1 Madrid, Spain 2 Udine, Italy
EditorWe read with interest the paper by Della Rocca and colleagues.1 They tested the accuracy of three methods of measurement of cardiac output (CO)the femoral artery bolus thermodilution method (COart), pulse contour-derived CO measurement (PCCO), and semicontinuous pulmonary artery pulsed warm thermodilution (CCO)against the clinical standard of pulmonary artery bolus thermodilution (COpa), during ortotopic liver transplantation. We would like to challenge some aspects of the study, related to the methods and the discussion of the results.
We question the method used by Della Rocca and colleagues to compare the pulse contour-derived cardiac output with the thermodilution method. Because of the automatic recalibration of the system made by the computer of the equipment in use, the value of PCCO used in the comparisons is an average of the real PCCO value taken before the femoral artery thermodilution measurement, and the corrected PCCO value taken immediately after the thermodilution measurement.2 The new thermodilution-derived calibration factor attenuates the variation of the pulse contour measurement of the CO in changing haemodynamic conditions, introducing an error in the conclusions about the agreement and precision of the PCCO vs the thermodilution method. This fact was not discussed by the authors and it was concluded, incorrectly, that there was a lack of dependence of PCCO on the system vascular resistance.
Furthermore, we think that the conclusion of Della Rocca and colleagues that good agreement and precision were found between all the methods of CO monitoring at each time point in the study period is questionable. We agree with the statistical method used for the comparisons, but the comment in the description by Bland and Altman3 did not give precise criteria for what is a good and what is a poor agreement in relative units. A possible way to evaluate the scatter in the paired measurements is by using the relative error of the bias and the limits of agreement as a percent of the absolute mean value of the standard method.4 Using the data in Tables 3 and 6 of Della Roccas study, we calculated the relative errors for all the paired comparisons (Table 1). Given the 15% variability in the pulmonary artery thermodilution method,5 we find the 30% scatter in the measurements taken in the femoral artery after vena cava clamping too great for good agreement to be concluded.
|
I. G. Tzenkov
J. Perez Peña
Madrid, Spain
EditorWe appreciate the interest of Drs Tzenkov and Perez Peña in our article,1 and the opportunity to reply. The letter confirms that interest in this field is still significant.
As previously reported by Rodig, Goedje and Bottiger,611 we measured pulse contour cardiac output (PCCO) and continuous cardiac output (CCO) immediately before and after intermittent CO measurements, and the mean of these continuous data pairs was recorded. As the pulse contour-based values might have changed during the thermodilution period, because of the online character of the system, the average of the values immediately before and after each set of thermodilutions were used for statistical analysis. It has been questioned whether the bolus thermodilution technique is suitable as a reference method for validating continuous thermodilution, because comparison of the two methods cannot be simultaneous, and an individual thermal washout curve derived from bolus injection reflects a shorter period of blood flow than that averaged for 36 min with CCO (for pulmonary artery catheter monitoring). To minimize this problem, we and others12 used the average of stable CCO values before and after bolus measurements for comparison.
As far as the dependence of the PCCO on the system vascular resistance is concerned, we mentioned this in the discussion section, where we agreed with the findings of Rödig and colleagues.8 They showed that changes in vascular tone of approximately 20% did not affect the pulse contour method, but large changes in arterial pressure may affect PCCO measurements, and a new calibration of the PCCO device is then necessary. The finding that moderate changes in systemic vascular resistance (SVR) did not necessarily affect the accuracy of the PCCO is supported by the findings of Irlbeck and colleagues.13 These authors studied PCCO and the intermittent pulmonary artery thermodilution technique (COpa) in critically ill patients and concluded that the PCCO is valid for clinical purposes only if the initial calibration is repeated every 4 h. We found no evidence that PCCO was inaccurate, even with substantial changes in SVR, but the changes in vascular tone in our patients were probably less and had no effect on the pulse contour method.
We used the Bland and Altman3 method for statistical analysis, of all the limitations (good and poor agreement in relative units). We confirm the good agreement between the techniques during the study period as a whole, but, as we reported in a previous letter,14 no significant difference in the agreement and precision of PCCO, CCO, or transpulmonary thermodilution (COart) vs bolus pulmonary artery thermodilution (COpa) was found during the study period. As previously reported by Bottiger and colleagues,10 11 during the cross-clamping and reperfusion phases, we also observed a lack of accuracy, even if the difference was not statistically significant. Moreover, changes in CO and their accuracy are clinically more important than the accuracy of stable values at different stages of the anaesthetic. For this reason, we analyzed the changes as bias, and 2 SD as the correlation coefficient.1 14 In our study, the analysis of the changes () confirmed the efficacy of the techniques during different stressing or extreme haemodynamic conditions, even if with a lesser degree of agreement and precision.14 This may be because of the corresponding lack of accuracy during the clamping and reperfusion phases.
Based on Tzenkov and Perez Peñas suggestion, we calculated the relative error of our collected data as described by Rodig and colleagues:8
100*(COpaCOart)/[(COpa+COart)/2]
100*(COpaPCCO)/[(COpa+PCCO)/2]
100*(COpaCCO)/[(COpa+CCO)/2]
The Students t-test was used to test whether there was a difference between CO values assessed by the different techniques, whether the bias within each group at each time point was different from zero, and to compare the bias and the relative error between the groups at each time point. The Bonferroni correction was applied to the level of significance to compensate for the effect of multiple comparisons. The relative error was within 15% for 70% of the comparisons between COart and COpa, for 75% of the comparisons between PCCO and COpa, and for 80% of the comparisons between CCO and COpa. The mean relative error in all patients for COpa vs COart, COpa vs PCCO and COpa vs CCO is given in Table 2. The mean relative error per phase is presented in Table 3. Bias, correlation and relative error were compared for all three measurements in the different surgical phases and at different levels of CO. No significant differences in the bias were found with regard to the different surgical phases. Independently of the values of cardiac output and the surgical phases, the mean relative error of the different methods studied was <6% in the whole group. But when analyzed at each surgical phase with regard to the relative errors of the COpa measurements, COart and PCCO increased. The standard deviations increased to >15% only during the anhepatic phase (T2).
|
|
Significant changes in COpa during the course of liver transplantation, defined as changes >15%, were previously demonstrated only occasionally using CCO.4 This difference can be explained by the large point-to-point variation in the bias of the measurements taken during the course of operation in the previous study.
The scatter in the measurements reported by Tzenkov and Perez Peña taken from the femoral artery recordings after inferior vena cava partial cross-clamping (piggy-back), had a lack of accuracy that has been previously reported.1 14 The piggy-back technique performed in this population study maintains caval flow without any effect on the transpulmonary artery thermodilution technique. However, based on the limits of the techniques studied (pulmonary artery catheter and PCCO System), which CO is the true CO cannot be detected by this study.
G. Della Rocca
M. G. Costa
Udine, Italy
References
1 Della Rocca G, Costa MG, Pompei L, Coccia C, Pietropaoli P. Continuous and intermittent cardiac output measurement: pulmonary artery catheter versus aortic transpulmonary technique. Br J Anaesth 2002; 88: 3506
2 System Description. In Operator's Manual, PULSION PiCCO, Version 4.1.x. München: PULSION Medical Systems, 2000; 1314
3 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; I: 30710
4 Greim CA, Roewer N, Thiel H, Laux G, Schulte am Esch J. Continuous cardiac output during adult liver transplantation: thermal filament technique versus bolus thermodilution. Anesth Analg 1997; 85: 4838[Abstract]
5 Jansen JRC, Schreuder JJ, Settles JJ, et al. Single injection thermodilution. A flow corrected method. Anesthesiology 1996; 85: 48190[ISI][Medline]
6 Gödje O, Höeke 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: 240712[ISI][Medline]
7 Gödje O, Thiel C, Lamm P, et al. Less invasive continuous hemodynamic monitoring during minimally invasive coronary surgery. Ann Thorac Surg 1999; 68: 15326
8 Rödig G, Prasser C, Keyl C, Liebold A, Hobbhahn J. Continuous cardiac output measurement: pulse contour analysis vs thermodilution technique in cardiac surgical patients. Br J Anaesth 1999; 82: 52530
9 Rödig G, Keyl C, Liebold A, Hobbhahn J. Intra-operative evaluation of a continuous versus intermittent bolus thermodilution technique of cardiac surgical patients. Eur J Anaesth 1998; 15: 196201[ISI][Medline]
10 Bottiger BW, Sinner B, Motsch J, Bach A, Bauer H, Martin E. Continuous versus intermittent thermodilution cardiac output measurements during orthotopic liver transplantation. Anaesthesia 1997; 52: 20714[CrossRef][ISI][Medline]
11 Bottiger BW, Rauch H, Bohrer H, et al. Continuous versus intermittent cardiac output measurements in cardiac surgical patients undergoing hypothermic cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1995; 9: 40511[ISI][Medline]
12 Schmid ER, Schmidlin D, Tornic M, Seifert B. Continuous thermodilution cardiac output: clinical validation against a reference technique of known accuracy. Intensive Care Med 1999; 25: 16672[CrossRef][ISI][Medline]
13 Irlbeck M, Forst H, Briegel J, Haller M, Peter K. Continuous measurement of cardiac output with pulse contour analysis. Anaesthesist 1995; 44: 493500[CrossRef][ISI][Medline]
14 Linton NWF, Linton RAF, Della Rocca G, Costa MG. Is comparison of changes in cardiac output, assessed by different methods, better than only comparing cardiac output to the reference method? Br J Anaesth 2002; 89: 3369
15 Taylor SH, Silke B. Is the measurement of cardiac output useful in clinical practice? Br J Anaesth 1988; 60: S9098[Medline]
16 Stetz CW, Miller RG, Kelly GE, Raffin TA. Reliability of the thermodilution method in the determination of cardiac output in clinical practice. Am Rev Resp Dis 1982; 126: 10014[ISI][Medline]