Anesth Analg 2005;100:748-753
© 2005 International Anesthesia Research Society
doi: 10.1213/01.ANE.0000144774.42408.05
TECHNOLOGY, COMPUTING, AND SIMULATION
Testing the Reliability of a New Ultrasonic Cardiac Output Monitor, the USCOM, by Using Aortic Flowprobes in Anesthetized Dogs
Lester A. Critchley, MD, FFARCSI*,
Zhi Y. Peng, MD, PhD*,
Benny S. Fok, BSc*,
Anna Lee, PhD, MPH*, and
Robert A. Phillips, FIR, DMU, AMS
*Department of Anaesthesia and Intensive Care, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, Hong Kong, China; and School of Medicine, University of Queensland, Brisbane, Queensland, Australia
Address correspondence and reprint requests to Lester A. Critchley, MD, FFARCSI, Department of Anesthesia and Intensive Care, Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong, China. Address e-mail to hcritchley{at}cuhk.edu.hk.
 |
Abstract
|
|---|
We have used an animal model to test the reliability of a new portable continuous-wave Doppler ultrasonic cardiac output monitor, the USCOM. In six anesthetized dogs, cardiac output was measured with a high-precision transit time ultrasonic flowprobe placed on the ascending aorta. The dogs’ cardiac output was increased with a dopamine infusion (0–15 µg · kg–1 · min–1). Simultaneous flowprobe and USCOM cardiac output measurements were made. Up to 64 pairs of readings were collected from each dog. Data were compared by using the Bland and Altman plot method and Lin’s concordance correlation coefficient. A total of 319 sets of paired readings were collected. The mean (±sd) cardiac output was 2.62 ± 1.04 L/min, and readings ranged from 0.79 to 5.73 L/min. The mean bias between the 2 sets of readings was –0.0l L/min, with limits of agreement (95% confidence intervals) of –0.34 to 0.31 L/min. This represents a ±13% error. In five of six dogs, there was a high degree of concordance, or agreement, between the 2 methods, with coefficients >0.9. The USCOM provided reliable measurements of cardiac output over a wide range of values. Clinical trials are needed to validate the device in humans.
|
Introduction
|
|---|
The two variables that most fundamentally describe the performance of the heart and systemic circulation are arterial blood pressure and cardiac output. Whereas arterial blood pressure is a homogenous variable throughout the arterial system and is easily measured, blood flow is rapidly dispersed throughout the circulation and is difficult to measure. It can be directly measured only at its origin, the heart and great vessels, which are difficult to access because these structures are situated deep in the thorax. Current methods used to measure cardiac output in a clinical setting either are too invasive or too difficult to use (such as thermodilution, dye dilution, and the Fick method) (1,2) or are too unreliable (such as bioimpedance) (3,4) to be used routinely. Hence, physicians often base their management of the circulation on arterial blood pressure and heart rate changes and ignore what may be happening to the cardiac output.
Thus, there is a real clinical need for the development of a simple, safe, and reliable method of measuring cardiac output at the bedside (2). One noninvasive method of measuring cardiac output that has attracted some attention over the years is continuous-wave (CW) Doppler. It usually involves insonating the aorta via the sternal notch. The early enthusiasm for the development of this method was dampened by inconsistent performances (5–7). The main problems were 1) inaccurate assessments of the aortic diameter, 2) failure to precisely determine the representative signal of blood flow across and along the aorta, 3) incorrectly defining the ejection period, and 4) angle deviations between the ultrasound beam and the axis of the vessel (8). These early attempts to develop a reliable cardiac output monitor were superseded by more invasive ultrasound techniques that imaged heart motion rather than blood flow detection, such as transesophageal echocardiography. The early CW Doppler devices were really quite simple, and 20 yr of improvements in transducer technology, signal processing, and software developments to enhance user capability have made clinically reliable noninvasive cardiac output monitoring by CW Doppler a distinct possibility. In particular, piezoelectric technology has improved, and transducers can be optimized to specifically measure the CW Doppler signal when there is no need to perform an imaging function. There is a tradeoff between imaging and Doppler response. Most large ultrasound devices, including transesophageal echocardiography, favor imaging at the expense of the Doppler signal.
The ultrasonic cardiac output monitor, or USCOM (USCOM Pty Ltd., Sydney, NSW, Australia), was developed in 2002 with the sole objective of measuring cardiac output (Fig. 1; http://www.uscom.com.au). It is based on the latest CW Doppler ultrasound technology. Although ultrasound imaging has a long and proven record in cardiac diagnosis, the USCOM is one of the first recently developed systems to be commercially produced that is dedicated solely to cardiac output monitoring.
View larger version (127K):
[in this window]
[in a new window]
|
Figure 1. The USCOM with Doppler probe and cable. The display shows real-time flow profiles, which have been saved. A selected profile has been outlined by using a touch-point pen (shown) to facilitate the calculation of cardiac output.
|
|
The objective of our study was to evaluate the reliability of USCOM measurements over a wide range of cardiac outputs. We used an dog model because reference cardiac output measurements could be accurately performed with a flowprobe placed on the aorta. Furthermore, a wide range of values of cardiac outputs could be pharmacologically produced.
|
Methods
|
|---|
Ethical approval for the study was obtained from the Animal Research Ethics Committee of the Chinese University of Hong Kong. Male mongrel dogs were provided by the Laboratory Animals Service Center of The Chinese University of Hong Kong. Anesthesia was induced with IM ketamine 10% (5 mg/kg) and xylazine 2% (2 mg/kg) and was maintained throughout the experiment with inhaled halothane 0.5%–1.0% in oxygen. The trachea was intubated, and the dog lungs were ventilated with a tidal volume of 10–15 mL/kg at a frequency of 12–15 breaths/min. Muscle relaxation was provided by increments of rocuronium (5 mg). IV access was secured in the forelimb and used to administer IV fluids (warmed saline 0.9% 2 mL · kg–1 · h–1) and drugs, which were infused with a syringe pump. The right femoral artery was cannulated to allow arterial blood pressure recordings. The left internal jugular vein was surgically exposed and cannulated to measure central venous pressure, which was kept constant by infusing additional volumes (100-mL increments) of warmed saline 0.9%. Body temperature was maintained by covering the dog with an insulated blanket.
A thoracotomy was performed at the left fourth intercostal space. The pericardium was incised longitudinally to expose the aortic root. The ascending aorta was separated from the pulmonary artery by 2–3 cm by blunt dissection with the finger. The free fat surrounding the aorta was carefully removed. A snugly fitting flowprobe—a 12-, 16-, or 20-mm A-series ultrasonic probe (Transonic Systems Inc., Ithaca, NY)—was placed around the ascending aorta, and ultrasonic gel was applied. The probe cable was brought out of the thorax posteriorly. The pericardium was closed with sutures, a chest drain connected to an underwater seal was inserted, the collapsed lung was reexpanded, and the chest wall was closed with sutures.
Aortic blood flow was measured by the flowprobe, which used a high-precision four-crystal array. It was connected to a T106 single-channel flowmeter (Transonic Systems) that also processed the transduced arterial blood pressure wave. A laptop computer collected the data and displayed it by using the software program WinDaq (DataQ Instruments, Akron, OH).
The USCOM uses a small handheld state-of-the-art piezoelectric CW Doppler ultrasound probe to insonate the aortic outflow via the thoracic inlet (Fig. 1), although the pulmonary artery, via the anterior chest wall, could also be used. The probe uses a wide acoustic beam to allow blood flow to be more easily tracked. The probe first needs to be used to explore the thoracic inlet to locate the source of optimum signal that originates from the aortic valve. This is a skill that takes time to acquire. The Doppler flow profile is displayed in real time on a robust touch screen (Fig. 1). The area of each flow profile corresponds to the velocity time integral (VTI). By outlining the profile with a simple five-point polygon touch-screen method and calculating its area, VTI is measured (Fig. 2). Stroke volume (SV) is calculated from the equation
View larger version (175K):
[in this window]
[in a new window]
|
Figure 2. Flow profiles from the USCOM. The main systolic flow profile is outlined, and the distance between profiles is marked by the operator. This enables the device to measure the velocity time integral and heart rate.
|
|
where CSA is the cross-sectional area of the aortic valve (9). The cosine of the angle of insonation (0–90°) is considered to be approximately 1.0. By defining the interval between consecutive profiles, the heart rate is derived and cardiac output is calculated (Fig. 2).
The USCOM was calibrated by inputting the diameter of the aorta, measured during the thoracotomy. In humans, this is done by inputting the subject’s height and using a formula to estimate the aortic valve size (10). Ultrasound gel was applied to the skin over the dog’s thoracic inlet, and the upper thorax was insonated with a 3.3-MHz CW Doppler ultrasound probe. The wave form of the flow profile from the aorta was identified on the monitor screen, and the probe position was held. Cardiac output was then measured as described above. The mean of three consecutive cardiac output measurements was used in the final analysis.
Simultaneous flowprobe and USCOM cardiac output measurements were made. The investigator performing the USCOM measurements was unaware of the flowprobe readings. The dog’s cardiac output was first increased in steps of 3 to 5 min by infusing dopamine (5–15 µg · kg–1 · min–1) at increasing rates. The cardiac output was then decreased by stopping the infusion and giving large concentrations of inhaled halothane (2%–4%). In the first dog experiment, we collected only 21 pairs of readings (learning phase). In subsequent experiments, we aimed to collect up to 64 pairs of readings. Data collection lasted 1–2 h.
Statistical analysis was performed with SPSS Version 11.5 (SPSS Inc., Chicago, IL) and Stata Version 8.2 (Stata Corp., College Station, TX) for Lin’s concordance. The Bland and Altman plot method (11) was used to assess agreement between the flowprobe (reference standard) and the USCOM measurements of cardiac output. The plot is a graphical representation of the data with the between-method difference (y axis) plotted against the average of the data (x axis). Bias is the mean difference between the two methods of measurement and represents the systematic error. The limits of agreement were defined as mean ± 1.96 sd and represent the range within which most differences between measurements by the two methods will lie (12). The method was modified to allow for unequal groups of replicates (11). Only the mean values from each experiment are shown, and the limits of agreement have been adjusted to compensate for differences in group size. This enables one to visually compare the calibration bias for each experiment. The two methods of measuring cardiac output were judged to be interchangeable if the limits of agreement did not exceed the threshold, set a priori at ±0.5 L/min or 20% of the mean cardiac output.
Total peripheral resistance was calculated from the mean arterial blood pressure divided by the flowprobe cardiac output. The total peripheral resistance (x axis) was then plotted against the bias (y axis) to determine whether changes in peripheral vascular resistance had a significant influence on USCOM measurements. A correlation and regression analysis was performed.
Lin’s concordance correlation coefficient (
) was estimated to evaluate the degree of reproducibility between the flowprobe and the USCOM for measuring cardiac output for each dog. This index is the correlation between the measurements that occur on the 45° line, or line of identity, through the origin (13).
|
Results
|
|---|
Data were collected from 6 male dogs aged 3–7 yr and weighing 11–22 kg. Aortic diameter ranged from 12 to 20 mm. Data collection took between 73 and 125 min.
A total of 319 sets of cardiac output comparisons were collected. Cardiac output measurements ranged from 0.90 to 5.60 L/min and 0.79 to 5.73 L/min for the flowprobe and USCOM, respectively. The overall mean ± sd cardiac outputs measured by the flowprobe and the USCOM were 2.60 ± 0.99 L/min and 2.62 ± 1.04 L/min, respectively. The mean bias was –0.01 L/min, and the 95% limits of agreement were –0.34 to 0.31 L/min (Fig. 3). Thus, a measurement by USCOM was unlikely to exceed a measurement by the flowprobe by more than 0.31 L/min or to be >0.34 L/min below it. This represented an error (1.96 sd/mean) for USCOM measurements of up to ±13%. There was no evidence that peripheral vascular resistance significantly affected the result.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 3. Bland and Altman plot of the mean difference or bias between the two sets of cardiac output measurements against their mean for each experiment (n = 6). The limits of agreement for the data are shown by the dashed lines, allowing for unequal groups of replicates.
|
|
There was high reproducibility in the measurements of cardiac output by the 2 methods in 5 of the 6 dogs, with coefficients of concordance () >0.9 (Table 1). A wide range of cardiac outputs was tested in each dog and were increased by up to 75%–260% of the initial value (Fig. 4). Individual USCOM measurements tended to lie on, or close to, the lines of identity, thus indicating a high degree of agreement with their corresponding flowprobe measurements.
View larger version (21K):
[in this window]
[in a new window]
|
Figure 4. Scatter plots for all six experiments comparing cardiac output (CO) measurements by the two methods. The line of identity, 45°, is shown.
|
|
|
Discussion
|
|---|
We were able to test the USCOM over a wide range of cardiac output values by using an animal model. The limits of agreement when compared with the flowprobe were small (±0.3 L/min, compared with 2.6 L/min for mean cardiac output). In 5 of 6 dogs the concordance was high, >0.9, and the overall percentage error was small (±13%). No similar data have been published evaluating the USCOM.
When a new method is compared with an established measurement technique, the primary aim is to determine whether the two methods agree sufficiently to be considered interchangeable (12). Because analysis with correlation and regression can be misleading, Bland and Altman (11) favored a different analytical approach based on calculating the bias and limits of agreement. If the limits of agreement are small and not considered clinically significant, then the two methods are interchangeable. In this study, we found that the USCOM data fulfilled these criteria.
Thermodilution has been the established method for many years of measuring cardiac output in a clinical setting. However, the accuracy of individual thermodilution measurements is not ideal and is approximately 10%–20%; these figures are similar for other modalities of cardiac output monitoring (14,15). On the basis of this level of accuracy and criteria established previously by our group, the limits of agreement for the interchangeability of thermodilution with a new technique are approximately 14%–28% (15). In a clinical setting, one would desire an accuracy of 10%, which is similar to that for noninvasive arterial blood pressure measurements (16,17). Thus, the error of 13% found in this study would appear to be clinically very acceptable. However, in this study we used a high-precision flowprobe with a quoted accuracy of 1%–2% (18,19), rather than the less accurate thermodilution method. Thus, the principle of interchangeability does not strictly apply to our data, and the 13% approaches the measurement error of the USCOM. However, when the noninvasive nature of the device is considered, the USCOM does compare very favorably with other noninvasive technologies, such as bioimpedance and older-generation Doppler technology, in which the quoted limits of agreement have been 30% or more (15). Admittedly, these figures have been judged against less reliable standards, such as thermodilution. Thus, our data support the use of the USCOM as a noninvasive clinical monitor of cardiac output.
In addition to evaluating precision, we also analyzed the ability of the USCOM to measure cardiac output over a range of values. Lin’s concordance was used in preference to traditional regression analysis because it gave a better indication of the device’s ability to track changes in cardiac output with respect to the line of identity (13). We found a high concordance in 5 of 6 dogs, indicating that in most cases the device reliably tracks changes in cardiac output.
Several factors were present that may have influenced the accuracy of the flowprobe measurements in Experiment 4. Cardiac output fluctuates between cardiac cycles; the most obvious cause is positive-pressure lung ventilation, in which cardiac output is impeded during the inspiratory phase. Hence, a failure to perform cardiac output comparisons simultaneously could lead to an increased error. Furthermore, if the flowprobe does not properly fit the aorta, its lumen may become distorted, and this results in turbulent, rather than laminar, flow. Turbulent flow does alter the flowprobe measurements. Similarly, pathologic defects of the aortic outflow tract, such as valve defects, may also affect measurements. Furthermore, in one dog that was not included in the study it was impossible to apply the flowprobe to the aorta because of abnormal enlargement. In a number of dogs, the heart rate also became irregular at the larger doses of dopamine. Because the USCOM calculates cardiac output by using one Doppler flow profile and the interpeak distance, irregularity in heart rate will result in beat-to-beat variations in cardiac output estimates. To overcome this problem, data were averaged over three consecutive readings. There were also a number of limitations with our use of the USCOM probe. During insonation, the aortic root was explored, and signals from a number of vessel sources were identified. The dominant signal was from the aortic valve. However, it is possible to insonate the wrong vessel or the wrong region of the aorta because of a lack of experience. Thus, a learning curve is anticipated with the USCOM.
The CSA of the aorta plays an important role in the calibration of the USCOM and in Doppler methods in general (9). In this study, the diameter of the aorta was measured directly during the placement of the flowprobe and was input into the device during calibration. In all but one experiment (Dog 4), the bias between the flowprobe and USCOM measurements was small and <3%, indicating that the USCOM and flowprobe measurements closely corresponded, when both measurements were correctly calibrated. However, when the USCOM is used in human subjects, the diameter of the aorta is estimated from a normogram that is based on the subject’s height. This normogram is derived from the Nidorf et al. (10) equation, which was rigorously evaluated by one of the authors (RAP) and was found to provide the most reliable estimate of valve dimensions. The 95% confidence intervals on the normogram represent a ±10%–20% variation in aortic diameters. Even small discrepancies in diameter—or, more correctly, radius, as CSA is related to 
r2—can cause quite large systematic errors in the cardiac output measurements. Thus, normogram-based estimates of the CSA of the aorta are bound to introduce some systematic error into the measurement of cardiac output in the clinical setting.
In older CW Doppler devices, such as the Datascope Accucom, blood flow was measured from the aortic arch, which distended during systole and altered the CSA (8). This variation in CSA affected the accuracy of such devices (8). The USCOM overcomes this problem by measuring the velocity profile at the aortic and pulmonary valves, which remain constant in diameter.
In conclusion, our data collected with a dog model indicate that the USCOM provides reliable cardiac output measurements over a wide range of values in dogs. Thus, the device has the potential to fill a much-needed gap in current clinical monitoring. Clinical trials are now needed to verify this potential in humans.
We thank Pacific Medical Systems Limited, Hong Kong, which provided the USCOM machine that was used in the study, and, in particular, Jules Flach and Simon C. Wong, who assisted with data collection.
|
References
|
|---|
- Connors AF, Speroff T, Dawson NV, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996;276:889–97.[Abstract]
- Van De Water JM, Miller TW, Vogel RL, et al. Impedance cardiography: the next vital sign technology? Chest 2003;123:2028–33.[Abstract/Free Full Text]
- Clarke DE, Raffin TA. Thoracic electrical bioimpedance measurement of cardiac output: not ready for prime time. Crit Care Med 1993;21:1111–2.[ISI][Medline]
- Weil MH. Electrical bioimpedance for noninvasive measurement of cardiac output. Crit Care Med 1997;25:1455.[ISI][Medline]
- Perrino AC Jr, Fleming J, LaMantia KR. Transesophageal Doppler ultrasonography: evidence for improved cardiac output monitoring. Anesth Analg 1990;71:651–7.[Abstract/Free Full Text]
- Spahn DR, Schmid ER, Tornic M, et al. Noninvasive versus invasive assessment of cardiac output after cardiac surgery: clinical validation. J Cardiothorac Anesth 1990;4:46–59.[Medline]
- Wong DH, Tremper KK, Stemmer EA, et al. Noninvasive cardiac output: simultaneous comparison of two different methods with thermodilution. Anesthesiology 1990;72:784–92.[ISI][Medline]
- Schmid ER, Spahn DR, Tornic M. Reliability of a new generation transesophageal Doppler device for cardiac output monitoring. Anesth Analg 1993;77:971–9.[Abstract/Free Full Text]
- Huntsman LL, Stewart DK, Barnes SR, et al. Noninvasive Doppler determination of cardiac output in man: clinical validation. Circulation 1983;67:593–602.[Abstract/Free Full Text]
- Nidorf SM, Picard MH, Triulzi MO, et al. New perspectives in the assessment of cardiac chamber dimensions during development and adulthood. J Am Coll Cardiol 1992;19:983–8.[Abstract]
- Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–60.[Abstract/Free Full Text]
- Mantha S, Roizen MF, Fleisher LA, et al. Comparing methods of clinical measurement: reporting standards for Bland and Altman analysis. Anesth Analg 2000;90:593–602.[Abstract/Free Full Text]
- Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989;45:255–68.[ISI][Medline]
- Stetz CW, Miller RG, Kelly GE, Raffin TA. Reliability of the thermodilution method in the determination of cardiac output in clinical practice. Am Rev Respir Dis 1982;126:1001–4.[ISI][Medline]
- Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput 1999;15:85–91.[ISI][Medline]
- Edmonds ZV, Mower WR, Lovato LM, Lomeli R. The reliability of vital sign measurements. Ann Emerg Med 2002;39:233–7.[ISI][Medline]
- Rutten AJ, Ilsley AH, Skowronski GA, Runciman WB. A comparative study of the measurement of mean arterial blood pressure using automatic oscillometers, arterial cannulation and auscultation. Anaesth Intensive Care 1986;14:58–65.[ISI][Medline]
- Bednarik JA, May CN. Evaluation of a transit-time system for the chronic measurement of blood flow in conscious sheep. J Appl Physiol 1995;78:524–30.[Abstract/Free Full Text]
- Dean DA, Jia CX, Cabreriza SE, et al. Validation study of a new transit time ultrasonic flow probe for continuous great vessel measurements. ASAIO J 1996;42:M671–6.[ISI][Medline]
Top
Abstract
Introduction
