Anesth Analg 2003;97:1283-1288
© 2003 International Anesthesia Research Society
PEDIATRIC ANESTHESIA
Pulse Contour Analysis for Cardiac Output Monitoring in Cardiac Surgery for Congenital Heart Disease
Aman Mahajan, MD,
Afshin Shabanie, MD,
Judi Turner, MD PhD,
Michael J. Sopher, MD, and
Jure Marijic, MD
Department of Anesthesiology, David Geffen School of Medicine, University of California Los Angeles
Address correspondence and reprint requests to Aman Mahajan, MD, Department of Anesthesiology, Box 951778, David Geffen School of Med at UCLA, Los Angeles, CA 90095. Address e-mail to amahajan{at}mednet.ucla.edu
 |
Abstract
|
|---|
Conventional methods of cardiac output monitoring using pulmonary artery catheters may not be feasible in patients with congenital heart disease because of patients’ small size or aberrant anatomy. We studied the accuracy of a new device, which uses pulse contour analysis to measure continuous cardiac output, in children and adults undergoing congenital heart surgery. Sixteen patients, median ages 7 yr old, were included in this prospective study. One-hundred-ninety-one data points were obtained in the pre- and postcardiopulmonary bypass periods and in the first 12 h after intensive care unit admission. We evaluated the relationship between cardiac index (CI) derived from transpulmonary thermodilution (TDCI) and CI derived from pulse contour analysis (PCCI). Bias and limits of agreement between TDCI and PCCI over all time periods were 0.1 ± 1.94, indicating a wide dispersion of the data. Coefficient of correlation (r) between the TDCI and PCCI was 0.7. Although in previous studies, PCCI has been suggested to be accurate in adult cardiac surgery, we found it to be less reliable in our study patients, even after shunt correction. The relationships of the volume and pressure based measures of preload, intrathoracic blood volume index (ITBI), and central venous pressure with CI were also investigated. After repair, correlation (r) between PCCI or TDCI and ITBI (0.56 and 0.71, respectively) was better than that between PCCI or TDCI and CVP (0.16 and 0.11, respectively), indicating greater validity of ITBI as a measure of preload.
IMPLICATIONS: Our results suggest that the pulse contour analysis cardiac output (CO) monitoring in patients undergoing congenital heart surgery may not provide as accurate or reliable measures of CO as previously suggested. The volume-based variable of preload intrathoracic blood volume index (ITBI) has better correlation with cardiac index (CI) than the central venous pressure, suggesting that ITBI may be a better indicator of preload.
|
Introduction
|
|---|
Cardiac output (CO) monitoring is one of the principal means of assessing the hemodynamic profile and tissue perfusion in the perioperative period. Although management of blood pressure is essential to optimize tissue perfusion, changes in blood pressure are not always reflective of the alterations in the CO. Thermodilution through a pulmonary artery catheter (PAC) has been used extensively in adult cardiac surgery to measure CO, and it has become a standard clinical monitor to which other methods of measuring CO may be compared. However, PACs are not devoid of risks, and their use is impractical, if not impossible, in small children and many patients with congenital cardiac lesions. In addition, thermodilution may not always be reliable in patients with complex congenital lesions involving intracardiac shunting because of indicator recirculation (1). Therefore, anesthesiologists have traditionally relied on other variables such as blood pressures, central venous pressures (CVP), mixed venous oxygen saturation, and arterial line waveform appearances as surrogate indicators of output during and immediately after congenital heart surgery.
A number of newer alternative technologies have been developed to allow CO monitoring without the use of a PAC. One such technique is transpulmonary thermodilution (TDCI) (2), made possible with the advent of more accurate thermistors, enabling us to measure CO measurements by injecting cold saline into a central venous line and monitoring the temperature in a central artery such as the femoral. The validity and accuracy of this methodology have been established in different patient populations and settings (2–4). Another new technology that uses algorithms that allow assessment of stroke volume and therefore CO by analysis of the contour of the arterial line waveform (pulse contour [PC] analysis CO measurement) has recently been introduced (5–7). This technology uses the principle that area under the curve of a central arterial waveform correlates with the stroke volume. The conversion factor of this correlation can be calculated after obtaining the CO value through the use of a different technology (i.e., calibrating the device). We used the PiCCO system (Pulsion Medical Systems, Munich, Germany), which incorporates thermodilution as well as the PC analysis technologies, to measure CO. This device uses a TDCI for initial and periodic calibration and then provides continuous measurements of the cardiac index (CI) using the PC analysis algorithm. The validity of this technique has been demonstrated in studies performed on adult cardiac surgery patients, suggesting that it has good correlations with PACs and other methods of CO monitoring (7–10). However, the validity of pulse contour analysis in the perioperative management of children and adults with congenital cardiac lesions has not been determined. This large subgroup of patients would benefit the most from such a device because there are no good alternatives available for measuring CO in these patients. This device also provides, through thermodilution, a number of volume-based measures of preload, including intrathoracic blood volume index (ITBI), stroke volume variation, and global end-diastolic volume of the heart, which may be useful in the management of these patients.
This study aimed to evaluate the validity of this technology in the intraoperative and postoperative management of patients (primarily the pediatric age group) undergoing corrective cardiac surgery for congenital heart lesions. We sought the relationship of CI derived from TDCI versus CI derived from PC analysis in children and adults with congenital heart disease at different time periods in the operating room and in the intensive care unit (ICU). In addition, we evaluated the correlation between ITBI, CVP, and CI.
|
Methods
|
|---|
After approval from the IRB and after informed consent or assent by the parents or patients, 16 patients undergoing corrective surgery for congenital heart lesions were enrolled in a prospective study. To be included in the study, patients were required to weigh at least 10 kg and have no contraindications to placement of femoral arterial lines. The spectrum of lesions included atrial septal defects, ventricular septal defects, tetralogy of Fallot, Fontan completion, subaortic membrane, pulmonary conduit-valve obstruction, and Ebstein anomaly. Eleven of the patients had lesions involving intracardiac shunting in the preoperative period. The corrective surgery included closure of all intra- or extracardiac arteriovenous shunts. The age range was 1–36 yr old, with a median age of 7 yr and a mean age of 10 yr. Two of the 16 patients were more than 18 yr old.
No aspect of surgical or anesthetic care was altered because of enrollment. A 4F femoral arterial catheter with thermistor (PiCCO systems) was inserted and automatic data acquisition begun after the initial calibration of the device using thermodilution measurements with 5–10 mL of cold saline depending upon patient size. Thermodilution injections were made at various points during the operation (baseline, after skin incision, after sternotomy, before and 15 min after bypass, and before transport to ICU) and at additional random points during surgery. Data points including TDCI, PC analysis, and ITBI were also obtained upon ICU admission and at 1, 3, 6, and 12 h thereafter. A total of 191 data points were obtained, including 61 points in the prebypass period (46 in shunting patients, and 15 in nonshunting patients), 69 points after bypass (all shunts corrected at this stage), and 61 data points in the ICU. Therefore, 145 of 191 data points were obtained with no intracardiac shunting involved (Table 1; Groups 2 and 3). The system operates in such a way that every time a thermodilution injection is performed, PC analysis readings automatically and immediately self-calibrate with the new value of TDCI. Because of this calibration process, we used the PC analysis value obtained immediately before such calibration for comparison with TDCI (termed PC analysis lag value).
View this table:
[in this window]
[in a new window]
|
Table 1. Accuracy of Pulse Controls analysis (PCCI) Compared with Transpulmonary Thermodilution (TDCI) During Different Time Periods of Congenital Cardiac Surgery
|
|
Statistical analysis of accuracy of PC analysis in comparison to TDCI was performed using the method described by Bland and Altman (10). Bias and limits of agreement (bias ± 2 SD) between TDCI and PC analysis were measured. We also determined the Pearson coefficient of correlation between the different variables (TDCI versus PC analysis, CVP versus CI, and ITBI versus CI).
|
Results
|
|---|
The mean PC analysis was 3.64 L · min-1 · m2 (range, 0.96–9.47 L · min-1 · m2) as compared with mean TDCI of 3.70 L · min-1 · m2 (range, 1.40–9.70) L · min-1 · m2. Bias, limits of agreement (bias ± 2 SD), and coefficient of correlation between TDCI and PC analysis for different patients and time periods are shown in Table 1 and Figure 1.
View larger version (32K):
[in this window]
[in a new window]
|
Figure 1. Top panels) Pre-cardiopulmonary bypass (CPB) period. (Middle panels) Post-CPB and intensive care unit (ICU). (Lower panels) All periods linear correlation (left) and mean difference between pulse contour analysis (PCCI) and transpulmonary thermodilution (TDCI) using Bland-Altman analysis (right). Each point represents the mean of three consecutive measures of TDCI and the respective PCCI_lag values. The right panels represent the bias between the two different techniques of cardiac output monitoring.
|
|
Coefficient of correlation was also determined between CVP, ITBI, and PC analysis in an attempt to (a) determine the existence of a correlation between the volume-based (ITBI) and the pressure based (CVP) measures of preload and (b) determine which one of the two correlates better with CO and hence is a better indicator of preload. The results are summarized in Table 2.
View this table:
[in this window]
[in a new window]
|
Table 2. Measure of Preload: Correlation of Pressure (Central Venous Pressure CVP) and Volume Based Intrathoacic Blood Volume Index (ITBI) Indicators of Preload with Cardiac Index (Pulse Control Analysis PCCI or Transpulmonary Thermodilution TDCI)
|
|
As seen in Table 2, no correlation is detectable between ITBI and CVP. Because intracardiac shunting might interfere with accurate measurement of ITBI before correction, the most reliable data for the comparison between ITBI and CVP as a measure of preload should be sought in Group 4 above (all patients and after shunt correction). In this group of patients, correlation coefficients between PC analysis and ITBI and PC analysis and CVP were 0.56 and 0.16, respectively. Correlation coefficients between TDCI and ITBI and TDCI and CVP were 0.71 and 0.11, respectively.
|
Discussion
|
|---|
Use of PC analysis to determine CO has generated interest in recent years. This technology has been evaluated in a few different clinical settings, such as surgery for coronary artery bypass grafting and noncardiac surgery as well as in the ICU setting (9,11–14). PC analysis CO monitoring is less invasive than PACs and also provides volume-based measures of preload that could be helpful in determining volume responsiveness of the patient. The use of this technology during congenital heart surgery can be even more beneficial because PACs are usually not an option, and yet, CO monitoring can be helpful for estimating cardiac function and tissue perfusion.
Our study was designed to look at the accuracy of PC-derived CO using the PiCCO systems in this group of patients. One way to describe the accuracy of two measurements in comparison to one another is to determine bias and the limits of agreement for a given confidence range. For a confidence of 95%, agreement is calculated as bias ± 2 SD. For all patients and all periods in our study, the mean bias was 0.1 L · min-1 · m2 (indicating the absence of any significant systematic error between the two measures), and the SD was 0.97, resulting in limits of agreement (calculated as bias ± 2 SD) of -1.84 to 2.04 L · min-1 · m2 between PC analysis and TDCI. The results when analyzed by different time periods were similar (Table 1). These results suggest that for any given value of PC analysis, there is a 95% chance that the TDCI would be within the above range as determined by PC analysis. For example, for a given PC analysis of 2.0, the TDCI might vary between 0.16 to 4.04 L · min-1 · m2. We believe that a range as wide as that obtained in our series can limit the clinical utility of the monitor because most anesthesiologists will treat a CI of 1.0 differently from a CI of 3.0. To report our data in another way, only 65% of PC analyses occur within a ±20% range of TDCI values after corrective surgery.
Studies of PC CO performed by others have described a wide spectrum of measured bias and agreement ranges between PC CO and femoral artery thermodilution. Values ranged from bias ± 2 SD = -0.12 ± 1.1 L/min reported by Gödje et al. (9,12) to 0.31 ± 2.50 reported by Zöllner et al. (13) and -0.14 ± 2.32 by Rauch et al (14). To compare our result of 0.1 ± 1.94 L · min-1 · m2 with those studies’ results, one must be aware that there is a clear difference in the patient population. Furthermore, our results are described as CI rather than CO, which increases our ranges for pediatric patients with a body surface area (BSA) of <1.0 m2 and decreases them for patients with a BSA >1.0 m2. We used CI because use of CO values without indexing them to BSA introduces the potential for tremendous statistical error when comparing values from small children to larger patients and tends to shift the results of all analyses towards the values obtained from the larger patients. Regarding the obtained correlation coefficients, the range was 0.88 by Zöllner et al. (13) to 0.95 by Buhre (11). Our result was r = 0.72 for all patients at all periods and 0.73 for all patients after the correction of intracardiac shunting.
Several factors, physiological or statistical-methodological, might be responsible for the differences between our results and the ones mentioned above. As explained earlier, it is erroneous to compare the TD CO with the simultaneous value of PC-derived CO because the latter value has just been calibrated to the thermodilution. We therefore used the PC analysis value immediately before the calibration (the lag value) for comparison to TDCI. Gödje et al. (9,12) recognized and explained the same concept, but these authors used an average of the values obtained just before and immediately after calibration for comparison. This methodology used by Gödje et al. (9,12) will make the correlation seem closer than they actually are. Buhre (11) also did not use the lag value for such comparisons. The above argument does not apply to studies comparing the PC and bolus PAC thermodilution CO because there is no PC auto-calibration with PAC bolus injections. Zöllner et al. (13) and Rauch et al. (14) compared PAC rather than transpulmonary bolus thermodilution for comparison with PC. Interestingly, the agreement ranges obtained by those two groups are considerably larger than those obtained by Gödje et al. or Buhre using TDCI. Furthermore, the complicated physiology of congenital lesions, especially intracardiac shunting and tricuspid valve disease, in our patients could be responsible for some of the discrepancy. However, the weak agreement between PC analysis and TDCI persisted even after the congenital lesions were corrected. Hemodynamic instability and drastic changes in the systemic vascular resistance can also render the PC analysis-derived CO inaccurate.
In a further attempt to explain the differences between our data and some of the previously published works, we repeated the analysis simply using the simultaneous values for PC analysis. This type of analysis produces results similar to those published previously, highlighting the importance of methodological factors. The correlation coefficient would be 0.90 and mean bias ± 2 SD would be 0.06 ± 1.26 using this type of analysis for all patients after surgical correction. However, we believe that the correct method for a true validation is using the lag value for PC analysis and not the simultaneous value because the latter has just been auto-calibrated to the TDCI.
A number of volume-based measures of preload can be derived from TDCI. More attention is being focused on stroke volume variation and ITBI in recent literature (15–19). The calculation of ITBI is based on the measurement of the transit time of the TDCI injectate and the slope characteristics of the thermodilution curve. Arguments have been made that the use of pressure-based measures such as CVP or left atrial pressure to evaluate preload assumes a normal or fixed compliance of the respective cardiac chambers and therefore could be misleading as a guide for fluid therapy in situations where such an assumption does not hold true.
In our series, ITBI had a consistently better correlation with PC analysis than did CVP (Table 2). This finding was consistent in each time period separately or in all periods combined and whether PC analysis or TDCI was used as a measure of CI. As stated earlier, intracardiac shunting can render thermodilution and all measures derived from it unreliable. We therefore believe that the most valid data to use for such a comparison are the data obtained after the correction of the shunts. In this time period, the correlation coefficient between PC analysis and ITBI were 0.56 as compared with those between PC analysis and CVP, which was 0.16. Correlation coefficient between TDCI and ITBI was 0.71 versus TDCI and CVP, which was 0.11.
Considering that in most clinical situations preload is one of the important factors influencing CO, it can be argued that a better correlation with CI indicates that ITBI may be a better and a more valid marker of preload. Similar results have been demonstrated in adults in different settings (16–20). In 57 patients with sepsis or septic shock, Sakka et al. (16) reported that ITBI was a more reliable indicator of cardiac preload than the filling pressures. In their study, correlation between CVP and CI or stroke volume index (SVI) was only 0.01 as compared with ITBI versus CI or SVI, which was 0.66. Moreover, a change in ITBI also correlated better with change in CI or SVI than did a change in CVP (r = 0.67 versus 0.05, respectively). In another prospective study of 40 patients undergoing heart transplantation, CVP and pulmonary artery occlusion pressures had a weaker relationship to SVI in these denervated hearts than did ITBI and global end-diastolic volume of the heart (20). Our data confirm that this relationship also applies to patients with congenital heart disease. However, one can also find studies contradicting this point. Mundigler et al. (21) suggest that in patients with poor left ventricular function, pressure based measures of preload more accurately reflect the changes in intravascular volume (demonstrated by response to fluid loading).
In summary, we demonstrate a large degree of variability between CI measured by PC analysis and CI measured by thermodilution in our group of patients with congenital heart lesions. This can limit the clinical usefulness of the PC measurements because of inadequate accuracy and precision. However, this monitor does provide the anesthesiologist the ability to measure CO using TDCI. This may be particularly helpful in these patients, because other methods of CO monitoring are not practical. In addition, this monitor allows measurement of ITBI, a volume-based index, which is likely superior to CVP as a marker of preload in patients undergoing cardiac surgery for correction of congenital lesions.
|
References
|
|---|
- Weyland A, Wietasch G, Hoeft A, et al. The effect of intracardiac left-right shunt on thermodilution measurements of cardiac output: an extracorporeal circulation model. Anaesthesist 1995; 44: 13–23.[Medline]
- Von Spiegel T. Cardiac output determination with transpulmonary thermodilution: an alternative to pulmonary catheterization? Anaesthesist 1996; 45: 1045–50.[ISI][Medline]
- McLuckie A, Marsh M, Andersen D, et al. A comparison of pulmonary and femoral artery thermodilution cardiac indices in paediatric intensive care patients. Acta Paediatr 1996; 85: 336–8.[ISI][Medline]
- Sakka SG, Reinhart K, Meier-Hellmann A. Comparison of pulmonary artery and arterial thermodilution cardiac output in critically ill patients. Intensive Care Med 1999; 25: 843–6.[ISI][Medline]
- Wesseling KH, Smith NT, Nicholas VW, et al. Beat to beat cardiac output from the arterial pressure pulse contour. In: Boerhave U, ed. Course on measurement in anesthesia. Leiden, University of Leiden Press, 1974: 150–64.
- Wesseling KH, Jansen JRC, Settels JJ, et al. Computation of aortic flow from pressure in humans using a nonlinear three element model. J Appl Physiol 1984; 74: 2566–73.
- Gödje O, Hoke K, Goetz AE, et al. Reliability of a new algorithm for continuous cardiac output determination by pulse-contour analysis during hemodynamic instability. Crit Care Med 2002; 30: 52–8.[ISI][Medline]
- Stok WJ. Noninvasive cardiac output measurement in orthostasis: pulse contour analysis compared with acetylene rebreathing. J Appl Physiol 1999; 87: 2266–73.[Abstract/Free Full Text]
- Gödje O, Thiel C, Lamm P. Less invasive, continuous hemodynamic monitoring during minimally invasive coronary surgery. Ann Thorac Surg 1999; 68: 1532–6.[Abstract/Free Full Text]
- Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307–10.[ISI][Medline]
- Buhre W. Comparison of cardiac output assessed by pulse-contour analysis and thermodilution in patients undergoing minimally invasive direct coronary artery bypass grafting. J Cardiothorac Vasc Anesth 1999; 13: 437–40.[ISI][Medline]
- Gödje O, Höke K, Lichtwark-Aschoff M. Continuous cardiac output by femoral arterial thermodilution calibrated pulse contour analysis: comparison with pulmonary arterial thermodilution. Crit Care Med 1999; 27: 2407–12.[ISI][Medline]
- Zöllner C, Haller M, Weis M. Beat to beat measurement of cardiac output by intravascular pulse contour analysis: a prospective criterion standard study in patients after cardiac surgery. J Cardiothorac Vasc Anesth 2000; 14: 125–9.[ISI][Medline]
- Rauch H, Muller F, Fleischer F, et al. Pulse contour analysis versus thermodilution in cardiac surgery patients. Acta Anaesthesiol Scand 2002; 46: 424–9.[ISI][Medline]
- Gödje O, Peyerl M, Seebauer T, et al. Central venous pressure, pulmonary capillary wedge pressure and intrathoracic blood volumes as preload indicators in cardiac surgery patients. Eur J Cardiothorac Surg 1998; 13: 533–8.
- Sakka SG, Bredle DL, Reinhart K, et al. Comparison between intrathoracic blood volume and cardiac filling pressures in the early phase of hemodynamic instability of patients with sepsis or septic shock. J Crit Care 1999; 514: 78–83.
- Bindels AJ, Van Der Hoeven JG, Graafland AD, et al. Relationship between volume and pressure measurements and stroke volume in critically ill patients. Crit Care 2000; 54: 193–9.
- Sakka SG, Meier-Hellman A. Evaluation of cardiac output and cardiac preload. In: Vincent JL, ed. Yearbook of intensive care and emergency medicine. New York: Springer Verlag, 2000; 671–9.
- Holm C, Melcer B, Horbrand F, et al. Intrathoracic blood volume as an end point in resuscitation of the severely burned: an observational study of 24 patients. J Trauma 2000; 548: 728–34.
- Gödje O, Seebauer T, Peyerl M, et al. Hemodynamic monitoring by double indicator dilution technique in patients after orthotopic heart transplantation. Chest 2000; 118: 775–81.[Abstract/Free Full Text]
- Mundigler G, Heinze G, Zehetgruber M. Limitations of transpulmonary indicator dilution method for assessment of preload changes in critically ill patients with reduced left ventricular function. Crit Care Med 2000; 28: 2231–7.[Medline]
Accepted for publication May 28, 2003.
This article has been cited by other articles:
|
|
|
|
 
O. J. Liakopoulos, J. K. Ho, A. Yezbick, E. Sanchez, C. Naddell, G. D. Buckberg, R. Crowley, and A. Mahajan
An Experimental and Clinical Evaluation of a Novel Central Venous Catheter with Integrated Oximetry for Pediatric Patients Undergoing Cardiac Surgery
Anesth. Analg.,
December 1, 2007;
105(6):
1598 - 1604.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
U. Fakler, Ch. Pauli, G. Balling, H.P. Lorenz, A. Eicken, M. Hennig, and J. Hess
Cardiac index monitoring by pulse contour analysis and thermodilution after pediatric cardiac surgery
J. Thorac. Cardiovasc. Surg.,
January 1, 2007;
133(1):
224 - 228.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
