending aorta at the site where the velocity is being measured. In peripheral vessels, the cross-sectional velocity profile has been shown to be parabolic, with higher velocities occurring in the center of the vessel; however, the cross-sectionalvelocity profile in the ascending aorta has been a point of some controversy. To establish the shape of this profile,a tiny ultrasound crystal was placed on the tip of a small needle that was introduced into the lumen of the ascending aorta of a group of experimental animals.A siliconerubber sheath anchored the device to the wall of the aorta. The tip of the probe was moved from wall to wall inside the aortic lumen, and measurements were made away from the tip of the needle to avoidprobe-induced turbulence." This study established that the cross-sectional velocity profile in the ascending aorta is flat, and it validated the concept of using a small-velocity sampling area in the ascending aorta as an indicator of mean velocity across the entire vessel. Because volume flow (cardiac output) is a product of mean velocity in a vessel times its cross-sectional area, it is necessary to know the diameter of the ascending aorta in addition to the velocity of blood flow. With regard to changes in the aortic diameter, Greenfield and Patel" demonstrated only a 1.6% ± 0.5% change in aortic diameter for each 20 rom Hg change in aortic pressure in adult humans. This minimal change in cardiac output was confirmed in our animal laboratory by placing ultrasound crystals on opposite sides of the aorta and varying the mean systemic blood pressure by volume
659
administration and pharmacologic manipulation." The study showed that for normal systolic-diastolic differences of 35 rom Hg, the average diameter of the ascending aorta changed 1.04 rom, which resulted in a volume change of 17%. However, if the pulsar component in diameter change is ignored, and only the average blood velocity and average aortic area are used, a relatively small mean error of 4.83% occurs at normal aortic pressures. Thus a single static aortic diameter measurement made at the time of probe application may A) with introducbe used in the flow equation (Q = tion of minimal error. Requirements for a successful postoperative cardiac output monitoring device include ease of application, safety, accuracy, and the ability to remove the device without a second surgical operation. In an effort to achieve these goals, a high-frequency removable ultrasound probe was devised that contained a miniature 20 MHz piezoelectric crystal mounted at 45 degrees in an epoxy cuff. Two small metal tines were incorporated into the tip of the probe to anchor it to the adventitia of the ascending aorta. The probe samples velocity in a 1 mm' area of the ascending aorta. The probes were placed in experimental animals and the results of flow measurements were compared with those obtained simultaneously with an electromagnetic flow probe. The ultrasound probe correlated extremely well with values obtained with the electromagnetic probe (average r = 0.98 ± 0.006).21 At this point, probe construction was undertaken by Medtronic, Inc., Minneapolis, Minnesota, and a series of implantations in humans was begun. The present report describes this initial human study in which an ultrasound probe was used to measure cardiac output in the postoperative period after cardiac operations.
v.
Patients and methods
This study was performed between August 1984 and September 1985, during which time probes were implanted in 20 patients: 14 male and six female patients. The ages ranged from 3 months to 47 years, with a mean age of 5.5 years; 17 patients were under 6 years of age. Operations performed on these individuals included atrial septal defect repairs (eight), procedures for tetralogy of Fallot (four), ventricular septal defect repairs (three), stenotic valve corrections (three), and Senning procedures (two). Approval for human implantation of these probes was obtained from the internal review board of the North Carolina Memorial Hospital, and each patient or his/her family signed an informed consent form before the operation. At the conclusion of the operative procedure, before
The Journal of Thoracic and Cardiovascular
6 6 0 Keagy et al.
Surgery
Connector Cable /
~ P",b. H.ad
~PCA2r:'M~M:::~~"~ne, " .... . '. I •• ••
~
/1 Aortic Wall
Fig. 2. Schematic representation of the probe anchored to the aortic wall with the metal tines.
Fig. 1. Two-plane photographs of the actual ultrasound probe.
closure of the sternum, the cardiac output was measured with a traditional electromagnetic flow probe. The twin metal tines were then used to anchor the ultrasound probe to the adventitia of the ascending aorta. Fig. 1, A and B shows photographs of the probe, and Fig. 2 represents the anchoring tines embedded in the aortic wall. The diameter of the vessel was ascertained by direct measurement, and cardiac output was recorded with the ultrasound probe. The comparison of these two measurements is presented in the Results section. The probe exited through a small stab wound in the anterior chest wall and was anchored to the skin with a small silk suture. A fiberoptic cable was used to attach the probe to the pulsed Doppler unit and to the signal analyzer. Cardiac output was monitored with the ultrasound probe during the patient's immediate postoperative period, and when the patient's condition stabilized, the small silk suture was cut and the probe was removed by gentle traction. The tines slide easily out of the aortic wall. The quadrature frequency data used in this study were obtained with a 20 MHz pulsed Doppler system designed and built by Dr. Craig Hartley, Baylor College of Medicine. Systems specifications included a pulse repetition frequency of 62.5 kHz, a transmitter pulse width of 0.4 usee, a receiver pulse width of 0.3 usee, and a variable range gate of 1 to 12 mm. Output from the Doppler system was amplified at the patient's bedside before being transmitted via fiberoptic cables to a remote PDP 11/23 minicomputer system, where 64
complex samples were obtained every 1/66 second at a rate of 50 kHz using a Data Translation Model 2782 analog-to-digital converter. After 192 sampling bursts were obtained (2.9 seconds), each burst was subjected to Fourier analysis and the power spectrum computed. The 50 kHz range could be adjusted to fit the frequency content of the signal; however, the range of -15 kHz to +35 kHz was adequate for most applications. The Doppler shift (kHz) of the signal during each sampling burst was computed for the shifted power spectrum via a first moment average technique centered around a robust mean." The estimated flow rate during the burst was calculated by multiplying the resultant kilohertz shift measurement by a constant that adjusted for both the Doppler shift equation and the cross-sectional area of the vessel. The cardiac output and heart rate for the 2.9 second interval were computed and displayed on the computer monitor along with their respective averages for the preceding 5 minutes. The pulsatile blood flow waveform for the data just processed arid the trend in cardiac output for the preceding 2 hours were shown graphically. The screen was updated every minute. All results were stored on the computer, and a printed copy of the screen could be obtained as desired.
Result One postoperative death occurred and five complications including postoperative atelectasis (one), pleural effusion (two), postoperative myocardial infarct (one), and respiratory insufficiency (one). None of these problems was related to use of the probe. Probe application was easily accomplished and signals remained constant throughout the postoperative period in 15 patients. Two probes had weak but usable signals, and one patient required replacement of the probe two times in the operating suite because of a weak signal. In
Volume 93 Number 5
Postoperative monitoring of cardiac output 6 6 1
May 1987
7
en
8
I-
z
w i= 6 Cl:
3=
0
...J
D
u,
Do
a-
u.. 0
a: 4 w
z
45 Degrees (USP/EMF Y intercept
-
Line of Least Squares
:::l
l:D
:e :::l
-
(J)
= 1) =0
2
o
4
5
6
7
EMF FLOW 10·11
12·13
14·15
16·17
18·19
20+
AORTIC DIAMETER (MILLIMETERS)
(Regression Equation)
y = 0.423 + 0.809x
Fig. 4. Relationship between cardiac outputs as obtained by the ultrasound technique and by the electromagnetic flow (EMF) probe. USP, Ultrasound probe.
Fig. 3. Diameter measurements made at the time of probe application in all patients in this study.
one individual the signal was lost after 12 hours, and one neonate had a small aorta that made probe application difficult. Diameter measurements were made at operation (Fig. 3), and the aortic diameter ranged from 10 to 22 mm, with a mean of 14.9 ± 3.3 mm. The largest aorta, 22 mm, was present in a 47-year-old patient who underwent correction of an atrial septal defect. An electromagnetic flow probe measurement on the ascending aorta was performed immediately before application of the ultrasound probe in all patients. The average cardiac output in the 20 patients, as measured with the ultrasound probe, was 2.2 ± 1.1 Lzmin (range 0.67 to 5.27 L/min). This is nearly identical to the results noted with the electromagnetic flow probe, the mean cardiac output being 2.3 ± 1.2 L/min (range 0.7 to 6 L/min). Regression analysis revealed a high linear correlation (r = 0.9) between the two techniques. Fig. 4 is a scattergram of all comparative measurements in the study. The thin line represents the linear regression line, and the thick line indicates the ideal situation with the ultrasound probe/electromagnetic flow probe ratio equal to 1. The probes were left in place from 1 to 67 hours after operation (mean 29 ± 13.8 hours) (Fig. 5).
Discussion Ultrasound technology has been previously used in the noninvasiveassessment of cardiac output. Velocity in the ascending aorta was monitored by a scanning probe
applied to the suprasternal notch, and this value was combined with a noninvasive echocardiographic measurement of aortic diameter to compute cardiac outpUt.9, 23-27 Several problems are related to this suprastern~l notch approach. The first concerns the probable aortic arch location of the sample volume used for the frequency (velocity) measurement. In this area, the turbulence involved with origins of the great vessels, the lack of a flat velocity profile as the distance from the aortic valve increases, and the inability to measure accurately the exact angle between the insonant ultrasound beam and the longitudinal axis of blood flow place some doubt on the validity of the velocity measurement. The ultrasound beam is assumed to be parallel with the vessel axis; however, because the geometry of the aortic arch area varies, this assumption of a zero degree angle is fraught with potential error. A second problem is the accuracy of the static diameter measurement, which also may be affected by the angle of the ultrasound scan head and by an unknown magnification factor. It is probable that there is a high linear correlation between values obtained with a suprasternal notch probe and the true cardiac output if the probe is kept in the same position. However, reapplication of a Doppler scan head will usually result in placement of the sample volume in a different location, which subjects it to the errors described earlier. Also, longer distance between ultrasound probe and site of measurement means that a lower frequency must be used, which reduces the signal-to-noise ratio. Much better signals are possible with the removable aortic probe. A recently developed intraesophageal echo Doppler
The Journal of Thoracic and Cardiovascular Surgery
6 6 2 Keagy et al.
en ....z
w
j:::
«
a. LI.
12 10
8
0
6
l:ll
4
a: w
:E
:;)
z
2
< 1 DAY
1-2 DAYS
2-3 DAYS
Fig. 5. Duration of probe application in all patients in the
study. probe has also been used to measure cardiac output. Velocity is measured with a continuous-wave Doppler probe placed in the esophageal lumen, and an A-scan unit allows measurement of the diameter of the thoracic aorta." One of the problems associated with this method is the assumption that descending thoracic aortic flow is directly proportional to the cardiac output, an assumption that ignores the effects of the great vessels in the aortic arch. In addition, the assumed angle for the ultrasound crystal may be in error. After our 1983 report of an extractable ultrasound probe in animals, a version of the probe for use in humans was reported by Matre, Segadal, and Engedal. 29 This device was placed under the adventitia of the ascending aorta after cardiac procedures, and changes in aortic velocity were recorded up to 24 hours postoperatively in 14 patients. Good correlation with thermodilution techniques (r = 0.79) was described. The authors concluded that this device allowed continuous monitoring of cardiac output, as well as short- and long-term trend analysis during the early postoperative period. Implantable ultrasound velocity probes have a number of advantages, including the lack of a baseline drift, a relative independence from hematocrit, and ease of removal when no longer needed. In addition, a constant angle of insonance (45 degrees) is maintained between the ultrasound beam and the longitudinal axis of blood flow, and the same sample volume site is used in all velocity calculations. This corrects one of the problems associated with the suprasternal notch approach, which requires repeated probe applications. In the system described in this report, a pulsed Doppler probe has been used to interrogate the ascending aorta. A pulsed Doppler probe" has one piezoelectric crystal that alternately acts as an ultrasonic transmitter
and receiver. Varying the time interval between these two functions (range gating) allows accurate placement of the Doppler sample volume at various locations in the lumen of the vessel. Also, varying the length of the ultrasound burst allows the size of the sample volume to be controlled within certain limits. This differs from a more commonly used continuous-wave Doppler probe, which contains two piezoelectric crystals, one constantly functioning as the transmitter of an ultrasound beam and the other acting as a continuous receiver. The continuous-wave Doppler probe has the advantage of a much larger sample volume, but the theoretical disadvantage of sampling other surrounding vascular structures such as accompanying veins. In addition, the need for two crystals requires a larger probe tip. The flat velocity profile in the ascending aorta allows the use of the pulsed Doppler techniques. Three additional probe implantations have been performed since this report was submitted. However, the electromagnetic flow probe correlation was not obtained because the validity of the device has been established. One of these patients had a cardiac arrest in close temporal proximity to the probe removal and underwent open chest resuscitation. The patient recovered without incident and was discharged. There was no blood around the site of attachment of the probe on the ascending aorta, and it was postulated that perhaps this episode was related to a severe hypercyanotic ("Tet") spell in this infant who had an uncorrected tetralogy of Fallot. In the remaining two patients the probe was removed without incident. The reflected aortic ultrasound signal can be evaluated with either fast Fourier transform sound-spectral analysis or zero crosser hardware. The latter method involves less expensive equipment and the results can be displayed faster, because extensive computer time is not required for signal analysis. However, the zero crosser becomes inaccurate in the presence of turbulence or with high signal frequencies. There has been controversy in the past about the velocity profile in the ascending aorta, but we think that our method of measuring the velocity profile is better than those of other investigators who have placed a probe on the outer surface of the aorta and range-gated across the lumen of the vessel. The problem with this range-gating method is that the velocity profile changes in size and configuration with increasing distance from the crystal, and the frequency results are amplitude dependent (the amplitude of the signal decreases with increasing distance from the probe). This present device appears to fulfill the criteria for a monitoring device to be used after cardiac operations
Volume 93 Number 5
May 1987
and has been successful in monitoring cardiac output after correction of congenital heart defects in 20 patients. The less than optimal association between ultrasound probe flow and electromagnetic probe flow in several of the 20 patients is undoubtedly due to an inferior Doppler signal, a problem that should be corrected as more precise manufacturing techniques are developed. Further testing is required to see whether the probe can be used to monitor cardiac output in adults after coronary artery bypass operations. Potential problems with bypass patients include difficulty in placing the probe because of the graft origins and the presence of turbulence in the area of the bypass grafts. In this report, there were minimal problems with probe application, although on some occasions the probe had to be repositioned several times. It is important to be sure that the flat portion of the probe is placed flush against the wall of the aorta, because this allows the maintenance of a 45 degree angle with the longitudinal axis of blood flow. The probe described in this report can constantly monitor the cardiac output in the postoperative period after cardiac operations. It can display a minuteby-minute recording of mean cardiac output as a function of time. This makes it extremely useful for following trends associated with fluid administration and making adjustments in intravenous drug therapy. REFERENCES I. Robinson PS, Crowther A, Jenkins BA. A computerized dichromatic earpiece densitometer for the measurement of cardiac output. J Cardiovasc Res 1979;13:420-6. 2. Landsman MLJ, Knop N, Mook GA, Zijlstra WG. A fiberoptic densitometer with cardiac output calculator. Pflugers Arch 1979;379:59-69. 3. Heneghan CPH, Branthwaite MA. Non-invasive measurement of cardiac output during anesthesia: an evaluation of the soluble gas uptake method. Br J Anaesth 1981;53:351-5. 4. Handt A, Farber MO, Szwed JJ. Intradialytic measure-
ment of cardiac output by thermodilution and impedance cardiography. Clin Nephrol 1977;7:61-4. 5. Farhi LE, Nesarajah MS, Olszowka LA, Metildi LA, Ellis AK. Cardiac output determination by simple onestep rebreathing technique. Respir Physiol 1976;28: 14159.
6. Bourdillon Pl, Becket 1M, Duffin PL. Saline conductivity method for measuring cardiac output simplified. Med BioI Eng Comput 1979;17:323-9. 7. Secher Nl, Thomsen A, Amsbo P. Measurement of rapid changes in cardiac stroke volume: an evaluation of the impedance cardiography method. Acta Anaesthesiol Scand 1977;21:353-8. 8. Katori R, Hatashi T, Kanamasa K, Ishikawa K. Mea-
Postoperative monitoring of cardiac output
663
surement of cardiac output by earpiece dye dilution method with automatic calibration of dye concentration. Tohoku J Exp Med 1977;122:51-8. 9. Huntsman LL, Stewart DK, Barnes SR, Franklin SB, Colocousis JS, Hessel EA. Noninvasive Doppler determination of cardiac output in man: clinical validation. Circulation 1983;67:593-602. 10. Alverson DC, Eldridge M, Dillon T, Yabek SM, Berman W Jr. Noninvasive pulsed Doppler determination of cardiac output in neonates and children. J Pediatr 1982;101:46-50. 11. Maruschak GF, Potter AM, Schauble JF, Rogers Me.
Overestimation of pediatric cardiac output by thermal dilution indicator loss. Circulation 1982;65:380-3. 12. Colgan FJ, Stewart S. An assessment of cardiac output by thermodilution in infants and children following cardiac surgery. Crit Care Med 1977;5:220-5. 13. Quinn K, Quebbeman EJ. Pulmonary artery pressure monitoring in the surgical intensive care unit. Arch Surg 1981;116:872-6. 14. Blackshear WM, Phillips DJ, Chikos PM, Harley JO,
Thiele BL, Strandness DE Jr. Carotid artery velocity patterns in normal and stenotic vessels. Stroke 1980; 11:67-71. 15. Blackshear WM, Phillips DJ, Thiele BL, et al. Detection
of carotid occlusive disease by ultrasonic imaging and pulsed Doppler spectrum analysis. Surgery 1979;86:698-
706. 16. Keagy BA, Pharr WF, Thomas D, Bowes DE. Objective
criteria for the interpretation of carotid artery spectral analysis patterns. Angiology 1982;33:213-20. 17. BodilyKC, Zierler RE, Marinelli MR, Thiele BL, Greene FM Jr, Strandness DE Jr. Flow disturbances following carotid endarterectomy. Surg Gynecol Obstet 1980; 151:77-80. 18. Lucas CL, Keagy BA, Hsiao HS, Johnson TA, Henry
GW, Wilcox BR. The velocity profile in the canine ascending aorta and its effects on the accuracy of pulsed Doppler determinations of mean blood velocity. Cardiavase Res 1984;18:282-93. 19. Greenfield JC, Patel DJ. Relation between pressure and diameter in the ascending aorta of man. Circ Res 1962;10:778-81. 20. Hsiao HS, Keagy BA, Lucas CL, Wilcox BR. Effects of
in vivo diameter changes on measurement of cardiac output in canine ascending aorta. J Cardiovasc Ultrasonogr 1983;2:395-400. 21. Keagy BA, Lucas CL, Hsiao HS, Wilcox BR: A removable extraluminal Doppler probe for continuous monitoring of changes in cardiac output. 1 Ultrasound Med 1983;2:357-62. 22. Lucas CL, Keagy BA, Hsiao HS, Wilcox BR. Software analysis of 20 MHz pulsed Doppler quadrature data. Ultrasound Med BioI 1983;9:641-55. 23. Ihlen H, Myhre E, Pamlie J, Forfang K, Larsen S.
Changes in left ventricular stroke volume measured by Doppler echocardiography. Br Heart J 1985;54:378-83.
The JoiJrnal of
6 6 4 Keagy et al.
24. Haites NE, McLennan FM, Mowat DH, Rawles JM. Assessment of cardiac output by the Doppler ultrasound technique alone. Br Heart J 1984;53:123-9. 25. Ascah KJ, Stewart WJ, Levine RA, Weyman AE. Doppler-echocardiographic assessment of cardiac output. Radiol Clin North Am 1985;23:659-70. 26. Walther FJ, Siassi B, Ramadan NA, Ananda AK, Wu PY. Pulsed Doppler determinations of cardiac output in neonates: normal standards for clinical use. Pediatrics 1985;76:829-33. 27. Rose JS, Nanna M, Rahimotoola SH, Elkayam U, McKay C, Chandraratna PA. Accuracy of determination of changes in cardiac output by transcutaneous continuous-wave Doppler computer. Am J CardioI1984;54:10991101. 28. Lavandier B, Cathignol D, Muchada R, Xan BB, Motin J. Noninvasive aortic blood flow measurement using an intraesophageal probe. Ultrasound Med Bioi 1985; 11:451-60. 29. Matre K, Segadal L, Engedal H. Continuous measurement of aortic blood velocity, after cardiac surgery, by means of an extractable Doppler ultrasound probe. J Biomed Eng 1984;7:84-8. 30. Baker DW. Pulsed ultrasonic Doppler blood flow sensing. IEEE Trans Sonics Ultrasonics 1970;SU-17:170.
Discussion DR. PHILLIP G. ASHMORE Vancouver, B, C. Canada
This is an interesting and important presentation because it focuses attention on the value of identifying trends in postoperative hemodynamics. We would agree that a means of assessing cardiac output in the early postoperative period might provide early warning of problems that might be prevented.
Thoracic and Cardiovascular Surgery
The advantage of the device described is that the probe is extraluminal. We have used the Oximetrix oximeter probe to provide similar information in patients with complex problems, and especially when we foresee the possibility of postoperative episodes of pulmonary hypertension. We introduce the probe into the pulmonary artery through the right ventricle. This allows us to follow the mixed venous saturation in the pulmonary artery and also to follow the pulmonary arterial pressure. Some years ago John Kirklin showed us that the mixed venous saturation in the pulmonary artery usually correlated well with the cardiac output, and our experience confirmsthis. A decrease in output can indicate a problem that often can be identified early and treated appropriately. In addition measurement of pulmonary artery pressure may help in dealing with the disturbing episodes of pulmonary hypertension seen after repair in some patients with major left-to-right shunts. I would endorse the concept of monitoring postoperative cardiac output and suggest that the indwelling Oximetrix probe is a useful alternative method of doing this and may provide additional information. I congratulate the authors on the ingenious method that they have developed and assessed.
DR. KEAGY (Closing) I thank Dr. Ashmore for his comments. We both recognize the need for monitoring cardiac output. All the other parameters that we monitor, such as urine output and blood pressures, are merely indirect reflectionsof cardiac output, and changes in them often occur later than the initial changes in cardiac output. We have not used the method that he described. We have had excellent results with our technique described, and our method also avoids placing any kind of foreign body in the bloodstream.
THoRAc CARDIOVASC SURG
1987;93:658-64
Constant postoperative monitoring of cardiac output after correction of congenital heart defects A new method has been developed that permits constant postoperative monitoring of mean and phasic cardiac output in patients after correction of congenital heart defects. A miniature ultrasound probe is attached to the adventitia of the ascending aorta at the conclusion of the operative procedure. This is connected to the monitoring equipment by means of polyurethane-covered wires that exit the chest wall through a small stab wound. The probe can easily be removed by gentle traction when the patient's condition is stable. The technique was developed, validated, and refined in extensive animal studies, and this report describes the first series of 20 consecutive human implants, performed between August 1984 and September 1985, in which the absolute cardiac output determination obtained with the ultrasound probe at the time of its application was correlated with cardiac output as measured with a standard electromagnetic flow probe. Fourteen male and six female patients (mean age 5.5 years) were studied. Operations performed included eight atrial septal defect repairs, four procedures for tetralogy of Fallot, three ventricular septal defect repairs, three stenotic valve corrections, and two Senning operations. One operative death occurred, but no complications were related to probe application or removal. The average cardiac output in the 20 patients as measured with the ultrasound probe was 2.2 ± 1.1 L/min (range 0.67 to 5.27 L/min). This is nearly identical to the results noted with the electromagnetic flow probe, where the mean cardiac output was 2.3 ± 1.2 L/min (range 0.7 to 6 L/min). Regression analysis revealed a high linear correlation (r = 0.9) between the two techniques. A monitor can display the cardiac output trend with 1 minute updates, which greatly enhance management of intravenous drug therapy and volume administration. In conclusion, this new extraluminal removable probe allows virtually continuous monitoring of the postoperative cardiac output after correction of congenital heart defects and should become a standard technique in the postoperative care of these patients.
Blair A. Keagy, M.D. (by invitation), Benson R. Wilcox, M.D., Carol L. Lucas, Ph.D. (by invitation), Henry S. Hsiao, Ph.D. (by invitation), G. William Henry, M.D. (by invitation), Michael Baudino, B.S. (by invitation), and Gene Bornzin, B.S. (by invitation), Chapel Hill. N. C; and Minneapolis, Minn.
A
knowledge of cardiac output is important in making appropriate patient care decisions after cardiac operations. A number of invasive and noninvasive techniques have been developed to measure this parameter.':" The standard against which most of these methods have been compared is the electromagnetic flow probe. Although the electromagnetic flow probe is acceptable for use in the experimental laboratory and in From the Department of Surgery, University of North Carolina, Chapel Hill, N. C., and Medtronic, Inc., Minneapolis, Minn. Read at the Sixty-sixth Annual Meeting of The American Association for Thoracic Surgery, New York, N. Y., April 28-30, 1986. Address for reprints: Blair A. Keagy, M.D., Divisionof Cardiothoracic Surgery, 108 Burnett-Womack Bldg., 229H, Chapel Hill, N. C. 27514.
658
the operating room, the size of the probe, the necessity for encircling the aorta with a rigid cuff, and the inability to remove the device outside of the operating room preclude its use in postoperative monitoring. Thermodilution is another means of monitoring cardiac output, but this technique fails to account for cardiac output fluctuations based on the respiratory cycle and provides only periodic information. In addition, some investigators have questioned the accuracy of this method. 1I • 13 More recently, ultrasound has been used to study cardiovascular hemodynamics. When an ultrasound beam is directed into the lumen of a blood vessel, its backscatter-or Doppler frequency shift by a moving column of red blood cells-is proportional to the velocity of blood flow. This relationship may be quantified by the
Volume 93
Postoperative monitoring of cardiac output
Number 5 May 1987
Dopplerequation, which states that the velocity of blood flow is directly proportional to the frequency of this reflected signal and inversely proportional to the cosine of the angle at which the ultrasound beam insonates the longitudinal axis of blood flOW. I4- 17 v=
cM 2focosO
where v = velocity of blood flow, c = average speed of sound in tissue (1.55 x 105 em/sec), M = frequency of the Doppler shifted signal, fo = frequency of the Doppler used, and () = angle of intersection between the Doppler beam and the longitudinal axis of the vessel. Two issues had to be addressed before an ultrasound device could be assumed to provide accurate cardiac output measurements. First, if velocity was sampled in only a small area of the ascending aorta and this value was used to extrapolate the velocity across the entire vessel, the cross-sectional velocity profile in the aorta had to be established. Second, since cardiac output is a product of velocity and cross-sectional area, it was necessary to establish a method of measuring accurately the area of the ascending aorta at the site where the velocity is being measured. In peripheral vessels, the cross-sectional velocity profile has been shown to be parabolic, with higher velocities occurring in the center of the vessel; however, the cross-sectionalvelocity profile in the ascending aorta has been a point of some controversy. To establish the shape of this profile,a tiny ultrasound crystal was placed on the tip of a small needle that was introduced into the lumen of the ascending aorta of a group of experimental animals.A siliconerubber sheath anchored the device to the wall of the aorta. The tip of the probe was moved from wall to wall inside the aortic lumen, and measurements were made away from the tip of the needle to avoidprobe-induced turbulence." This study established that the cross-sectional velocity profile in the ascending aorta is flat, and it validated the concept of using a small-velocity sampling area in the ascending aorta as an indicator of mean velocity across the entire vessel. Because volume flow (cardiac output) is a product of mean velocity in a vessel times its cross-sectional area, it is necessary to know the diameter of the ascending aorta in addition to the velocity of blood flow. With regard to changes in the aortic diameter, Greenfield and Patel" demonstrated only a 1.6% ± 0.5% change in aortic diameter for each 20 rom Hg change in aortic pressure in adult humans. This minimal change in cardiac output was confirmed in our animal laboratory by placing ultrasound crystals on opposite sides of the aorta and varying the mean systemic blood pressure by volume
659
administration and pharmacologic manipulation." The study showed that for normal systolic-diastolic differences of 35 rom Hg, the average diameter of the ascending aorta changed 1.04 rom, which resulted in a volume change of 17%. However, if the pulsar component in diameter change is ignored, and only the average blood velocity and average aortic area are used, a relatively small mean error of 4.83% occurs at normal aortic pressures. Thus a single static aortic diameter measurement made at the time of probe application may A) with introducbe used in the flow equation (Q = tion of minimal error. Requirements for a successful postoperative cardiac output monitoring device include ease of application, safety, accuracy, and the ability to remove the device without a second surgical operation. In an effort to achieve these goals, a high-frequency removable ultrasound probe was devised that contained a miniature 20 MHz piezoelectric crystal mounted at 45 degrees in an epoxy cuff. Two small metal tines were incorporated into the tip of the probe to anchor it to the adventitia of the ascending aorta. The probe samples velocity in a 1 mm' area of the ascending aorta. The probes were placed in experimental animals and the results of flow measurements were compared with those obtained simultaneously with an electromagnetic flow probe. The ultrasound probe correlated extremely well with values obtained with the electromagnetic probe (average r = 0.98 ± 0.006).21 At this point, probe construction was undertaken by Medtronic, Inc., Minneapolis, Minnesota, and a series of implantations in humans was begun. The present report describes this initial human study in which an ultrasound probe was used to measure cardiac output in the postoperative period after cardiac operations.
v.
Patients and methods
This study was performed between August 1984 and September 1985, during which time probes were implanted in 20 patients: 14 male and six female patients. The ages ranged from 3 months to 47 years, with a mean age of 5.5 years; 17 patients were under 6 years of age. Operations performed on these individuals included atrial septal defect repairs (eight), procedures for tetralogy of Fallot (four), ventricular septal defect repairs (three), stenotic valve corrections (three), and Senning procedures (two). Approval for human implantation of these probes was obtained from the internal review board of the North Carolina Memorial Hospital, and each patient or his/her family signed an informed consent form before the operation. At the conclusion of the operative procedure, before
The Journal of Thoracic and Cardiovascular
6 6 0 Keagy et al.
Surgery
Connector Cable /
~ P",b. H.ad
~PCA2r:'M~M:::~~"~ne, " .... . '. I •• ••
~
/1 Aortic Wall
Fig. 2. Schematic representation of the probe anchored to the aortic wall with the metal tines.
Fig. 1. Two-plane photographs of the actual ultrasound probe.
closure of the sternum, the cardiac output was measured with a traditional electromagnetic flow probe. The twin metal tines were then used to anchor the ultrasound probe to the adventitia of the ascending aorta. Fig. 1, A and B shows photographs of the probe, and Fig. 2 represents the anchoring tines embedded in the aortic wall. The diameter of the vessel was ascertained by direct measurement, and cardiac output was recorded with the ultrasound probe. The comparison of these two measurements is presented in the Results section. The probe exited through a small stab wound in the anterior chest wall and was anchored to the skin with a small silk suture. A fiberoptic cable was used to attach the probe to the pulsed Doppler unit and to the signal analyzer. Cardiac output was monitored with the ultrasound probe during the patient's immediate postoperative period, and when the patient's condition stabilized, the small silk suture was cut and the probe was removed by gentle traction. The tines slide easily out of the aortic wall. The quadrature frequency data used in this study were obtained with a 20 MHz pulsed Doppler system designed and built by Dr. Craig Hartley, Baylor College of Medicine. Systems specifications included a pulse repetition frequency of 62.5 kHz, a transmitter pulse width of 0.4 usee, a receiver pulse width of 0.3 usee, and a variable range gate of 1 to 12 mm. Output from the Doppler system was amplified at the patient's bedside before being transmitted via fiberoptic cables to a remote PDP 11/23 minicomputer system, where 64
complex samples were obtained every 1/66 second at a rate of 50 kHz using a Data Translation Model 2782 analog-to-digital converter. After 192 sampling bursts were obtained (2.9 seconds), each burst was subjected to Fourier analysis and the power spectrum computed. The 50 kHz range could be adjusted to fit the frequency content of the signal; however, the range of -15 kHz to +35 kHz was adequate for most applications. The Doppler shift (kHz) of the signal during each sampling burst was computed for the shifted power spectrum via a first moment average technique centered around a robust mean." The estimated flow rate during the burst was calculated by multiplying the resultant kilohertz shift measurement by a constant that adjusted for both the Doppler shift equation and the cross-sectional area of the vessel. The cardiac output and heart rate for the 2.9 second interval were computed and displayed on the computer monitor along with their respective averages for the preceding 5 minutes. The pulsatile blood flow waveform for the data just processed arid the trend in cardiac output for the preceding 2 hours were shown graphically. The screen was updated every minute. All results were stored on the computer, and a printed copy of the screen could be obtained as desired.
Result One postoperative death occurred and five complications including postoperative atelectasis (one), pleural effusion (two), postoperative myocardial infarct (one), and respiratory insufficiency (one). None of these problems was related to use of the probe. Probe application was easily accomplished and signals remained constant throughout the postoperative period in 15 patients. Two probes had weak but usable signals, and one patient required replacement of the probe two times in the operating suite because of a weak signal. In
Volume 93 Number 5
Postoperative monitoring of cardiac output 6 6 1
May 1987
7
en
8
I-
z
w i= 6 Cl:
3=
0
...J
D
u,
Do
a-
u.. 0
a: 4 w
z
45 Degrees (USP/EMF Y intercept
-
Line of Least Squares
:::l
l:D
:e :::l
-
(J)
= 1) =0
2
o
4
5
6
7
EMF FLOW 10·11
12·13
14·15
16·17
18·19
20+
AORTIC DIAMETER (MILLIMETERS)
(Regression Equation)
y = 0.423 + 0.809x
Fig. 4. Relationship between cardiac outputs as obtained by the ultrasound technique and by the electromagnetic flow (EMF) probe. USP, Ultrasound probe.
Fig. 3. Diameter measurements made at the time of probe application in all patients in this study.
one individual the signal was lost after 12 hours, and one neonate had a small aorta that made probe application difficult. Diameter measurements were made at operation (Fig. 3), and the aortic diameter ranged from 10 to 22 mm, with a mean of 14.9 ± 3.3 mm. The largest aorta, 22 mm, was present in a 47-year-old patient who underwent correction of an atrial septal defect. An electromagnetic flow probe measurement on the ascending aorta was performed immediately before application of the ultrasound probe in all patients. The average cardiac output in the 20 patients, as measured with the ultrasound probe, was 2.2 ± 1.1 Lzmin (range 0.67 to 5.27 L/min). This is nearly identical to the results noted with the electromagnetic flow probe, the mean cardiac output being 2.3 ± 1.2 L/min (range 0.7 to 6 L/min). Regression analysis revealed a high linear correlation (r = 0.9) between the two techniques. Fig. 4 is a scattergram of all comparative measurements in the study. The thin line represents the linear regression line, and the thick line indicates the ideal situation with the ultrasound probe/electromagnetic flow probe ratio equal to 1. The probes were left in place from 1 to 67 hours after operation (mean 29 ± 13.8 hours) (Fig. 5).
Discussion Ultrasound technology has been previously used in the noninvasiveassessment of cardiac output. Velocity in the ascending aorta was monitored by a scanning probe
applied to the suprasternal notch, and this value was combined with a noninvasive echocardiographic measurement of aortic diameter to compute cardiac outpUt.9, 23-27 Several problems are related to this suprastern~l notch approach. The first concerns the probable aortic arch location of the sample volume used for the frequency (velocity) measurement. In this area, the turbulence involved with origins of the great vessels, the lack of a flat velocity profile as the distance from the aortic valve increases, and the inability to measure accurately the exact angle between the insonant ultrasound beam and the longitudinal axis of blood flow place some doubt on the validity of the velocity measurement. The ultrasound beam is assumed to be parallel with the vessel axis; however, because the geometry of the aortic arch area varies, this assumption of a zero degree angle is fraught with potential error. A second problem is the accuracy of the static diameter measurement, which also may be affected by the angle of the ultrasound scan head and by an unknown magnification factor. It is probable that there is a high linear correlation between values obtained with a suprasternal notch probe and the true cardiac output if the probe is kept in the same position. However, reapplication of a Doppler scan head will usually result in placement of the sample volume in a different location, which subjects it to the errors described earlier. Also, longer distance between ultrasound probe and site of measurement means that a lower frequency must be used, which reduces the signal-to-noise ratio. Much better signals are possible with the removable aortic probe. A recently developed intraesophageal echo Doppler
The Journal of Thoracic and Cardiovascular Surgery
6 6 2 Keagy et al.
en ....z
w
j:::
«
a. LI.
12 10
8
0
6
l:ll
4
a: w
:E
:;)
z
2
< 1 DAY
1-2 DAYS
2-3 DAYS
Fig. 5. Duration of probe application in all patients in the
study. probe has also been used to measure cardiac output. Velocity is measured with a continuous-wave Doppler probe placed in the esophageal lumen, and an A-scan unit allows measurement of the diameter of the thoracic aorta." One of the problems associated with this method is the assumption that descending thoracic aortic flow is directly proportional to the cardiac output, an assumption that ignores the effects of the great vessels in the aortic arch. In addition, the assumed angle for the ultrasound crystal may be in error. After our 1983 report of an extractable ultrasound probe in animals, a version of the probe for use in humans was reported by Matre, Segadal, and Engedal. 29 This device was placed under the adventitia of the ascending aorta after cardiac procedures, and changes in aortic velocity were recorded up to 24 hours postoperatively in 14 patients. Good correlation with thermodilution techniques (r = 0.79) was described. The authors concluded that this device allowed continuous monitoring of cardiac output, as well as short- and long-term trend analysis during the early postoperative period. Implantable ultrasound velocity probes have a number of advantages, including the lack of a baseline drift, a relative independence from hematocrit, and ease of removal when no longer needed. In addition, a constant angle of insonance (45 degrees) is maintained between the ultrasound beam and the longitudinal axis of blood flow, and the same sample volume site is used in all velocity calculations. This corrects one of the problems associated with the suprasternal notch approach, which requires repeated probe applications. In the system described in this report, a pulsed Doppler probe has been used to interrogate the ascending aorta. A pulsed Doppler probe" has one piezoelectric crystal that alternately acts as an ultrasonic transmitter
and receiver. Varying the time interval between these two functions (range gating) allows accurate placement of the Doppler sample volume at various locations in the lumen of the vessel. Also, varying the length of the ultrasound burst allows the size of the sample volume to be controlled within certain limits. This differs from a more commonly used continuous-wave Doppler probe, which contains two piezoelectric crystals, one constantly functioning as the transmitter of an ultrasound beam and the other acting as a continuous receiver. The continuous-wave Doppler probe has the advantage of a much larger sample volume, but the theoretical disadvantage of sampling other surrounding vascular structures such as accompanying veins. In addition, the need for two crystals requires a larger probe tip. The flat velocity profile in the ascending aorta allows the use of the pulsed Doppler techniques. Three additional probe implantations have been performed since this report was submitted. However, the electromagnetic flow probe correlation was not obtained because the validity of the device has been established. One of these patients had a cardiac arrest in close temporal proximity to the probe removal and underwent open chest resuscitation. The patient recovered without incident and was discharged. There was no blood around the site of attachment of the probe on the ascending aorta, and it was postulated that perhaps this episode was related to a severe hypercyanotic ("Tet") spell in this infant who had an uncorrected tetralogy of Fallot. In the remaining two patients the probe was removed without incident. The reflected aortic ultrasound signal can be evaluated with either fast Fourier transform sound-spectral analysis or zero crosser hardware. The latter method involves less expensive equipment and the results can be displayed faster, because extensive computer time is not required for signal analysis. However, the zero crosser becomes inaccurate in the presence of turbulence or with high signal frequencies. There has been controversy in the past about the velocity profile in the ascending aorta, but we think that our method of measuring the velocity profile is better than those of other investigators who have placed a probe on the outer surface of the aorta and range-gated across the lumen of the vessel. The problem with this range-gating method is that the velocity profile changes in size and configuration with increasing distance from the crystal, and the frequency results are amplitude dependent (the amplitude of the signal decreases with increasing distance from the probe). This present device appears to fulfill the criteria for a monitoring device to be used after cardiac operations
Volume 93 Number 5
May 1987
and has been successful in monitoring cardiac output after correction of congenital heart defects in 20 patients. The less than optimal association between ultrasound probe flow and electromagnetic probe flow in several of the 20 patients is undoubtedly due to an inferior Doppler signal, a problem that should be corrected as more precise manufacturing techniques are developed. Further testing is required to see whether the probe can be used to monitor cardiac output in adults after coronary artery bypass operations. Potential problems with bypass patients include difficulty in placing the probe because of the graft origins and the presence of turbulence in the area of the bypass grafts. In this report, there were minimal problems with probe application, although on some occasions the probe had to be repositioned several times. It is important to be sure that the flat portion of the probe is placed flush against the wall of the aorta, because this allows the maintenance of a 45 degree angle with the longitudinal axis of blood flow. The probe described in this report can constantly monitor the cardiac output in the postoperative period after cardiac operations. It can display a minuteby-minute recording of mean cardiac output as a function of time. This makes it extremely useful for following trends associated with fluid administration and making adjustments in intravenous drug therapy. REFERENCES I. Robinson PS, Crowther A, Jenkins BA. A computerized dichromatic earpiece densitometer for the measurement of cardiac output. J Cardiovasc Res 1979;13:420-6. 2. Landsman MLJ, Knop N, Mook GA, Zijlstra WG. A fiberoptic densitometer with cardiac output calculator. Pflugers Arch 1979;379:59-69. 3. Heneghan CPH, Branthwaite MA. Non-invasive measurement of cardiac output during anesthesia: an evaluation of the soluble gas uptake method. Br J Anaesth 1981;53:351-5. 4. Handt A, Farber MO, Szwed JJ. Intradialytic measure-
ment of cardiac output by thermodilution and impedance cardiography. Clin Nephrol 1977;7:61-4. 5. Farhi LE, Nesarajah MS, Olszowka LA, Metildi LA, Ellis AK. Cardiac output determination by simple onestep rebreathing technique. Respir Physiol 1976;28: 14159.
6. Bourdillon Pl, Becket 1M, Duffin PL. Saline conductivity method for measuring cardiac output simplified. Med BioI Eng Comput 1979;17:323-9. 7. Secher Nl, Thomsen A, Amsbo P. Measurement of rapid changes in cardiac stroke volume: an evaluation of the impedance cardiography method. Acta Anaesthesiol Scand 1977;21:353-8. 8. Katori R, Hatashi T, Kanamasa K, Ishikawa K. Mea-
Postoperative monitoring of cardiac output
663
surement of cardiac output by earpiece dye dilution method with automatic calibration of dye concentration. Tohoku J Exp Med 1977;122:51-8. 9. Huntsman LL, Stewart DK, Barnes SR, Franklin SB, Colocousis JS, Hessel EA. Noninvasive Doppler determination of cardiac output in man: clinical validation. Circulation 1983;67:593-602. 10. Alverson DC, Eldridge M, Dillon T, Yabek SM, Berman W Jr. Noninvasive pulsed Doppler determination of cardiac output in neonates and children. J Pediatr 1982;101:46-50. 11. Maruschak GF, Potter AM, Schauble JF, Rogers Me.
Overestimation of pediatric cardiac output by thermal dilution indicator loss. Circulation 1982;65:380-3. 12. Colgan FJ, Stewart S. An assessment of cardiac output by thermodilution in infants and children following cardiac surgery. Crit Care Med 1977;5:220-5. 13. Quinn K, Quebbeman EJ. Pulmonary artery pressure monitoring in the surgical intensive care unit. Arch Surg 1981;116:872-6. 14. Blackshear WM, Phillips DJ, Chikos PM, Harley JO,
Thiele BL, Strandness DE Jr. Carotid artery velocity patterns in normal and stenotic vessels. Stroke 1980; 11:67-71. 15. Blackshear WM, Phillips DJ, Thiele BL, et al. Detection
of carotid occlusive disease by ultrasonic imaging and pulsed Doppler spectrum analysis. Surgery 1979;86:698-
706. 16. Keagy BA, Pharr WF, Thomas D, Bowes DE. Objective
criteria for the interpretation of carotid artery spectral analysis patterns. Angiology 1982;33:213-20. 17. BodilyKC, Zierler RE, Marinelli MR, Thiele BL, Greene FM Jr, Strandness DE Jr. Flow disturbances following carotid endarterectomy. Surg Gynecol Obstet 1980; 151:77-80. 18. Lucas CL, Keagy BA, Hsiao HS, Johnson TA, Henry
GW, Wilcox BR. The velocity profile in the canine ascending aorta and its effects on the accuracy of pulsed Doppler determinations of mean blood velocity. Cardiavase Res 1984;18:282-93. 19. Greenfield JC, Patel DJ. Relation between pressure and diameter in the ascending aorta of man. Circ Res 1962;10:778-81. 20. Hsiao HS, Keagy BA, Lucas CL, Wilcox BR. Effects of
in vivo diameter changes on measurement of cardiac output in canine ascending aorta. J Cardiovasc Ultrasonogr 1983;2:395-400. 21. Keagy BA, Lucas CL, Hsiao HS, Wilcox BR: A removable extraluminal Doppler probe for continuous monitoring of changes in cardiac output. 1 Ultrasound Med 1983;2:357-62. 22. Lucas CL, Keagy BA, Hsiao HS, Wilcox BR. Software analysis of 20 MHz pulsed Doppler quadrature data. Ultrasound Med BioI 1983;9:641-55. 23. Ihlen H, Myhre E, Pamlie J, Forfang K, Larsen S.
Changes in left ventricular stroke volume measured by Doppler echocardiography. Br Heart J 1985;54:378-83.
The JoiJrnal of
6 6 4 Keagy et al.
24. Haites NE, McLennan FM, Mowat DH, Rawles JM. Assessment of cardiac output by the Doppler ultrasound technique alone. Br Heart J 1984;53:123-9. 25. Ascah KJ, Stewart WJ, Levine RA, Weyman AE. Doppler-echocardiographic assessment of cardiac output. Radiol Clin North Am 1985;23:659-70. 26. Walther FJ, Siassi B, Ramadan NA, Ananda AK, Wu PY. Pulsed Doppler determinations of cardiac output in neonates: normal standards for clinical use. Pediatrics 1985;76:829-33. 27. Rose JS, Nanna M, Rahimotoola SH, Elkayam U, McKay C, Chandraratna PA. Accuracy of determination of changes in cardiac output by transcutaneous continuous-wave Doppler computer. Am J CardioI1984;54:10991101. 28. Lavandier B, Cathignol D, Muchada R, Xan BB, Motin J. Noninvasive aortic blood flow measurement using an intraesophageal probe. Ultrasound Med Bioi 1985; 11:451-60. 29. Matre K, Segadal L, Engedal H. Continuous measurement of aortic blood velocity, after cardiac surgery, by means of an extractable Doppler ultrasound probe. J Biomed Eng 1984;7:84-8. 30. Baker DW. Pulsed ultrasonic Doppler blood flow sensing. IEEE Trans Sonics Ultrasonics 1970;SU-17:170.
Discussion DR. PHILLIP G. ASHMORE Vancouver, B, C. Canada
This is an interesting and important presentation because it focuses attention on the value of identifying trends in postoperative hemodynamics. We would agree that a means of assessing cardiac output in the early postoperative period might provide early warning of problems that might be prevented.
Thoracic and Cardiovascular Surgery
The advantage of the device described is that the probe is extraluminal. We have used the Oximetrix oximeter probe to provide similar information in patients with complex problems, and especially when we foresee the possibility of postoperative episodes of pulmonary hypertension. We introduce the probe into the pulmonary artery through the right ventricle. This allows us to follow the mixed venous saturation in the pulmonary artery and also to follow the pulmonary arterial pressure. Some years ago John Kirklin showed us that the mixed venous saturation in the pulmonary artery usually correlated well with the cardiac output, and our experience confirmsthis. A decrease in output can indicate a problem that often can be identified early and treated appropriately. In addition measurement of pulmonary artery pressure may help in dealing with the disturbing episodes of pulmonary hypertension seen after repair in some patients with major left-to-right shunts. I would endorse the concept of monitoring postoperative cardiac output and suggest that the indwelling Oximetrix probe is a useful alternative method of doing this and may provide additional information. I congratulate the authors on the ingenious method that they have developed and assessed.
DR. KEAGY (Closing) I thank Dr. Ashmore for his comments. We both recognize the need for monitoring cardiac output. All the other parameters that we monitor, such as urine output and blood pressures, are merely indirect reflectionsof cardiac output, and changes in them often occur later than the initial changes in cardiac output. We have not used the method that he described. We have had excellent results with our technique described, and our method also avoids placing any kind of foreign body in the bloodstream.