Assessment of cardiac output (CO) is crucial in the management of the critically ill, especially in post cardiac surgery intensive care unit (ICU) patients.1,2 Thermodilution method introduced by Ganz and coworkers3 in 1970 is the current clinical standard to measure CO. This method requires insertion of a pulmonary artery catheter (PAC) and most commonly cold saline is injected as a bolus through the PAC into the right atrium for CO measurements. However, PAC is associated with complications related to their invasiveness, such as pneumothorax, arterial puncture, arrhythmias, bacteremia, and hence its use is under constant clinical criticism.4,5 Current less-invasive methods have limitations such as use of nonphysiologic indicator and blood loss (LiDCO6) or require insertion of a dedicated femoral artery catheter for transpulmonary thermodilution (Pulsion7). Thus, there exists no simple and safe method to routinely measure CO in ICU patients. This is especially true for pediatric and neonatal ICU patients for whom small blood vessels create additional problems.
A novel COstatus system (Transonic Systems Inc., Ithaca, NY) based on ultrasound dilution (UD) technology has been developed8 to measure CO and blood volume parameters in ICU patients using an extracorporeal arteriovenous (AV) loop connected between in situ arterial and central venous catheters. The purpose of this study was to compare CO measurements by ultrasound dilution (COUD) with the “clinical standard” thermodilution (COTD) in adult cardiosurgical ICU patients.
This study was approved by the institutional review board at the National Research Centre for Surgery, Russian Academy of Medical Science, Moscow, and was carried out per the ethical standards set forth in the Helsinki Declaration of 1975. Written informed consent was obtained from all the patients studied. Patients with in situ PAC and radial artery catheter, placed for their routine medical care, were eligible for the study. Patients were excluded if they had history of heparin allergy. For induction of anesthesia, atracurium besylate 0.3–0.6 mg/kg and sevoflurane up to 8% were used. Base anesthesia was supported with propofol 2–12 mg/kg/h and fentanyl and sevoflurane 0.5%–3%. Twenty one patients were on dopamine or dobutamine infusion (2–12 μg/kg/min), and in the remaining five, norepinephrine (0.1–0.5 μg/kg/min) was also used. Fluid resuscitation was performed under the control of patient's hemodynamic profile.
The COstatus system (Transonic Systems Inc., Ithaca, NY) was used for COUD determination by UD method. Ultrasound velocity in blood (1,560–1,590 m/s) is primarily a function of total protein and ion concentrations. Injections of isotonic saline (1,533 m/s) cause a decrease in ultrasound velocity, which makes it useful as an indicator1,9–12 (Figure 1). To apply this technology in the ICU setting, the following steps were followed.
a. A disposable extracorporeal AV loop primed with heparinized saline (priming volume ∼5 ml) was connected between the radial artery catheter and the introducer of the PA catheter (Figure 2). The AV loop is made with microbore tubing to have minimal priming volume.
b. Two reusable ultrasound flow-dilution sensors that measure changes in blood ultrasound velocity and blood flow in the AV loop (Figure 1) were clamped onto the arterial and venous limbs of the loop (Figure 2). The venous sensor is used to measure the volume and time of injection and record blood properties. The arterial sensor is used to record the dilution curve after the indicator passes through the cardiopulmonary system. Using the information recorded by both the venous and arterial sensors, COUD is measured using Stewart-Hamilton principle.
c. For each measurement session, a roller pump was turned on to circulate blood from the artery to the vein through the AV loop at a rate of 8–12 ml/min for 5–7 minutes. Twenty-five milliliters of isotonic saline, prewarmed to body temperature in a saline bag warmer (HFW 1000, Transonic Systems Inc., Ithaca, NY), was injected into the venous side of the AV loop.
d. One measurement session included three COUD measurements performed randomly throughout the breathing cycle. One extra measurement was performed if a “Repeat” message was displayed on the monitor.
e. At the conclusion of a measurement session, the blood in the AV loop was returned to the patient, and the AV loop was filled with heparinized saline to be ready for the next measurement session.
Cardiac output determination by thermodilution method was performed by injecting 10 ml of ice-cold saline into the proximal injectate port of a thermistor-tipped, flow-directed PAC (Edwards Lifesciences, Irvine, CA). Per our standard clinical practice and also as suggested by other researchers,13 the largest and smallest measurements of the five injections performed for COTD were omitted. The remaining measurements were averaged for comparison with COUD measurements.
Cardiac output by the ultrasound dilution measurements were performed within 3–5 minutes after thermodilution measurements. Based on the recommendation from the manufacturer, three injections were made to get COUD measurements. Both COUD and COTD measurements were performed by the same operator. All COUD measurements were averaged for comparison except for those with a “Repeat” message. All patients were calm with stable hemodynamics and restful throughout the study, and for those who were receiving mechanical ventilation, no changes were made in the ventilator settings.
Sample size was determined using the paired t test method. For a mean CO of 5.0 L/min and the clinically acceptable percentage error to be <30%,14 the SD of differences was assumed to be 0.75 L/min. Based on this assumption, a sample of at least 22 patients was required to detect a mean difference of 0.5 L/min between the methods with a statistical power of 85% and at a significance level of 0.05.15,16
Correlation between COUD and COTD was determined by linear regression analysis using the least squares method. Bias was defined as the mean difference between the two methods. Limits of agreement were calculated as the bias ± 2SD between the methods for the 95% confidence limits. Comparison of the bias and limits of agreement between the two methods was analyzed using the Bland-Altman approach.17 The percentage error was calculated as 2SD of the bias (2SD) divided by the mean COTD.14 As suggested by Critchley and Critchley,14 a percentage error of <30% was defined as the acceptable limit for COUD to be considered equivalent and hence interchangeable with COTD.
In addition, the data was categorized into three groups based on thermodilution data: low CO, medium CO, and high CO. Percentage error14 was calculated for each of these groups and also for total measurements. Reproducibility between measurements by each method was assessed by calculating coefficient of variance (CV).
Twenty-six post cardiac surgery ICU patients (Table 1) with indications for PA catheterization (Table 2) were studied. A total of 77 COUD and COTD averaged measurement sets were compared. Cardiac output measured by thermodilution ranged from 3.28 to 9.4 L/min, whereas COUD ranged from 2.85 to 10.1 L/min. The correlation between the methods was found to be r = 0.91, COUD = 0.93(COTD) + 0.42 L/min (Figure 3). Bland-Altman analysis showed that the bias and precision (mean difference ± 2SDs) between the two methods was −0.004 ± 1.34 L/min (Figure 4). Percentage error (Table 3) for the total data was found to be 22.2%. Maximum variation in the percentage error from each of the three groups when compared with the total error was 3.3%. The CV for COUD measurements was found to be 5.84%, whereas the CV for COTD measurements was found to be 12% for five injection sets and 6% for three injection sets.
Ultrasound Dilution Technology for Hemodynamic Measurements
Ultrasound dilution technology9 was introduced in 1995 and uses isotonic saline as an indicator to change the ultrasound velocity of blood. Since then, this technology is widely used (more than 150 full papers) in hemodialysis for measurement of access recirculation, AV access flow, and CO. Previous studies showed that COUD correlated well (r = 0.97) with pulmonary artery (PA) thermodilution measurements during hemodialysis18 and with “gold standard” perivascular flow probe (r = 0.99) that measures blood flow directly by placing an ultrasound flow probe on the ascending aorta in a pig hemodialysis model.19
In our earlier study,1 an AV tubing loop was used without a roller pump. In that study, the flow between the radial artery and specially cannulated cubital vein was driven by the natural arterial venous pressure gradient. A special calibration was performed with the AV loop, and intravenous saline injections were performed into a separate central venous catheter. Although the comparison with thermodilution produced good correlation (r = 0.97), it was found impractical for routine use, primarily because the blood flow in the loop was periodically unstable during measurement procedure.
In this new approach using the COstatus system, a roller pump drives a small (8–12 ml/min) but stable flow between the standard preexisting peripheral artery and central venous catheters in the patient. So, in addition to the advantage of having stable flow in the AV loop, the need for cubital vein catheterization is eliminated.
To our knowledge, this study is the first CO validation that uses the new extracorporeal AV loop concept verses PA thermodilution, which is considered as the “clinical standard” for assessment of CO in ICU. Our results (Figures 2 and 3) showed that CO measured by UD closely agreed with the thermodilution measurements in the heterogeneous post cardiac surgery ICU patients we studied.
To evaluate the equivalence between the two methods, we used the Critchley and Critchley14 criteria, similar to other CO comparison studies.20–21,27 This criterion suggests that a new technique is clinically acceptable if the percentage error was within 30% when compared with the current reference methods. They defined the percentage error as 2SD of the bias (mean of the difference between the two methods) divided by the mean of the reference method (COTD in this study). Our data showed that the percentage error was 22.2% and hence COUD produced results are clinically acceptable14 in our patient population.
The data was categorized into three groups based on thermodilution measurements. The limits of these three categories were chosen arbitrarily. This categorization was performed to investigate the agreement between the two methods at different levels and to check whether there was any noticeable difference in bias at different levels. Our data as summarized in Table 3 showed clinically acceptable agreement in all the three groups, suggesting that the two methods agree well over the entire range of CO measured.
The Place of New UD CO Technology Among Others in a Clinical Field
The use of PA thermodilution, the current “clinical standard” for hemodynamic assessment, is currently diminishing because of its association with various complications.1,4–5 Less-invasive dilution technologies such as transpulmonary thermodilution and lithium dilution are now available. Transpulmonary thermodilution requires femoral artery catheterization,7 and hence its use is limited with other arterial sites such as the radial artery. Lithium dilution requires injection of nonphysiologic substance, involves blood loss, and causes problems in terms of lithium accumulation if the measurements were to be repeated.20,22 Moreover, use of this technology is limited in some countries, such as in United States and Japan to patients weighing 40 kg or above.
Less-invasive arterial pressure-based technologies have been recently developed to monitor CO. Some of these require calibration, whereas a few claim that they do not need calibration. Frequent recalibration of the pulse contour methods with intermittent indicator dilution methods such as thermodilution are recommended before attempting any major treatment changes.20,23 In case of methods that do not require calibration, the aortic compliance is assumed based on patient demographics, which may lead to errors because the compliance is patient and condition specific. Hence, these methods might be good to monitor dynamic changes but not for accurate absolute results under the difficult clinical conditions seen during intraoperative and postoperative care of cardiac surgery patients.24 Unreliability of these methods was also found in patients with congenital heart disease25 and in septic shock patients.26
The major advantage of COstatus system is in its extracorporeal nature with the ability to be connected to preexisting standard peripheral arterial, and central venous catheters. Once the disposable set is primed and connected to the catheters, it takes 5–6 minutes to perform 2–3 measurements.
Harmless indicator (isotonic saline) and the absence of blood loss allow the use of COstatus in patients of any age including small children. Indeed, there are reports of using this technology in rats and in piglets with neonatal setting (umbilical caterers)28–29 and in small children of 2.6 kg and more.30–31
Limitation of the COstatus system is that it can be used only with patients having both in situ arterial and central venous catheters. Although these catheters are standard and are available for routine medical care, the requirement to have both catheters in the patient limits its use to only such patients. Also, for the duration of COstatus use (5–6 minutes when the pump is operated), arterial pressures is not available because we closed the stopcock to the pressure line and opened it to the AV loop. Finally, the volume of injections may be an issue in patients who are extremely hypervolemic. Our data showed that the coefficient of variation (SD/mean) from the 271 COUD measurements was 6.0%. This means that in almost 90% of the cases, the results from two consecutive injections will be within 10%. Thus, in a majority of the sessions, it is possible to limit the number of injections to two if the two readings are within 10%. This gives an opportunity to inject less amount of saline per measurement session.
A limitation of our study protocol is that only post cardiac surgery ICU patients with stable conditions were recruited, and hence the study does not represent patients with changing clinical conditions.
We found that CO determined by the new COstatus system based on UD method, and the clinical standard PA thermodilution method are interchangeable in heterogeneous post cardiac surgery critically ill patients. Being less invasive and safe, COstatus has the potential to be used not only in adults but also in younger age patients. Further studies should be performed to assess the reliability of COstatus in patients with different clinical conditions such as shock, sepsis.
The authors thank Transonic Systems Inc., Ithaca, NY, for lending the equipment for this study.
1.Eremenko A, Balykov I, Chaus N, et al
: Use of an extracorporeal arteriovenous tubing loop to measure cardiac output in intensive care unit patients by ultrasound velocity dilution. ASAIO J
44: M462–M464, 1998.
2.Booth JH, Quinn T, Richmond ME, et al
: Cardiac output measurement by arterial pressure waveform analysis during optimization of biventricular pacing after cardiac surgery. ASAIO J
55: 587–591, 2009.
3.Swan HJ, Ganz W, Forrester J, et al
: Catheterization of the heart in man with use of a flow-directed ballon-tipped catheter. N Engl J Med
283: 447–451, 1970.
4.Abreu AR, Campos MA, Krieger BP: Pulmonary artery rupture induced by pulmonary artery catheter: A case report and review of the literature. J Intensive Care Med
19: 291–296, 2004.
5.Sandham JD, Hull RD, Brant RF, et al
: A randomized, controlled trial of the use of pulmonary-artery catheters in high-risk surgical patients. N Engl J Med
348: 5–14, 2003.
6.Linton RAF, Band DM, Haire KM: A new method of measuring cardiac output in man using lithium dilution. Br J Anaesth
71: 262–266, 1993.
7.Orme RM, Pigott DW, Mihm FG: Measurement of cardiac output by transpulmonary arterial thermodilution using a long radial artery catheter. A comparison with intermittent pulmonary artery thermodilution. Anaesthesia
59: 590–594, 2004.
8.Krivitski NM, Kislukhin VV, Thuramalla NV: Theory and in-vitro validation of a new extracorporeal approach for hemodynamic assessment in pediatric and neonatal ICU patients. Pediatr Crit Care Med J
9: 423–428, 2008.
9.Krivitski NM: Novel method to measure access flow during hemodialysis by ultrasound dilution technique. ASAIO J
41: M741–M745, 1995.
10.Krivitski NM, Starostin D, Smith TL: Extracorporeal recording of mouse hemodynamic parameters by ultrasound velocity dilution. ASAIO J
45: 32–36, 1999.
11.Moser M, Kenner T: Blood flow and blood volume determinations in aorta and in coronary circulation by density dilution. Basic Res Cardiol
83: 577–589, 1988.
12.Moser M, Kenner T: Distribution spaces of intravenously injected saline, plasma and red cell concentrate. Pflugers Arch
394S: R48, 1982.
13.Nilsson LB, Nilsson JC, Skovgaard LT, Berthelsen PG: Thermodilution cardiac output: Are three injections enough? Acta Anaesthesiol Scand
48: 1322–1327, 2004.
14.Critchley LA, Critchley JA: A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. J Clin Monit Comput
15: 85–91, 1999.
15.Dupont WD, Plummer WD: Power and sample size calculations: A review and computer program. Control Clin Trials
11: 116–128, 1990.
16.Friedman LM, Furberg CD, DeMets DL: Fundamentals of Clinical Trials
, 3rd ed. New York, NY, Springer, 1999, p. 111.
17.Bland JM, Altman DG: Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
1: 307–310, 1986.
18.Nikiforov YV, Kisluchine VV, Chaus NI: Validation of a new method to measure cardiac output during extracorporeal detoxification. ASAIO J
42: M903–M905, 1996.
19.Kisloukhine VV, Dean DA: Validation of a novel ultrasound dilution method to measure cardiac output during hemodialysis. ASAIO J
42: M906–M907, 1996.
20.Cooper ES, Muir WW: Continuous cardiac output monitoring via arterial pressure waveform analysis following severe hemorrhagic shock in dogs. Crit Care Med
35: 1724–1729, 2007.
21.Sander M, Spies CD, Grubitzsch H, et al
: Comparison of uncalibrated arterial waveform analysis in cardiac surgery patients with thermodilution cardiac output measurements. Crit Care
10: R164, 2006.
22.Freeman MP, Freeman SA: Lithium: Clinical considerations in internal medicine. Am J Med
119: 478–481, 2006.
23.Østergaard M, Nielsen J, Rasmussen JP, Berthelsen PG: Cardiac output—Pulse contour analysis vs. pulmonary artery thermodilution. Acta Anaesthesiol Scand
50: 1044–1049, 2006.
24.Rauch H, Muller M, Fleischer F, et al
: Pulse contour analysis versus thermodilution in cardiac surgery patients. Acta Anaesthesiol Scand
46: 424–429, 2002.
25.Mahajan A, Shabanie A, Turner J, et al
: Pulse contour analysis for cardiac output monitoring in cardiac surgery for congenital heart disease. Anesth Analg
97: 1283–1288, 2003.
26.Berthelse PG: Pulse-contour-derived cardiac output is unreliable in septic shock. Acta Anaesthesiol Scand
27.Engoren M, Barbee D: Comparison of cardiac output determined by bioimpedance, thermodilution, and the Fick method. Am J Crit Care
14: 40–45, 2005.
28.de Boode WP, van Heijst AF, Hopman JC, et al
: Cardiac output measurement using an ultrasound dilution method: A validation study in ventilated piglets. Pediatr Crit Care Med
11: 103–108, 2010.
29.Gleed R, Smith T, Callahan M, et al
: Validation of novel ultrasound dilution cardiac output method for pediatric and neonatal patients. Intensive Care Med
32: S172, 2006.
30.Schulenberg A, Harmon W, Rubenstein J, et al
: A novel method to measure cardiac output in the pediatric ICU: Animal validation and preliminary clinical study. Crit Care Med
34: A12, 2007.
Copyright © 2010 by the American Society for Artificial Internal Organs
31.Crittendon I, Dreyer W, Kim J: Validation of ultrasound dilution cardiac output measurements in pediatric patients. Crit Care Med
37: A90, 2009.