Cardiac Output Determination Using A Widely Available Direct Continuous Oxygen Consumption Measuring Device: A Practical Way To Get Back To The Gold Standard (2025)

Abstract

Introduction

Cardiac output (CO) is an important parameter of cardiac performance, an accurate assessment of which is an important function of cardiac catheterization laboratories (CCL).(1) This is especially true in valvular disorders, e.g. in patients with equivocal echo findings where aortic valve area must be calculated for suspected aortic stenosis (AS). (2) A precise measurement of CO and consequently the degree of AS severity is crucial to ensure that only appropriate patients are referred for valve replacement and to prevent both premature or inappropriately delayed aortic valve replacement. (3)

A widely used CO method is the Fick equation, in which the total uptake or release of a substance, such as oxygen, by an organ is the product of the blood flow through the organ and the arteriovenous concentration difference of the substance. (4) When oxygen consumption is directly measured, this is called direct oxygen consumption (Direct, VO2) and is widely considered the gold standard for CO measurement. Alternatively CO can be calculated using estimated VO2.

Historically VO2 was directly measured using a Douglas bag over many minutes, a cumbersome affair. Therefore an estimation of VO2 is used instead in many CCL, with VO2 values typically estimated from tables or published predictive equations. (5-8) Reliability and use of predictive equations for CO measurement have been questioned in the CCL setting because of large discrepancies between measured and estimated values, (9-11) which of course have an influence on subsequent hemodynamic and valve area calculations. (12, 13)

Portable devices are available that measure breath-by-breath oxygen and carbon dioxide levels using a facemask and provide direct continuous VO2 measurements (Figure 1). These devices are well validated to produce very reliable VO2 assessment similar to these assessed by Douglas bag while being less cumbersome to patients and staff. (14) Importantly, VO2 can vary during the course of a catheterization. We hypothesized that measuring CO at a time when the VO2 was not only stable but precisely known (i.e., not an average over time) might give the most reliable calculations of aortic valvular dimensions. We investigated the correlation of the CO based on direct continuous VO2 measurement by this device to the more commonly used techniques of estimated VO2 and thermodilution CO. (15, 16).

Methods And Materials

This study was approved by the institutional review board (IRB) of Christiana Care Health System. Thirty patients scheduled for right and left heart catheterization were included in the study. Catheterization was performed in the usual fashion including measurement of right and left sided pressures. Intravenous midazolam and fentanyl was given to patients as needed. Patients were not placed on supplemental oxygen.

In order to measure VO2 directly and continuously, patients wore a fitted facemask with an airtight seal over the nose and mouth, which was connected to Ultima CardiO2 breathing analyzer (Medgraphics, St. Paul, Minnesota) The device allows comparison of inspired and expired air to determine VO2. To minimize the time of mask use, it was placed only after all catheters were in place for pressure measurements. After a baseline of at least 3 minutes to assure patient comfort and a steady-state VO2, immediate measurements of femoral artery and pulmonary artery saturations were obtained in triplicate, immediately followed by three thermodilution CO measurements. We could complete these measurements in approximately 90-120 seconds. This process was repeated again after 3 minutes after a new steady-state VO2 was confirmed in the same manner to get a second set of values. We waited until the continuous VO2 was relatively stable (Figure 1) before performing the CO determinations in triplicate. If the VO2 became variable subsequently, we waited for a return of a relatively stable VO2 before performing the repeat CO determinations (Figure 1). Oxygen consumption was measured continuously for these 2 minutes with the mean VO2 over that time period used for analysis.

The cardiac output can be assessed in the fick methods utilizing this equation:

$$\text{Cardiac Output}=\frac{\text{Oxygen Consumption}}{\text{A-VO2 Difference}}$$

A-VO2 difference is the difference in the arterial and venous oxygen blood content. Oxygen blood content can be assessed using this equation:

$$\text{Oxygen blood content}=\text{Oxygen saturation}\times 1.36\times \text{Hemoglobin (g/dl)}$$

While direct VO2 was assessed as described above, estimated fick calculation oxygen consumption was assessed using this equation:

$$\mathrm{O}2\text{consumption}=3\mathrm{ml}\mathrm{O}2\times \text{weight (Kg)}$$

All patients save one had suitable VO2 tracings for these measurements. This one patient had unreliable VO2 tracings due to Cheyne-Stokes respiration. The three values from each cardiac output method were averaged, and those values were compared. Another separate set of direct continuous VO2, assumed VO2 and TD CO was then repeated, after steady-state VO2 was confirmed, to be used as a validation of reproducibility of CO and AVA. The Phillips (XIMS, Philips Xper Connect, Melbourne, FL) system is in use at our CCL.

Results

The baseline characteristics of the 30 patients included in the study patients are displayed in table-1. Direct continuously measured VO2 CO and estimated VO2 CO calculations correlated poorly (Figure 2A, R= 0.57; ICC =0.59). Direct continuously measured VO2 CO and thermodilution CO calculations correlated poorly as well (Figure 2B, R= 0.51; ICC =0.60).

Using a constant assumed VO2 value led to a large discrepancy in repeated CO measurement during the case (Figure 3A, R= 0.62; ICC =0.57). Repeated TD CO was associated with less discrepancy (Figure 3B, R= 0.78; ICC =0.84). Although direct continuously measured VO2 changed at times during the course of a procedure, requiring a pause to await a relatively stable value in approximately 1/5 of patients, (although measurements were only made when a steady state was achieved), the resulting repeated CO measurements were closely correlated (Figure 3C, R=0.93; ICC=0.96) (Table 2).

Discussion

CO calculated by direct continuous VO2 measurement differs significantly from those based on estimated VO2 values and thermodilution. The superb reproducibility of CO using this direct continuous VO2 technique suggests this is the superior CO technique. Although direct VO2 based CO was felt to be the gold standard historically, cumbersome measurement techniques led to the adoption of the estimated VO2 determination.(9) However, many CCL computer systems, including our widely used software program (Phillips Xper, Melbourne FL) make an assumption of a patient's VO2 based on an equation developed by LaFarge and Miettinenin. (5) This equation was developed with regression analysis using patients aged 3-40 years with congenital heart disease and an average age of 12 years.(5) The authors themselves recommended against using this equation for patients older than 40 years of age.(5) Unfortunately using the assumed Fick calculation in patients older than 40 years of age, as were all our patients, can have a significant impact on the validity of the hemodynamic assessment. Even if the estimated VO2 had a better theoretical underpinning, the variations in direct continuously measured VO2 occasionally seen during catheterizations would make a single static estimation of VO2 inaccurate. The large variability in repeated calculated CO using a static value suggests that changes in actual VO2 during the procedure is a contributing factor. And, of course, any difference in the CO measurement will have a direct continuous linear effect on the valve area calculation, i.e., a 25% difference in CO measurement will alter the valve area by 25%. At least a 25 % variation in CO was seen in 56% of measurements with both assumed VO2 and TD when compared to direct continuous VO2. (11, 13) Similarly Chase et al showed that the rate of ≥25% error in the equations by LaFarge and Miettinen, Dehmer et al, and Bergstra et al occurred in 11%, 23%, and 45% of patients, respectively. Misclassification of cardiac index derived from each equation for 2 clinically important classifications: cardiogenic shock-21%, 23%, and 32% and hypoperfusion-16%, 16%, and 25%; respectively. (17) Recent studies have showed that the correlation of direct continuous VO2 and assumed VO2 derived cardiac output was not impressive even in children. In both of these studies the direct VO2 was assessed utilizing devices while children underwent general anesthesia, techniques not readily available to adult cardiologists. (18, 19)

Comparisons of CO determination using estimated and direct measured values have previously differed in patients undergoing cardiac catheterization.(9-13, 18-23) In a recent study, Narang et al compared direct VO2 CO calculated as an average over time using Douglas bags to different formulas of estimated VO2 CO. The direct and estimated VO2 CO values differed by >25% in 17% to 25% of patients depending on the formula used. (11) demonstrated Gertz et al, using a facemask device, recently showed that estimates of VO2 using of the LaFarge formula significantly underestimated CO and hence underestimated AVA, resulting in a high rate of misclassification of AS severity. (3) In their study the difference in calculated valve area in patients with aortic stenosis between using Lafarge table and measured VO2 was as much as 0.5cm². (3) In our study we found instances where the CO was both under- and over-estimated by the assumed Fick technique.

The other commonly used CO method is thermodilution. TD Technique can show variability in measurement and therefore is not reliable in the setting of many underlying cardiac abnormalities especially tricuspid regurgitation and atrial fibrillation. (24, 25) Despite these limitations, it is still probably more reliable than estimated Fick in those without significant tricuspid regurgitation and/or atrial fibrillation. (3) We found this technique to be reproducible but CO values did not correlate well with direct continuous VO2.

We used a device that allows an easy reproducible way to measure direct continuous VO2. The patient wears a full-face mask that is connected via tubing to the actual metabolic cart for collection of data. The patient does not need to be intubated or heavily sedated while undergoing the procedure. VO2 assessed using metabolic cart has been validated versus those assessed by Douglas bag and was found to give extremely high correlating values (R= 0.99). (14) Both the American Thoracic Society (ATS) and European Respiratory Society (ERS) approve the use of the device, and it is already used routinely during cardiopulmonary testing in many hospitals, as was ours. In our experience, the device is easy to use and the measurements did not cause a significant increase to the procedural time, since the VO2 measurements may be taken very quickly once all catheters are in place. The fact that all measurements of the CO equation (VO2 and blood saturations) can be done at exactly the same time may be important, as both of these parameters can change during a procedure (Figure 1B). The accuracy and ease of the use of this device, the availability of this device in most hospitals and the lack of need for general anesthesia and intubation make it practical to use especially when making decisions about advanced heart failure therapies or valvular surgery.

This study has several limitations. First, it included only a small number of patients. Second; we often used mild sedation, which may affect VO2 or depress cardiac output. Third, although a reflection of clinical practice, our study included patients with significant tricuspid regurgitation, which may be responsible for underestimation of CO by TD (24, 25). Finally, some patients may be unable to carry out this test if they develop discomfort from wearing the mask or develop anxiety from a sense of claustrophobia, (although all of our patients, were able to tolerate the mask for the duration of the procedure).

Conclusion

Cardiac output calculated from direct continuously measured VO2 varies substantially from those based on assumed VO2 values and thermodilution, which are the default methods used in most CCL. Direct continuously measured VO2-derived CO gives highly reproducible valve area measurements and may be more accurate.

Acknowledgements

This study was funded in part by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number U54-GM104941 (PI: Binder-Macleod).

Footnotes

References