Published in European Journal of Anaesthesiology January 2015 – Volume 32 – Issue 1 – p 5–12
Authors: Slagt et al
BACKGROUND: Cardiac output (CO) measurement is often required in critically ill patients. The performances of newer, less invasive techniques require evaluation in patients with severe sepsis and septic shock.
OBJECTIVES: To compare calibrated arterial pressure waveform analysis-derived CO(COap, VolumeView/EV1000) and the uncalibrated form (COfv, FloTrac/Vigileo) with transpulmonary thermodilution derived CO (COtptd).
DESIGN: A prospective, observational, single-centre study.
SETTING: ICU of a general teaching hospital.
PATIENTS: Twenty consecutive patients with severe sepsis or septic shock requiring haemodynamic monitoring by VolumeView/EV1000 and receiving mechanical ventilation.
INTERVENTION: Connection of FloTrac/Vigileo to radial artery catheter already in situ.
MAIN OUTCOME MEASURES: Radial (COfv) and femoral (COap) arterial waveform-derivedCO measurements were compared with COtptd with respect to bias, precision, limits of agreement and percentage error, and the percentage error in the course of time since the last calibration of COap by COtptd.
RESULTS: In comparing COap with COtptd (n = 267 paired measurements), the bias was 0.02 and limits of agreement were −2.49 to 2.52 l min−1, with a percentage error of 31%. The percentage error between COap and COtptd remained less than 30% until 8 h after calibration. In comparing COfv with COtptd (n = 301), the bias was −0.86 l min−1 and limits of agreement were −4.48 to 2.77 l min−1, with a percentage error of 48%. The biases of COap and COfv correlated with systemic vascular resistance [r = 0.13 (P = 0.029) and r = 0.42 (P < 0.001), respectively]. Clinically significant changes in COap and COfv correlated positively with COtptd at r = 0.51 (P < 0.001) and r = 0.64 (P < 0.001), respectively.
CONCLUSION: There was moderate agreement when measuring CO with either arterial waveform analysis technique. Compared with the uncalibrated COfv, the recently introduced calibrated arterial pressure waveform analysis-derived COap was more accurate and less dependent on vascular tone for up to 8 hours after callibation when monitoring CO in patients with severe sepsis and septic shock. The COap and COfv methods have poor to moderateCO-tracking abilities.
This article is accompanied by the following Invited Commentary:
Crossingham I, Columb M. Moderate agreement for cardiac output monitors. Moderately good or moderately bad? Eur J Anaesthesiol 2015; 32:1–2.
Introduction
The VolumeView/EV1000 (Edwards Lifesciences, Irvine, California, USA) is a new transpulmonary thermodilution (TPTD) device that has recently been introduced to monitor cardiac output (CO). It utilizes TPTD to calculate CO (COtptd) and, using this result, calibrates the system which estimates cardiac output from the arterial pressure waveform (COap). Thereafter, arterial pressure waveform analysis is used to estimate continuous CO (CCO), similar to the technique embedded in the FloTrac/Vigileo (Edwards Lifesciences) monitor (COfv). With each TPTD measurement, the COap is recalibrated.1There is a single validation study in relatively haemodynamically stable patients, suggesting that COap is as accurate as the TPTD reference and superior to pulse contour-derived (PiCCO2; Pulsion Medical Systems, Munich, Germany).1 Both techniques (PiCCO2 and VolumeView) use TPTD with the Stewart-Hamilton equation to calculate CO. We previously suggested2,3 that even with the most recent third-generation FloTrac/Vigileo software, the calculated (uncalibrated) CO is still too inaccurate in vasodilated and septic patients to allow for clinically meaningful CO monitoring and therapeutic decision making.
The aim of the current prospective, observational, single-centre study was to compare calibrated COap and uncalibrated COfv with COtptd in critically ill patients with severe sepsis or septic shock in the ICU. We hypothesised that calibrated would outperform uncalibrated COmeasurement in this setting.
Materials and methods
After Medical Ethics Review Committee approval (Ethics Committee, Noord-Holland, Alkmaar, The Netherlands No. M011-019; 26 April 2011) and written informed consent, 20 consecutive patients were included in this observational study. All patients with severe sepsis or septic shock (as defined by the American College of Chest Physicians/Society of Critical Care Medicine consensus conference) requiring vasoactive therapy along with monitoring of CO, radial arterial pressure and central venous pressure (CVP) were eligible for this study. Additional inclusion criteria that had to be met were the use of norepinephrine, organ failure and mechanical ventilation. The study was performed in the ICU of a general teaching hospital from June 2011 to April 2013. Exclusion criteria were age less than 18 years, contraindications for a femoral artery catheter and known severe tricuspid or aortic valvular regurgitation. The study did not otherwise alter the standard of care provided.
Protocol
Patient characteristics were recorded, including disease severity scores. Paired CO measurements were performed as clinically required, at least once every shift and with any changes to therapy with fluids and vasoactive agents. A FloTrac sensor was connected to the 20-guage radial arterial cannula already in situ and then connected to the Vigileo monitor. After placement of the TPTD catheter (VolumeView/EV1000 system; Edwards Lifescience) in the femoral artery, all time clocks from the Spacelab monitor (Spacelabs Medical Inc, Issaquah, Washington, USA), VolumeView/EV1000 and FloTrac/Vigileo monitors were synchronised. The TPTD measurement was performed in sets of three to five bolus injections of 20 ml iced isotonic saline through the central venous catheter irrespective of the ventilator cycle. All individual bolus measurements had to be validated before being averaged. The mean value was recorded and regarded as the COtptd. At the start of each TPTD CO measurement, the COfv was measured and the mean value was recorded. All haemodynamic data stored in the EV1000 computer and FloTrac/Vigileo monitor were downloaded for analysis. All paired COap and COtptd measurements were also analysed in relation to the time after the last calibration to establish the calibration-free period in which COap remains clinically acceptable with a percentage error less than 30%.4 The mean arterial pressure (MAP) was monitored using the femoral arterial catheter and the central venous pressure (CVP) from a central venous catheter (internal jugular or subclavian vein), inserted for clinical reasons. The systemic vascular resistance (SVRtptd) was calculated from (MAP - CVP) x 80/COtptd, dyne s cm−5. The electrocardiogram was monitored throughout and heart rate (HR) was recorded.
Description of techniques
The TPTD measurement using the VolumeView/EV1000 system uses a bolus injection through a central venous catheter situated above the diaphragm and a femoral arterial catheter with a specific thermistor tip subsequently measures the thermodilution curve. This method provides CO measurements as well as volumetric indices such as global end-diastolic volume (GEDV), intrathoracic blood volume (ITBV), extravascular lung water (EVLW), global ejection fraction and pulmonary vascular permeability index.5 The CO is estimated from the bolus TPTD measurements using the Stewart–Hamilton equation.1 For the CCO measurements, the VolumeView/EV1000 monitor combines the area under the systolic part of the arterial pressure waveform and waveform analysis as used in the FloTrac/Vigileo system. The exact method of how both algorithms are integrated into the COap measurement has not been disclosed at this time. After each intermittent bolus TPTD measurement, the COap is recalibrated.1
The FloTrac/Vigileo system estimates CO by using the standard deviation of the pulse pressure, incorporating actual vascular tone based on waveform analysis and patient characteristics.6 The arterial waveform is analysed over 20 s with a frequency of 100 Hz. The third-generation software version (3.02) includes two separate models for the arterial tone factor: first, an arterial tone model that was developed predominantly from patients who did not have a hyperdynamic circulation (this is the same model used in the previous version 1.10); and second, an arterial tone model that was developed predominantly from patients who had a hyperdynamic circulation. The need for using two separate models is because of the differences in relating peripheral arterial pressure to flow during nonhyperdynamic and hyperdynamic circulations.7 The switching between the two models is automated using an algorithm that analyses 14 parameters of the arterial pressure waveform to detect the occurrence of hyperdynamic circulation.
Statistical analyses
After confirming normal distribution of data using the Kolmogorov–Smirnov test, data were pooled and summarised as mean (SD) and parametric analyses were performed. Pearson correlation coefficients of pooled data were estimated using Statistical Package for Social Sciences (SPSS) version 21 (SPSS Inc, Chicago, Illinois, USA). To assess agreement, a Bland-Altman analysis8 was performed and adjusted for repeated measurements (Medcalc software version 12.2.1.0; Mariakerke, Belgium). Bias was defined as the mean difference between CO derived from two methods. Limits of agreement were calculated from ±1.96 SD of the bias. The percentage error (1.96 SD/mean CO) was calculated with 30% taken as clinically acceptable.9 The precision of the reference COtptd was calculated using the method proposed by Cecconi et al.4 to calculate the precision of the test methods. Polar plots (SigmaPlot software version 11, San Jose, California, USA) were also used to analyse the agreement in CO trend monitoring between methods.10 In the polar plot, the changes of CO data are converted to a radial vector wherein the degree of agreement between two devices becomes the angle between radial vector and horizontal axis (i.e. the polar axis). If agreement is perfect, the radial vector lies along the polar axis and the angle is zero; the angle between the vector and the horizontal axis represents disagreement. Agreement for trending is acceptable when points lie between either 150° and 210° or 30° and 330°. The distance from the centre of the plot (vector) represents the mean changes inCO. Concordances of clinically significant changes in COmeasurements are reported. As COap is recalibrated with COtptd, we analysed the effect of time on the difference between the two methods. A P value less than 0.05 was considered statistically significant.
Results
The characteristics of the 20 patients included in the study are summarised in Table 1. The haemodynamic data at inclusion are given in Table 2. A total of 301 paired measurements were obtained, with the number of paired measurements per patient ranging from 5 to 24. The COap, COfv and COtptd were 8.2 (2.5), 7.3 (3.6) and 8.2 (2.3) l min−1, respectively, and the SVRtptd was 636 (246) dyne s cm−5. Eighty-five percent of the intermittent measurements were performed during norepinephrine infusions at a median (range) dose of 0.20 (0.02 to 1.70) μg kg−1 min−1.
As the first COtptd measurement in each patient was used to calibrate COap (so not used for analysis) and with data storage errors for 14 other COap measurements, 267 COap measurements were available for comparison with COtptd. COap ranged from 2.7 to 15.1 l min−1 and COtptd ranged from 2.6 to 15.9 l min−1. Bland–Altman analysis corrected for repeated measurements showed a bias (COap − COtptd) of 0.02 l min−1 and limits of agreement of −2.49 to 2.52 l min−1, with a percentage error of 31%. The precision of the COtptd was 6.7%. Using the formulae presented by Cecconi et al.,4we estimated that the precision of COap is 30%, from the shared percentage error of 31%. This implies that the COap has a relative reduction in precision by a factor of 4.5 compared with COtptd. The pooled bias of COap − COtptd was correlated with SVRtptd (r = 0.13, P = 0.029).
The pooled data for each 2 h time period since last calibration are summarised in Table 4; the COap percentage error remained within the defined clinically acceptable range of 30%9 up to 8 h after the last calibration.
After excluding CO changes less than 1.2 l min−1 or less than 15% of mean CO11 in either reading, 130 measurements remained. The concordance of changes was 71%. When restricted to CO changes of greater than 1.2 l min−1 in both measurements, only 39 paired measurements remained with a concordance of 74%. Expressing changes in CO in the polar plot after elimination of central noise (COchange less than 1.2 l min−1), 46 measurements (68%) out of 68 were within the 30° to 330° or 150° to 210° of the polar axis (Fig. 3), suggesting moderate trending capability.
For the comparison COfv with COtptd, 301 paired measurements were available (Figs 1c and 1d and Table 3). COfv ranged from 2.9 to 17.1 l min−1and COtptd ranged from 2.6 to 15.9 l min−1. Bland–Altman analysis corrected for repeated measurements showed a bias of −0.86 l min−1 and limits of agreement of −4.48 to 2.77 l min−1, with a percentage error of 48%. With a precision of 6.7% for COtptd, the COfv precision was estimated at 47.5% and this represented a 7.1-fold relative reduction in precision with COfv compared with COtptd. The pooled bias of COfv − COtptd was correlated with SVRtptd (r = 0.42, P < 0.001).
After excluding CO changes less than 1.2 l min−1 or less than 15% of mean CO11 in either measurement, 142 measurements remained. The concordance of these changes was 77%. When only CO changes of greater than 1.2 l min−1 in both readings were used, 40 measurements remained with a concordance of 88%. Expressing changes in CO in the polar plot after elimination of central noise (CO less than 1.2 l min−1), 62 out of 82 (76%) were within the 30° to 330° or 150° to 210° of the polar axis as shown in Fig. 3. There was, however, no significant difference for COfv compared with COap, (68 versus 76%; P = 0.36 by Fisher exact test). This reflects moderate trending capabilities during severe sepsis and septic shock.
Discussion
To our knowledge, this is the first study using the VolumeView/EV1000 monitor in patients with severe sepsis or septic shock. Our results show that calibrated arterial pressure wave analysis (COap) derived from this monitor is more accurate than uncalibrated pressure waveform analysis COfv. However, COap still has a percentage error of 31% compared with the COtptd. COap remains within the clinically acceptable range (percentage error <30%)4 up 8 h after calibration. Trending capacity of both measurements are only considered moderate.
The VolumeView/EV1000 system has been validated in animal12 and human studies.1 Both studies reported good agreement and interchangeability with the reference measuring device (TPTD). Our results are comparable to those obtained by Bendjelid et al.1 who compared the pressure waveform analysis CO measured by the EV1000 with the pulse contour CO measured by PiCCO2 in 72 critically ill patients. However, the clinical conditions were different in that less than 10% of patients were septic, more than 50% underwent cardiac surgery, and measurements were performed under various modes of ventilatory support. This contrasts with our study where all patients were suffering from septic shock.
Our comparison of calibrated COap with COtptd in severe sepsis and septic shock revealed a percentage error of 31%, almost reaching the Critchley and Critchley criterion for interchangeability of 30%.9Analysing the time interval between COtptd measurements (recalibration of the COap) showed that the percentage error between COtptd and COap remained within 30% until 8 h after last calibration. Nevertheless, in clinical practice, earlier recalibration may be indicated during periods of haemodynamic instability when resuscitation with fluids, inotropes or vasopressors may result in large changes in vascular tone. However, the percentage errors of 21 and 27% for the initial two periods following calibration suggest that the calibrated waveform analysis may cope with haemodynamic instability. Conversely, our results suggest that haemodynamic stability does not prevent the need for recalibration at least once every 8 h. Hamzaoui et al.13 investigated the accuracy of the PiCCO2 pulse contour method (Pulsion Medical Systems, Munich, Germany) in critically ill patients, most of whom had sepsis or septic shock. They found that the percentage error increased above 30% 1 h following calibration and, therefore, recommended early recalibration. Jellema et al.14 and Gödjeet al.15 suggested that CCO (Modelflow and PiCCO, respectively) remained accurate up to 44 h after initial calibration, but it is unclear whether the Critchley criteria were met as no percentage errors were reported. Hamzaoui et al.13 and our data suggest a shorter time period, which seems more realistic. Critchley and Critchley9 have shown that the percentage error between two CO measurement methods consists of the precision of both the reference and the new method.9 The precision of the TPTD in our study was good (6.7%) in line with previous reports in the literature.13,16 The precisions of the COap and COfv were not as impressive in our study at 30 and 47.5%, respectively. Recently, the Critchley criteria have been challenged, as many noninvasive CO measuring devices do not meet the 30% percentage error, commonly regarded as acceptable.17 In addition, Columb18 suggested using agreement and tolerability intervals to estimate the agreement:tolerability index (ATI) and define clinical agreement. An ATI of 1.0 or less suggests acceptable agreement, 1.0 to 2.0 implies marginal agreement and greater than 2.0 is outwith the limits of acceptable agreement. If we define a tolerability interval forCO of 4.0 l min−1 (assuming a hypodynamic circulation below 4.0 and a hyperdynamic circulation greater than 8.0 l min−1), the ATI is 1.2 for COap. Using the COfv, the ATI is 1.8, suggesting marginal agreement for both but supporting our conclusion that the calibrated arterial pressure waveform analysis performs better than the uncalibrated technique, when compared with TPTD in severe sepsis or septic shock.
The FloTrac/Vigile has proven accuracy in stable haemodynamic conditions3,19–22 but not in severe sepsis, septic shock2,3,11,23–25or liver surgery.3,26–31 Our results are consistent with the data in the literature. Large limits of agreement and a poor precision and percentage error of the COfv (which are time-dependent) do not render the technique interchangeable with COtptd in patients suffering from sepsis and septic shock. Indeed, our data emphasise the need for calibration of arterial pressure derived haemodynamic measuring devices in patients suffering from septic shock. As discussed previously,2,3 the SVR is an important factor influencing bias and precision and the COfv is unreliable at low SVR. Indeed, in septic patients with large changes in vascular resistance, most similar devices attempting to integrate vascular tone and arterial compliance to calculate CO6,14,32 are unreliable when uncalibrated.2,3,24,26–29 Our results are consistent with those obtained by Jellema et al.14 who compared the Modelflow method for arterial pressure derived COestimation with bolus thermodilution in patients with sepsis. They found that uncalibrated measurements were less precise compared with calibrated measurements (percentage errors 55 and 18%, respectively) and concluded that calibration of the Modelflow method was needed especially during hyperdynamic states.
Changes in CO following interventions are suggested to be of more clinical use than absolute values,10 even though the definitions of clinically significant changes are controversial.8,10,11 The VolumeView/EV1000 monitor combines TPTD and pressure waveform analysis to calculate COap and, as shown in our study, this gives a more robust CO measurement. However, both COfv and COap show only poor to moderate trending abilities. Changes in vascular tone are associated with larger changes in CO when measured by COfv.11,32–35 Changes in CO following therapeutic interventions are reported to be important clinical responses in validating new monitoring devices.36 Concordance in trending of at least 90% implies clinically acceptable tracking capability.10 This was not reached for either device in our study and this also is consistent with the literature.2,3,11,25
Limitations of our study include the simple observational design and the use of a single centre. The generalisability of our results is therefore limited. An important consideration when assessing the effect of the elapsed time since last calibration is the limited availability of data, particularly beyond 8 h. The issues of repeated measurements and pragmatic pooling of data, although incompletely addressed, are unlikely to materially affect our findings. Using the arterial pressure waveform to calculate CO places a high demand on the quality of the signal and this can result in loss of data. In addition, the sites wherein the arterial signals are sampled, namely radial and femoral arteries, may affect measurements.37–40 As COfv was derived from the radial and COap from the femoral artery, we cannot exclude systematic errors by measurement site.
Conclusion
There was moderate agreement with thermodilution when measuringCO with either arterial waveform analysis technique. Compared with the uncalibrated COfv, the recently introduced calibrated arterial pressure waveform analysis-derived COap was more accurate and less dependent on vascular tone for up to 8 hours after callibation when monitoring CO in patients with severe sepsis and septic shock. The tracking abilities of both COap and COfv are poor to moderate.
References
1. Bendjelid K, Marx G, Kiefer N, et al. Performance of a new pulse contour method for continuous cardiac output monitoring: validation in critically ill patients. Br J Anaesth 2013; 111:573–579.
2. Slagt C, de Leeuw MA, Beute J, et al. Cardiac output measured by uncalibrated arterial pressure waveform analysis by recently released software version 3.02 versus thermodilution in septic shock. J Clin Monit Comput 2013; 27:171–177.
3. Slagt C, Malagon I, Groeneveld AB. Systematic review of uncalibrated arterial pressure waveform analysis to determine cardiac output and stroke volume variation. Br J Anaesth 2014; 112:626–637.
4. Cecconi M, Rhodes A, Poloniecki J, et al. Bench-to-bedside review: the importance of the precision of the reference technique in method comparison studies: with specific reference to the measurement of cardiac output. Crit Care 2009; 13:201.
5. Kiefer N, Hofer CK, Marx G, et al. Clinical validation of a new thermodilution system for the assessment of cardiac output and volumetric parameters. Crit Care 2012; 16:R98.
6. Pratt B, Roteliuk L, Hatib F, et al. Calculating arterial pressure-based cardiac output using a novel measurement and analysis method. Biomed Instrum Technol 2007; 41:403–411.
7. Hatib F, Jansen JR, Pinsky MR. Peripheral vascular decoupling in porcine endotoxic shock. J Appl Physiol 2011; 111:853–860.
8. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307–310.
9. 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.
10. Critchley LA, Yang XX, Lee A. Assessment of trending ability of cardiac output monitors by polar plot methodology. J Cardiothorac Vasc Anesth 2011; 25:536–546.
11. Monnet X, Anguel N, Naudin B, et al. Arterial pressure based cardiac output in septic patients: different accuracy of pulse contour and uncalibrated pressure waveform devices. Crit Care 2010; 14:R109.
12. Bendjelid K, Giraud R, Sigenthaler N, Michard F. Validation of a new transpulmonary thermodilution system to assess global end-diastolic volume and extravascular lung water. Crit Care 2010; 14:R209.
13. Hamzaoui O, Monnet X, Richard C, et al. Effects of changes in vascular tone on the agreement between pulse contour and transpulmonary thermodilution cardiac output measurements within an up to 6-h calibration-free period. Crit Care Med 2008; 36:434–440.
14. Jellema WT, Wesseling KH, Groeneveld AB, et al. Continuous cardiac output in septic shock by simulating a model of the aortic input impedance: a comparison with bolus injection thermodilution.Anesthesiology 1999; 90:1317–1328.
15. Gödje O, Höke 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; 30:52–58.
16. Holm C, Mayr M, Hörbrand F, et al. Reproducibility of transpulmonary thermodilution measurements in patients with burn shock and hypothermia. J Burn Care Rehabil 2005; 26:260–265.
17. Peyton PJ, Chong SW. Minimally invasive measurement of cardiac output during surgery and critical care: a meta-analysis of accuracy and precision. Anesthesiology 2010; 113:1220–1235.
18. Columb MO. Clinical measurement and assessing agreement. Curr Anaesth Crit Care 2008; 19:328–329.
19. Jo YY, Song JW, Yoo YC, et al. The uncalibrated pulse contour cardiac output during off-pump coronary bypass surgery: performance in patients with a low cardiac output status and a reduced left ventricular function. Kor J Anesthesiol 2011; 60:237–243.
20. Broch O, Renner J, Gruenewald M, et al. A comparison of third-generation semi-invasive arterial waveform analysis with thermodilution in patients undergoing coronary surgery. Sci World J 2012; 2012:45108119.
21. Mutoh T, Ishikawa T, Kobayashi S, et al. Performance of third-generation FloTrac/Vigileo system during hyperdynamic therapy for delayed cerebral ischemia after subarachnoid hemorrhage. Surg Neurol Int 2012; 3:99.
22. Vasdev S, Chauhan S, Choudhury M, et al. Arterial pressure waveform derived cardiac output FloTrac/Vigileo system (third generation software): comparison of two monitoring sites with the thermodilution cardiac output. J Clin Monit Comput 2012; 26:115–120.
23. Sakka SG, Kozieras J, Thuemer O, van Hout N. Measurement of cardiac output: a comparison between transpulmonary thermodilution and uncalibrated pulse contour analysis. Br J Anaesth 2007; 99:337–342.
24. Slagt C, Beute J, Hoeksema M, et al. Cardiac output derived from arterial pressure waveform analysis without calibration versus thermodilution in septic shock: evolving accuracy of software versions. Eur J Anaesthesiol 2010; 7:550–554.
25. Monnet X, Anguel N, Jozwiak M, et al. Third-generation FloTrac/Vigileo does not reliably track changes in cardiac output induced by norepinephrine in critically ill patients. Br J Anaesth 2012; 108:615–622.
26. Biais M, Nouette-Gaulain K, Cottenceau V, et al. Cardiac output measurement in patients undergoing liver transplantation: pulmonary artery catheter versus uncalibrated arterial pressure waveform analysis. Anesth Analg 2008; 106:1480–1486.
27. Della Rocca G, Costa MG, Chiarandini P, et al. Arterial pulse cardiac output agreement with thermodilution in patients in hyperdynamic circulation. J Cardiothorac Vasc Anesth 2008; 22:681–687.
28. Krejci V, Vannucci A, Abbas A, et al. Comparison of calibrated and uncalibrated arterial pressure-based cardiac output monitors during orthotopic liver transplantation. Liver Transpl 2010; 16:773–782.
29. Akiyoshi K, Kandabashi T, Kaji J, et al. Accuracy of arterial pressure waveform analysis for cardiac output measurement in comparison with thermodilution methods in patients undergoing living donor liver transplantation. J Anesth 2011; 25:178–183.
30. Biancofiore G, Critchley LA, Lee A, et al. Evaluation of a new software version of the FloTrac/Vigileo (version 3.02) and a comparison with previous data in cirrhotic patients undergoing liver transplant surgery. Anesth Analg 2011; 113:515–522.
31. Su BC, Tsai YF, Chen CY, et al. Cardiac output derived from arterial pressure waveform analysis in patients undergoing liver transplantation: validity of a third-generation device. Transplant Proc 2012; 4:424–428.
32. Wesseling KH, Jansen JR, Settels JJ, Schreuder JJ. Computation of aortic flow from pressure in humans using a nonlinear, three-element model. J Appl Physiol 1993; 74:2566–2573.
33. Lorsomradee S, Lorsomradee SR, Cromheecke S, De Hert SG. Continuous cardiac output measurement: arterial pressure analysis versus thermodilution technique during cardiac surgery with cardiopulmonary bypass. Anaesthesia 2007; 62:979–983.
34. Eleftheriadis S, Galatoudis Z, Didilis V, et al. Variations in arterial blood pressure are associated with parallel changes in FloTrac/Vigileo-derived cardiac output measurements: a prospective comparison study. Crit Care 2009; 13:R179.
35. Meng L, Tran NP, Alexander BS, et al. The impact of phenylephrine, ephedrine, and increased preload on third-generation Vigileo-FloTrac and esophageal Doppler cardiac output measurements.Anesth Analg 2011; 113:751–757.
36. Squara P, Cecconi M, Rhodes A, et al. Tracking changes in cardiac output: methodological considerations for the validation of monitoring devices. Intensive Care Med 2009; 35:1801–1808.
\37. 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 2006; 10:R164.
38. Sander M, Soies CD, Foer A, Von Heymann C. Cardiac output measurement by arterial waveform analysis in cardiac surgery: a comparison of measurements derived from waveforms of the radial artery versus the ascending aorta. J Int Med Res 2008; 36:414–419.
39. Hamm J-B, Nguyen B-V, Kiss G, et al. Assessment of a cardiac output device using arterial pulse waveform analysis, Vigileo, in cardiac surgery compared to pulmonary arterial thermodilution. Anaesth Intensive Care 2010; 38:295–301.
40. Hofer CK, Button D, Weibel L, et al. Uncalibrated radial and femoral arterial pressure waveform analysis for continuous cardiac output measurement: an evaluation in cardiac surgery patients. J Cardiothorac Vasc Anesth 2010; 24:257–264.