Measurement of cardiac output: a comparison between transpulmonary thermodilution and uncalibrated pulse contour analysis

Br J Anaesth 2007;99:337-342

Sakka SG, Kozieras J, Thuemer O, van Hout N




Purpose            The purpose of this study was to evaluate the accuracy of Vigileo (Edwards Lifesciences, Irvine, CA) pulse contour cardiac output (PCCO) measurements in patients being treated for sepsis.

Background            Cardiac output monitoring is one of the most important hemodynamic variables in critically ill patients. Vigileo Pulse Contour Cardiac Output (PCCO) monitoring generates a cardiac output value by analysis of an arterial pressure waveform. It requires a peripheral arterial line and demographic information specific to the patient it is being used to monitor (e.g. height, weight). The Vigileo system does not use in vivo calibration with a reference source, such as thermodilution cardiac output. Studies evaluating the accuracy of PCCO measurements have yielded contradictory results.

Methodology            This prospective study included 24 sedated and mechanically ventilated septic patients. All patients were being monitored with a transpulmonary cardiac output (TpCO) monitor and all were receiving norepinephrine as part of their treatment. Vigileo Pulse Contour Cardiac Output monitoring was added for the purpose of the study and its cardiac output values compared to those of the TpCO monitor. Each patient served as their own control.

Baseline cardiac output (CO) measurements were obtained with TpCO and PCCO. The mean arterial pressure (MAP) was then increased with norepinephrine and a second set of CO measurements were recorded. A third set of control CO measurements were recorded after the norepinephrine infusion was reduced to its previous rate. Each measurement was taken five minutes after a stable MAP was achieved. Fluid status, airway pressure, and FIO2, were unchanged during the study period.

Result            The mean APACHE II score of the subjects was 26. Their ages ranged from 26 to 77 years. There were 16 males and 8 females. Vigileo PCCO correlated poorly with TpCO over all measurements (r2=0.26, P<0.0001). In general Vigileo PCCO measurements were lower than TpCO, but in some cases they were much higher.

Conclusion            The Vigileo PCCO system does not correlate well with transpulmonary thermodilution cardiac output.



The previous abstract and comment in this issue examined the same minimally invasive cardiac output technology with one distinction, it was made by a different manufacturer. (See Reliability of continuous pulse contour cardiac output measurement during hemodynamic instability elsewhere in this issue.) In addition to the device being made by a different manufacturer, there are a couple important differences in the studies themselves. First, this study looked at human patients, rather than an animal model. Second, the patients included in this study were examined during a period of hemodynamic stability. Past studies of PCCO technology have shown better accuracy during periods of hemodynamic stability than during unstable hemodynamics. (This fact alone is troubling. Why else would we want to monitor vital signs except to know when they had changed?) The last difference is the most informative. The previous study examined the difference between the PCCO values and thermodilution cardiac output values. They found that the differences were quite large much of the time. This study looked at the correlation between PCCO values and thermodilution values. Even if PCCO was not accurate, if it went up when thermodilution went up and down when thermodilution went down it could still correlate well. In that case, the PCCO numbers might not mean much by themselves but we would at least know when the cardiac output was going up or down. We’d know something qualitative about cardiac output. The first study showed that PCCO wasn’t accurate. This one showed that it didn’t correlate well with thermodilution either.

Two words of caution are warranted. It will take more than two studies to determine whether or not PCCO is accurate and/or reliable. Some other studies have shown better results. I don’t see a consensus yet. Often times, when a new product is under development different manufacturers produce more or less accurate devices of the same type. To know if the technology works all the devices will have to be tested. And, I feel obligated to reiterate, both of these studies compared PCCO to thermodilution cardiac output. Thermodilution cardiac output has a significant level of inaccuracy. It may be accurate enough to compare the current crop of fairly inaccurate PCCO devices against. But if PCCO devices get more accurate, they will need to be compared to a better “gold standard,” such as fick dye dilution cardiac output.


Michael Fiedler, PhD, CRNA




Correlation examines whether or not two things change together. If there is no correlation between HR and BP the correlation coefficient is 0. If HR and BP both move up equally at the same time the correlation coefficient would be a perfect 1. If HR moved down half as much as BP moved up the correlation would be -0.5.