Anesth Analg 2008; 106:171-174
© 2008 International Anesthesia Research Society
doi: 10.1213/01.ane.0000297440.52059.2c
TECHNOLOGY, COMPUTING, AND SIMULATION
Noninvasive Cardiac Output Determination Using Applanation Tonometry-Derived Radial Artery Pulse Contour Analysis in Critically Ill Patients
Friederike Compton, MD,
Marc Wittrock,
Juergen-Heiner Schaefer, MD,
Walter Zidek, MD,
Martin Tepel, MD, and
Alexandra Scholze, MD
From the Division of Nephrology, Department of Nephrology and Endocrinology, Charité University Medicine Berlin, Campus Benjamin Franklin, Germany.
Address correspondence to Dr. Friederike Compton, Medizinische Klinik IV, Nephrologie, Hindenburgdamm 30, D-12200 Berlin, Germany. Address e-mail to friederike.compton{at}charite.de.
 |
Abstract
|
|---|
Conventional thermodilution cardiac output (CO) monitoring is limited mainly to intensive care units and operating rooms because it requires the use of invasive techniques. To reduce the potential for complications and to broaden the applicability of hemodynamic monitoring, noninvasive methods for CO determination are being sought. Applanation tonometry allows noninvasive CO estimation through pulse contour analysis, but the method has not been evaluated in critically ill patients. We therefore performed noninvasive radial artery applanation tonometry in 49 critically ill medical intensive care unit patients and compared CO estimates to invasive CO measurements obtained using a pulmonary artery catheter or the PiCCO® transpulmonary thermodilution system. One-hundred-sixteen measurements were performed, and patients were receiving vasopressor support during 78 measurements. When the data were analyzed with bias and precision statistics, a large bias of 2.03 L · min–1 · m–2 and a high percentage error of 85% were found between the invasive measurements and applanation tonometry-derived CO estimates, with the noninvasive CO results being significantly lower than the invasive ones (P < 0.001). There was no significant difference in bias between the patients who were receiving vasopressor support and those who were not (P = 0.874) or between patients with good and poor applanation tonometry pressure waveform signal quality (P = 0.071). Whereas a significant increase in the invasively determined CO was observed when a fluid bolus was administered (n = 7, P = 0.016), these changes were not reflected by the noninvasive method. We conclude that radial artery applanation tonometry is not suitable to determine CO in critically ill hemodynamically unstable patients.
|
Introduction
|
|---|
Monitoring and optimization of cardiac output (CO) as well as other hemodynamic variables are an integral part of the management of critically ill hemodynamically unstable patients. Thermodilution measurement of CO using a pulmonary artery catheter (PAC) remains a standard for hemodynamic monitoring in the intensive care unit (ICU), but is associated with a number of potentially severe complications.1,2 Even though transpulmonary thermodilution is comparably less invasive and the method is increasingly being used in the ICU, it still requires placement of a central venous and an arterial catheter.3,4 Either technique is therefore difficult to apply in the initial evaluation of the hemodynamically unstable patient, which often takes place in the emergency department or elsewhere outside the ICU.
Noninvasive estimation of CO is possible through analysis of the radial artery pulse contour obtained with applanation tonometry. Until now, the procedure has mainly been used in the diagnostic evaluation of patients with cardiovascular risk factors and hypertension.5–7 The CO estimation algorithm has been validated against invasive measurements in healthy subjects, but there are no data evaluating applanation tonometry in critically ill hemodynamically unstable patients.8
We therefore studied 49 ICU patients and compared CO determinations obtained with the HDI/Pulse Wave CR-2000 Cardiovascular Profiling Instrument®, which uses applanation tonometry to derive the radial artery pulse pressure waveform, with invasive CO measurements obtained either through pulmonary artery thermodilution or transpulmonary thermodilution.
|
METHODS
|
|---|
Patients
Forty-nine patients consecutively treated in the medical ICU of the Charité Campus Benjamin Franklin in Berlin, Germany, between February 2005 and July 2006 were enrolled in the study. In all patients, invasive thermodilution CO monitoring was performed as part of the clinical treatment and was compared with the noninvasive CO determinations made with the HDI/Pulse Wave CR-2000 Cardiovascular Profiling Instrument (Hypertension Diagnostics, Inc., Eagan, MN).
The study was approved by the local ethics committee, which waived the need for informed consent.
Invasive CO Measurement
The decision to use invasive hemodynamic monitoring was at the discretion of the treating physician. Indications included differential diagnosis of shock, severe sepsis with multiple organ failure, and assessment of intravascular volume status.
Thermodilution CO measurements were either performed using a PAC (7.5 F 831F75 Edwards Lifesciences, Unterschleißheim, Germany) or the PiCCO® (Pulsiocath® PV2015L20, Pulsion medical systems, Munich, Germany) transpulmonary thermodilution catheter inserted into the femoral artery.
Noninvasive CO Determination with the HDI/Pulse Wave CR-2000 Cardiovascular Profiling Instrument
With the HDI/Pulse Wave CR-2000 Cardiovascular Profiling Instrument (Hypertension Diagnostics Inc.), radial artery pressure waveforms are obtained noninvasively using a proprietary tonometer applied to the skin of the distal forearm overlying the radial artery. The forearm is kept in a constant position using a "wrist stabilizer," and the tonometer is housed in a holding and positioning device, which is wrapped around the distal forearm. Pressure waveform signal quality is displayed on a scale from 1 (=poor signal) to 25 (=excellent signal). The arterial pressure waveforms are calibrated with a blood pressure cuff and calibration system integral to the device. Cardiac ejection time is derived from the pulse wave and used to obtain CO applying a multivariate computer algorithm considering patient age, heart rate, and body surface area. Details of the procedure and equations have been described elsewhere.8
Statistical Analysis
CO results were indexed to total body surface area and are referred as cardiac index. The data were analyzed using GraphPad prism (version 3.0, GraphPad software, San Diego, CA) and SPSS for windows (version 11.5, SPSS Inc., Chicago, IL). Results were expressed as means with standard deviations. Nonparametric testing was used to compare patient groups and analyze changes within groups, and a two-sided P value <0.05 was considered statistically significant. To compare invasively and noninvasively obtained CO results, bias and precision statistics were performed and data were displayed as Bland–Altman plots, where the mean difference between paired readings is referred to as the bias, and the 95% confidence limits calculated from the individual standard deviations as the limits of agreement.9 The percentage error was then calculated as suggested by Critchley and Critchley.10
|
RESULTS
|
|---|
Demographic data of the 49 patients are shown in Table 1. One-hundred-sixteen noninvasive CO determinations were attempted. In 14 instances, no radial pulse could be derived with applanation tonometry; in 13 of these cases, the patients were receiving vasopressors. Thus, 102 noninvasive CO estimates were available for analysis.
Patients were receiving vasopressor support during 78 (77%) of the measurements with a mean norepinephrine dose (n = 72) of 0.30 ± 0.27 µg · kg–1 · min–1 and epinephrine (n = 15) 0.22 ± 0.36 µg · kg–1 · min–1, respectively.
Invasive thermodilution CO measurements were obtained with a PAC in 36 instances and with a PiCCO system in 79 cases. In 13 measurements, both invasive methods were performed, and in these cases, the PiCCO CO results were used for comparison with applanation tonometry.
Figure 1 shows the agreement between the invasive and the noninvasive methods. For all invasive measurements taken together (n = 102, panel C), the bias was 2.03 L · min–1 · m–2 with limits of agreement of ±2.67 L · min–1 · m–2. The overall percentage error was ±85 percent. Bias and agreement was similar for each of the two invasive methods (PAC 2.17 ± 3.89, PiCCO 2.13 ± 2.58), as displayed in panels A and B. Means and standard deviations of the methods compared are displayed in Table 2. The CO estimates obtained noninvasively were significantly lower than the invasive CO measurements (n = 102, P < 0.001).
View larger version (11K):
[in this window]
[in a new window]
|
Figure 1. Agreement between invasive and noninvasive cardiac output (CO) measurements (panel A: pulmonary artery catheter measurements, panel B: PiCCO® measurements, panel C: all invasive measurements taken together).
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Cardiac Output Results (Mean ± Standard Deviation) Obtained with the Invasive Versus the Noninvasive Methods
|
|
When the patients who were receiving vasopressor therapy at the time of measurement were compared with those who were not, no significant difference in bias was found between the subgroups (mean bias in patients receiving vasopressors 1.99 L · min–1 · m–2, mean bias in patients without vasopressors 2.04 L · min–1 · m–2, P = 0.874).
Subgroup analysis with respect to applanation tonometer-derived blood pressure waveform signal quality did not yield significant differences in bias between patients with good signal quality (
10, n = 40) and poor signal quality (<10, n = 62). Mean signal quality of all measurements was 10 ± 5.8, mean bias in patients with good signal quality was 2.35 L · min–1 · m–2, mean bias with poor signal quality was 1.73 L · min–1 · m–2 (P = 0.071).
In seven cases, CO was determined before and after a fluid bolus of 250 mL of 4% gelatin polysuccinate (Gelafundin®, B. Braun Melsungen AG, Melsungen, Germany). Although there was a significant increase in the CO determined by thermodilution techniques after a fluid bolus (P = 0.016), no difference was observed when CO was determined noninvasively (Fig. 2).
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Invasive (panel A) and noninvasive (panel B) cardiac output (CO) determinations before and after a fluid bolus.
|
|
|
DISCUSSION
|
|---|
Noninvasive CO determination can be attempted using several different methods.4,11 Transesophageal echocardiography allows estimation of the CO, but operator-dependence and the invasive nature of the procedure have limited its use mainly to the operating room. Impedance cardiography provides beat-to-beat operator-independent CO determinations, but the thoracic impedance method has especially been shown to be less accurate than invasive and noninvasive pulse contour analysis CO determinations in critically ill patients.12
Noninvasive pulse contour analysis can be performed with a system using the finger blood pressure waveform to obtain the arterial pulse contour noninvasively. The method has been evaluated in critically ill patients, but was judged to be inaccurate compared with PAC CO.12,13 Difficulties with the finger blood pressure registration in critically ill patients as well as limitations of the underlying computer algorithm were hypothesized to be responsible for the results. With applanation tonometry, the radial artery is used to derive the blood pressure waveform, and the method is also based on a different methodology.
Our results show that noninvasive radial artery applanation tonometry is technically possible in critically ill patients, but that the method is not suitable to replace invasive hemodynamic monitoring. The large bias, poor agreement, and high percentage error preclude the use of noninvasive applanation tonometry to estimate CO in critically ill hemodynamically unstable patients.10 In addition, when used to track individual CO changes with fluid administration in order to indicate the hemodynamic "trend," applanation tonometry-derived CO readings did not reflect the CO increases measured invasively.
Subgroup analyses could neither identify vasopressor-induced vasoconstriction nor poor signal quality as a reason for the poor performance of the method, but the noninvasive approach of applanation tonometry could still preclude accurate pressure readings from the radial artery in hemodynamically unstable patients. Our data do not allow a more detailed analysis of the factors responsible for the results obtained. More specific studies involving more homogeneous patient groups will be necessary to address the differences we found between the invasive methods and noninvasive applanation tonometry.
|
Footnotes
|
|---|
Accepted for publication September 11, 2007.
Reprints will not be available from the author.
|
REFERENCES
|
|---|
- Connors AF Jr, Seroff T, Dawson NV, Thomas C, Harrell FE Jr, Wagner D, Desbiens N, Goldman L, Wu AW, Califf RM, Fulkerson WJ Jr, Vidallet H, Broste S, Bellamy P, Lynn J, Knaus WA. The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. JAMA 1996;276:889–97[Abstract/Free Full Text]
- Hadian M, Pinsky MR. Evidence-based review of the use of the pulmonary artery catheter: impact data and complications. Crit Care 2006;10(suppl 3):S8[Medline]
- Goedje O, Hoeke K, Lichtwarck-Aschoff M, Faltchauser A, Lamm P, Reichart B. Continuous cardiac output by femoral arterial thermodilution calibrated pulse contour analysis: comparison with pulmonary arterial thermodilution. Crit Care Med 1999;27:2407–12[Web of Science][Medline]
- Parmley CL, Pousman RM. Noninvasive cardiac output monitoring. Curr Opin Anaesthesiol 2002;15:675–80[Medline]
- Cohn JN, Finkelstein S, McVeigh G, Morgan D, LeMay L, Robinson J, Mock J. Noninvasive pulse wave analysis for the early detection of vascular disease. Hypertension 1995;26: 503–8[Abstract/Free Full Text]
- McVeigh GE, Bratteli CW, Morgan DJ, Alinder CM, Glasser SP, Finkelstein SM, Cohn JN. Age-related abnormalities in arterial compliance identified by pressure pulse contour analysis: aging and arterial compliance. Hypertension 1999;33:1392–8[Abstract/Free Full Text]
- Kneifel M, Scholze A, Burkert A, Offermann G, Rothermund L, Zidek W, Tepel M. Impaired renal allograft function is associated with increased arterial stiffness in renal transplant recipients. Am J Transplant 2006;6:1624–30[Web of Science][Medline]
- Rietzschel ER, Boeykens E, De Buyzere ML, Duprez DA, Clement DL. A comparison between systolic and diastolic pulse contour analysis in the evaluation of arterial stiffness. Hypertension 2001;37:e15–e22[Abstract/Free Full Text]
- Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10[Web of Science][Medline]
- Critchley LAH, Critchley JAJH. A meta-analysis of studies using bias and precision statistics to compare cardiac output meaurement techniques. J Clin Monit Comput 1999;15:85–91[Medline]
- Kööbi T. Non-invasive cardiac output determination: state of the art. Curr Opin Anaesthesiol 1999;12:9–13[Medline]
- Hirschl MM, Kittler H, Woisetschläger C, Siostrzonek P, Staudinger T, Kofler J, Oschatz E, Bur A, Gwechenberger M, Laggner AN. Simultaneous comparison of thoracic bioimpedance and arterial pulse waveform-derived cardiac output with thermodilution measurement. Crit Care Med 2000;28: 1798–802[Web of Science][Medline]
- Hirschl MM, Binder M, Gwechenberger M, Herkner H, Bur A, Kittler H, Laggner AN. Noninvasive assessment of cardiac output in critically ill patients by analysis of the finger blood pressure waveform. Crit Care Med 1997;25:1909–14[Web of Science][Medline]
This article has been cited by other articles:
|
|
|
|
 
F. Compton and J.-H. Schafer
Noninvasive Cardiac Output Determination: Broadening the Applicability of Hemodynamic Monitoring
Seminars in Cardiothoracic and Vascular Anesthesia,
March 1, 2009;
13(1):
44 - 55.
[Abstract]
[PDF]
|
|
|
Top
Abstract
Introduction
