In Reply:—

We thank Kück et al.  for their response to our study of the NICO2monitor, 1because dialogue between the designers of equipment and those who put it to practical use is valuable. In this case, the clinical implications of our findings do not seem to be apparent to the authors, possibly because of different interpretations of the role of continuous monitoring devices. For the NICO2to be of clinical use as a continuous cardiac output (CO) monitor, it should display changes in CO values accurately immediately after minute ventilation (VV̇E) is decreased. In fact, we found that after 15 min it still underestimates CO. 1,2Claiming that the decrease in VV̇Ewas outside the typical range of change, Kück et al.  then suggest that we should wait even longer. Their case would be strengthened if they could specify appropriate delay periods for different levels of change: for example, how long must we wait for correct readings when VV̇Ereduction is less than 20%? Clinicians need to acquire reliable data in real time or, at least, know how to interpret values. If we must wait before getting reliable CO data from the NICO2and can trust the values only when VV̇Eis stable, its clinical value as a continuous monitoring device in all situations is open to question.

Having already reported the implications of VV̇Eas a factor in decreased precision and accuracy of NICO2monitoring, 1we were already aware that 15 min was insufficient time for venous carbon dioxide (Pv̄co2) to stabilize. We do not, however, fully agree when Kück et al.  claim that a baseline drift in Pv̄co2(ΔPv̄co2) leads to the underestimation by NICO2. We assume that Pv̄co2changes exponentially with a time constant of 35 min 3and that Pv̄co2and Paco2change in parallel to reach eventual values. Thus, because Paco2increased from 35.5 mmHg to 51.9 mmHg, 1ΔPv̄co2during the 50-s rebreathing period is calculated as 0.25 mmHg.

The NICO2system uses the following equation:


Here, ΔV̇CO2is the change in carbon dioxide production between normal breathing and carbon dioxide rebreathing; ΔCv̄CO2is the change in venous carbon dioxide content; and ΔCaCO2is the change in arterial carbon dioxide content. If the carbon dioxide dissociation curve is linear, the error caused by baseline drift in Cv̄CO2should mirror the ratio of ΔCv̄CO2/ΔCaCO2, which is equal to ΔPv̄co2/ΔPaco2. Because ΔPaco2during the rebreathing is approximately 2 to 6 mmHg, 4baseline drift of Pv̄co2at 0.25 mmHg may be responsible for 4–13% underestimation, much smaller than the value we observed (30% underestimation). 1The designers have proposed a new algorithm in which it is unnecessary for Pv̄co2to reach a constant during carbon dioxide rebreathing CO measurement. 5We do not know if the system that we evaluated incorporated this revision, but even if it did, something else may be causing the discrepant underestimation. 6 

In clinical situations, whether intentional or unintentional, reduction of VV̇Eto half is common. Mechanical ventilation at low tidal volume is a standard technique in cases of acute lung injury or adult respiratory distress syndrome. For example, in treating a patient with adult respiratory distress syndrome due to sepsis, during ongoing CO monitoring with NICO2, tidal volume would be decreased. In an actual example, Amato et al.  reported decreasing tidal volume from 661 ml to 362 ml and observed Paco2to increase from 38.1 mmHg to 58.2 mmHg. 7In instances of pneumothorax, asthma, pulmonary bleeding, endotracheal tube misplacement, and numerous other clinical situations, VV̇Eis likely to suddenly decrease.

Although it seems reasonable to reject a single outlier in scatter plots (figs. 1C and 2C of the article) 1for linear correlation and bias analysis, the statistical effects of disregarding this point are minor: correlation coefficient, 0.34 to 0.40; slope of linear regression, 0.70 to 0.68; bias, −1.73 to −1.88; and precision, 1.27 to 1.07. Ultimately, reducing n from 25 to 24 is useful because it enables us to observe a more consistent underestimation of CO after decreased minute ventilation.

The purpose of our study was to probe the limits of clinical usefulness of the NICO2system. We appreciate that because it is noninvasive, easy to use, and works well when VV̇Edoes not dramatically change, the system generally provides convenient and effective clinical monitoring. Without first evaluating the parameters within which a device provides useful data, however, it would not be prudent, in critical situations, to rely on information from any monitor. Unless clinicians are aware of the limits of the NICO2system, blind trust may result in unnecessary use of catecholamines and other less-than-optimal judgment. We thank the designers for providing a piece of equipment that makes it easier for physicians to provide attentive care to patients, and we welcome this opportunity to reiterate that it is crucially relevant for clinicians to be aware that the real-time monitoring accuracy of the NICO2system is affected by changes in VV̇E.

Tachibana K, Imanaka H, Takeuchi M, Takauchi Y, Miyano H, Nishimura M: Noninvasive cardiac output measurement using partial carbon dioxide rebreathing is less accurate at settings of reduced minute ventilation and when spontaneous breathing is present. A nesthesiology 2003; 98: 830–7
Tachibana K, Imanaka H, Miyano H, Takeuchi M, Kumon K, Nishimura M: Effect of ventilatory settings on accuracy of cardiac output measurement using partial CO2rebreathing. A nesthesiology 2002; 96: 96–102
Taskar V, John J, Larsson A, Wetterberg T, Jonson B: Dynamics of carbon dioxide elimination following ventilator resetting. Chest 1995; 108: 196–202
Haryadi DG, Orr JA, Kuck K, McJames S, Westenskow DR: Partial CO2rebreathing indirect Fick technique for non-invasive measurement of cardiac output. J Clin Monit 2000; 16: 361–74
Kuck K, Haryadi DG, Brewer LM, Orr JA: Mixed venous CO2does not need to remain constant during CO2rebreathing cardiac output measurements (abstract). A nesthesiology 2000; 93: A331
Yem JS, Tang Y, Turner MJ, Baker AB: Sources of error in noninvasive pulmonary blood flow measurements by partial rebreathing: A computer model study A nesthesiology 2003; 98: 881–7
Amato MBP, Barbas CSV, Medeiros DM, Magaldi RB, Schettino GPP, Lorenzi-Filho G, Kairalla RA, Deheinzelin D, Munoz C, Oliveira R, Takagaki TY, Carvalho CRR: Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998; 338: 347–54