Chest ACCP Member Benefits
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sun, X.-G.
Right arrow Articles by Wasserman, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sun, X.-G.
Right arrow Articles by Wasserman, K.
(Chest. 2000;118:631-640.)
© 2000 American College of Chest Physicians

Comparison of Exercise Cardiac Output by the Fick Principle Using Oxygen and Carbon Dioxide*

Xing-Guo Sun, MD; James E. Hansen, MD, FCCP; Hua Ting, MD; Ming-Lung Chuang, MD; William W. Stringer, MD, FCCP; David Adame, RCP, CRTT and Karlman Wasserman, MD, PhD, FCCP

Correspondence to: Karlman Wasserman, MD, PhD, FCCP, Division of Respiratory and Critical Care Physiology and Medicine, Department of Medicine, Harbor-UCLA Medical Center, PO Box 405, St. John’s Cardiovascular Research Center, 1000 W Carson St, Torrance, CA 90509-2910; e-mail: kwasserm{at}ucla.edu


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background and study objective: Theoretically, cardiac output (CO) calculated by the Fick principle should be the same using O2 (CO[O2]) or CO2 (CO[CO2]) as the test gas. However, agreement depends on the accuracy of gas exchange and blood gas measurements and the validity of the equations to convert measured variables into blood gas contents. Considering the widespread use of indirect estimates of pulmonary artery blood PCO2 and CO2 content to measure Fick principle CO during exercise, we wished to determine whether CO[O2] and CO[CO2] were equal during exercise and whether CO[CO2] could be accurately and precisely determined using direct measures of pulmonary artery blood.

Preparation and methods: Five healthy young nonsmoking volunteer men performed incremental exercise from rest to peak exercise on two separate occasions with intervening rest. Catheters were placed in brachial and pulmonary arteries to allow repeated blood sampling every minute during concurrent breath-by-breath gas exchange measurements from rest to peak exercise. CO[O2] was compared with CO[CO2] at multiple levels of exercise. Using standard equations, arterial and mixed venous O2 contents were calculated from hemoglobin concentration (Hb), oxyhemoglobin saturation (SO2), and PO2, whereas CO2 contents were calculated from PCO2, pH, Hb, and SO2. Blood gas analyzers were used for measurement of pH, PCO2, and PO2, and a co-oximeter was used for measurement of Hb and SO2. Initial calculations suggested that exercise CO[CO2] was 14% higher than CO[O2] and helped disclose small systematic measurement errors in PCO2 for values > 45 mm Hg detected by proficiency testing surveys and documented with blood tonometry in the blood gas analyzer.

Results: After correcting PCO2 for the small systematic measurement error found, the measures and equations used to calculate arterial and mixed venous O2 and CO2 contents were adequate to provide mean CO values that are reasonably similar. However CO[CO2] values were more than twice as variable as CO[O2].

Conclusions: The increased variability of Fick principle CO[CO2] compared with CO[O2] is attributable to the much lower extraction ratio for CO2 and the greater complexity in calculation of blood CO2 than O2 contents. These results raise concerns about the accuracy and precision of estimating CO and stroke volume using CO2 as a test gas, even with direct measurement of blood CO2 contents in normal subjects.

Key Words: blood CO2 content • cardiac output • exercise • Fick principle • PCO2 • proficiency testing • tonometered blood


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
In 1870, Fick proposed a method to measure cardiac output (CO) based on the principle of conservation of mass,1 now known as the Fick principle. Using O2 as the test gas, it is the reference method or "gold standard" for CO measurements (CO[O2]) to which all other methods are referred. Advances in technology allow the major variables in this method to be measured intermittently from on-line measurement of O2 uptake (O2) together with arterial (SaO2) and mixed venous oxyhemoglobin saturation (SO2).2 3 4 5 6

Fick stated that the "corresponding calculation with CO2 quantities gives a determination of the same value, which provides a control for the other calculation."1 As Butler7 pointed out several decades ago, methods using CO2 became more numerous than those using O2 because of the apparent simplicity of dealing with CO2, a gas that is virtually absent from inspired air, has high solubility, is relatively unaffected by nonuniformity of alveolar CO2 in the lung, and has a practically straight slope to its dissociation curve over the physiologic range. However, the direct Fick principle for CO using CO2 as the test gas (CO[CO2]) has rarely been reported and apparently not actually compared with CO[O2]. Conceivable differences between CO[O2] and CO[CO2] might occur because of errors of measurements or incorrect equations for calculating gas contents.8 9 Our objective was to compare CO[O2] and CO[CO2] at rest and during progressive ramp pattern cycle ergometer exercise while continuously measuring O2 and CO2 output (CO2), breath-by-breath, and intermittently sampling arterial and mixed venous blood measurement of PCO2, PO2, pH, hemoglobin concentration (Hb), and oxyhemoglobin saturation (SO2) during exercise. The degree of agreement could be used to determine the reliability of measurements and equations for calculating arterial and mixed venous contents of the two gases over a wide range of partial pressures, pH, Hb, and SO2.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Subjects and Protocol
Subjects and Preliminary Ramp Exercise Testing:
The Human Subjects Committee at Harbor-UCLA Medical Center approved the research protocol. After informed consent, five healthy nonsmoking male subjects performed a preliminary noninvasive increasing work rate exercise test on an electromagnetically braked cycle ergometer (type 18070; Gould-Godart; Bilthoven, Netherlands) to determine their maximum work rate, anaerobic threshold (AT), and maximal O2 (O2max). During each test, a pedal frequency of 60 revolutions/min was maintained with the aid of a visual pedal rate indicator.

Catheter Placement:
A flow-directed pulmonary artery catheter (Arrow International; Reading, PA) was introduced via the sheath (Cordis Corporation; Miami, FL), which was inserted percutaneously into the right femoral vein, and positioned in the main pulmonary artery under direct fluoroscopic guidance. Another arterial catheter was placed percutaneously into the left brachial artery. Both catheters were attached to infusion devices (Continu-Flo; Baxter Health Care; Deerfield, IL), which provided a slow continuous flow (15 mL/h) of heparinized normal saline solution (1,000 U/L) as well as periodic bolus flushing of the catheter.

Exercise Protocols:
Two ramp exercise tests (tests A and B) were performed to exhaustion by each of the five subjects. The rate of work rate increase (range, 25 to 40 W/min) was chosen, depending on fitness, so that the subject would exhaust in about 10 min during which work rate was progressively increased. Gas exchange and heart rate (HR) measurements were made breath-by-breath. Thirty-second averages were calculated during 3 min of rest, 3 min of unloaded pedaling, and during the progressively increasing work rate test. At least 1 h of rest separated the two tests.

Blood Samples:
After clearing approximately three catheter dead space volumes, blood was sampled simultaneously from the pulmonary artery and brachial artery during rest, unloaded cycling, and during the last half of each minute of increasing work rate exercise. Blood gas samples were drawn for 15 to 20 s in the midportion of the half-minute period matching the appropriate gas exchange measures. This minimized fluctuations in gas exchange measures related to breath-by-breath and beat-by-beat volume changes as well as fluctuations in PCO2, PO2, and pH related to the ventilatory cycle. The samples were collected in 3-mL glass syringes, which contained a small amount (mean, 0.14 mL) of liquid heparin (1,000 U/mL) to prevent clotting before analysis.

Measurements
Blood Analysis:
The blood samples were agitated and immediately chilled in ice slurry. Blood gas analysis was performed with an Instrumentation Laboratory (IL) 1306 blood gas analyzer (Instrumentation Laboratory; Lexington, MA) for pH, PCO2, and PO2, and with an IL 482 co-oximeter (Instrumentation Laboratory) for total Hb and SO2. The measured values obtained were corrected for heparin dilution. The analyzer precision was verified with quality control materials every 20 to 30 min.

Respired Gas Analysis:
The subjects respired through a mouthpiece during each exercise period. Expired air was directed to a Fleisch-type No. 3 pneumotachograph via a two-way breathing valve (100 mL dead space). Respired O2, CO2, and N2 partial pressures at the mouthpiece were continuously measured by mass spectrometry (MGA-1100; Perkin Elmer; Pomona, CA). Minute ventilation (E; body temperature pressure, saturated) and O2 and CO2 (both standard temperature and pressure, dry) were calculated as previously reported.10 The AT was determined from a plot of CO2 vs O2 (V-slope plot) as described by Beaver et al.11

Calculations
Calculations of Blood O2 and CO2 Contents:
Blood O2 contents (CbO2) in milliliters per deciliter were calculated from the following equation:2 3 12 13 14 15 16

(1)
where Hb is hemoglobin concentration in grams per deciliter, SO2 is O2 saturation in percent, and PO2 is O2 partial pressure in millimeters of mercury.

Plasma CO2 contents (CplCO2) were calculated from the standard formula derived from the Henderson–Hasselbach equation17 18 19 20 :

(2)
where 2.226 is the conversion factor for CO2 (standard temperature and pressure, dry) in millimoles per liter to milliliters per deciliter, s is the plasma solubility coefficient in millimoles per liter per millimeter of mercury of CO2, and pK' is the apparent pK of the CO2–bicarbonate system. The variables s and pK' are 0.0307 and 6.0907 in plasma at 37°C and pH 7.4 in normal human subjects.21 Because they are dependent on conditions of temperature (T) and pH, values appropriate to the relevant blood conditions were first determined from the following equations17 21 22 :

(3)

(4)

Total blood CO2 contents (CbCO2) in milliliters per deciliter were calculated from the equation of Douglas et al,17 which takes into account the effects on s and pK' of changes in T and pH during exercise21 22 :

(5)
From the above equations, the arterial and mixed venous O2 contents (CaO2 and CO2) and CO2 contents (CaCO2 and CCO2) were calculated to determine the arterial–mixed venous O2 content difference [C(a-)O2] and mixed venous–arterial CO2 content difference [C(-a)CO2], respectively.

Calculations of CO and Stroke Volume:
From blood and gas exchange values, standard Fick equations and HR were used to calculate CO[CO2], CO[O2], and stroke volume (SV).

Initial Comparison of CO[O2] with CO[CO2]
At rest, the ratio of CO[CO2]/CO[O2] was nearly similar, at 1.02 ± 0.05. In contrast, at all levels of exercise, the ratio of CO[CO2]/CO[O2] was 1.14 ± 0.13, which was significantly different from that at rest (Fig 1 ). Because we could not ascertain a physiologic reason for this change from a 2% difference to a 14% difference, we carefully reexamined our measurements, equations, and calculations.



View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. CO differences during exercise uncorrected for PCO2 analyzer bias. The CO[CO2] values are calculated from the directly measured data from the IL-1306 blood gas analyzer. The CO[CO2] is plotted against CO[O2] (top) and percent deviation of CO[CO2] from CO[O2] is plotted against CO[O2] (bottom) during ramp exercise. The solid lines are the regression lines. The dotted line in the upper panel is the slope of 1.00 and in the lower panel is zero deviation. The CO[CO2] is larger than CO[O2] with a slope of 1.14 (p < 0.0001). The lower panel shows that the average deviation of the CO[CO2] from the CO[O2] during exercise is 14%.

 
Correction of Measured PCO2
Reanalysis of PCO2 Proficiency Testing Data:
We considered data previously reported9 for 30 lots of perfluorocarbon proficiency testing material on a nationwide basis with lot mean values ranging from 22 to 72 mm Hg. The raw data from the > 900 instruments, regardless of the number of instruments per each model, were used. The all instrument means (AIM) from all 900 instruments and model means of three IL models (IL-1306, IL-1420, and IL-1620), each of which included 40 to 100 instruments per model, were used. As is apparent from Figure 2 , there are significant differences in measurement of perfluorocarbon proficiency testing material among instrument models at the lower and higher PCO2 levels (by analysis of variance [ANOVA]; p < 0.001). There were no significant differences among the three IL models from the AIM PCO2 values of 35 to 45 mm Hg (p > 0.05). For the IL-1306 model, comparing partial ranges of AIM PCO2 < 35 mm Hg (eight lots of values) and > 45 mm Hg (19 lots of values) separately, we obtained the following two different equations using measured PCO2 as the independent variable. For PCO2 < 35 mm Hg, the equation is:

with R = 0.9999; SD = 0.05; and p < 0.0001. For PCO2 > 45 mm Hg, the equation is:

with R = 0.999; SD = 0.45; and p < 0.0001.



View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Deviation of PCO2 using proficiency testing material for different models of blood gas analyzers. The deviations from AIM of PCO2 measurements for 30 lots of perfluorocarbon proficiency testing material for three IL models are plotted against the AIM PCO2 values ranging from 22 to 72 mm Hg. The figure has been modified from Figure 5 of a previous publication.9 There are no significant differences among the three IL models from the AIM PCO2 values of 35 to 45 mm Hg (p > 0.05). The IL-1420 values do not deviate significantly from the AIM over the full range of AIM PCO2 values (p > 0.05). The IL-1620 values are significantly higher than the AIM only at AIM PCO2 values > 45 mm Hg (p < 0.001). However, the IL-1306 values (the model used in this study) are significantly higher than the AIM at values < 35 mm Hg and significantly lower than the AIM at values > 45 mm Hg. Using the appropriate equations, the IL-1306 model overestimates the PCO2 value by 2.06 mm Hg at a PCO2 of 21 mm Hg and underestimates the PCO2 value by 2.47 mm Hg at a PCO2 of 70 mm Hg. Thus, high PCO2 values (found only in mixed venous blood in our study) using the IL-1306 are lower than those that would have been measured and reported using the IL-1420 model or the AIM of all blood gas analyzer models.

 
Using these two equations, we found that the IL-1306 model overestimated the PCO2 by 2.06 mm Hg at a PCO2 of 21 mm Hg and underestimated the PCO2 by 2.47 mm Hg at PCO2 of 70 mm Hg.

Blood Tonometry:
To ascertain whether our IL-1306 instruments in our laboratory had similar systematic deviations in measured PCO2, tonometered blood values were compared with target blood values over several years including the period for the data in this report. Three instruments of IL-1306 model and four instruments of IL-1600 model series were used in our laboratory. Fresh blood from the same person was tonometered for 45 min using the EQUILibrator (QC 433) 300 model tonometer (RNA Medical; Acton, MA) to three PCO2 targeted levels: low (mean, approximately 21 mm Hg; range, 21 to 22 mm Hg), normal (mean, approximately 40 mm Hg; range, 37 to 44 mm Hg), and high (mean, approximately 70 mm Hg; range, 68 to 74 mm Hg). The blood samples were then promptly measured. On each day of measurement, targeted values were calculated from primary standard partial pressures of gases in the tanks and measured barometric pressure (NOVA 469 model; Princo Instruments; Southampton, PA). Approximately 100 (89 to 166) measurements of PCO2 were made at each level for each model over several years. Deviations of measured PCO2 from the targeted values were calculated and plotted against the targeted PCO2. At the targeted PCO2 values of approximately 40 mm Hg, the IL-1306 and IL-1600 series blood gas analyzers in our laboratory precisely and accurately measured PCO2 in tonometered blood. However, measurement accuracy at both lower and higher targeted levels were dependent on the model used (p < 0.001). The IL-1306 model overestimated the PCO2 by about 2.07 mm Hg at average PCO2 values of 21 mm Hg and underestimated the PCO2 by 2.44 mm Hg at average PCO2 values of 70 mm Hg. For the IL-1306 model PCO2, measurement errors from tonometered values at both lower and higher PCO2 level were nearly identical to the same model deviations from AIM (2.06 vs 2.07 and 2.47 vs 2.44) found for the perfluorocarbon proficiency testing material.

Correction of Measured PCO2 Values:
Because the measurement errors of PCO2 with IL-1306 model at three levels in tonometered blood were nearly identical to the deviation from AIM in perfluorocarbon proficiency testing, the corrected PCO2 [PCO2(cor.)] values were calculated from measured PCO2 values > 45 mm Hg, using the following equation:

(6)
Measured PCO2 values in the 35 to 45 mm Hg range were not corrected. There were no arterial or mixed venous PCO2 values < 35 mm Hg.

These corrected values were then used for final calculations of plasma and blood CO2 contents, CO, and related values.

Data Analysis and Statistics
Data from the first and second test from each subject were treated separately. The CO[CO2] vs CO[O2] and the percent deviations vs CO[O2] at rest and during exercise were calculated and plotted separately.2 The deviations vs average of CO[O2] and CO[CO2] at rest and during exercise were calculated before and after correction of PCO2 separately.23 Additionally, the ratios of SV each minute to mean exercise SV of each study using CO2 or O2 were calculated and plotted. The statistical analyses of data were performed using the repeated-measures ANOVA or the paired Student’s t test. Relationships among variables were studied using linear regression techniques and by calculating Pearson product-moment correlation coefficients. A p < 0.05 was considered significant. All data are expressed as mean ± SD, unless otherwise specified.


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Exercise Studies and Blood Values
The subject’s physical characteristics and the work rate at AT and maximal exercise of test A and test B are shown in Table 1 . There was no significant difference between the two tests. O2 increased at a rate of 10.30 ± 0.70 mL/min/W during exercise.


View this table:
[in this window]
[in a new window]

  		Table 1. Subjects’ Physical Characteristics, and Work Rate at AT and Maximal Exercise*

 
Because the duration of exercise varied slightly between subjects, blood values were grouped according to exercise intensity (Table 2 ). Both arterial and venous Hb significantly increased from resting values during exercise from AT to maximal exercise compared with rest values (p < 0.05). SaO2 decreased slightly at maximal exercise (p < 0.05), but PaO2 did not change significantly during exercise. Both SO2 and PO2 progressively decreased from rest to maximal exercise (p < 0.001). PaCO2 increased from rest to AT (p < 0.05), and then decreased from AT to maximal exercise (p < 0.001) to a value significantly lower than that at rest (p < 0.01). In contrast, PCO2 progressively increased, and arterial and venous pH decreased from rest to maximal exercise (p < 0.05).


View this table:
[in this window]
[in a new window]

  		Table 2. Selected Blood Measurements at Rest and During Exercise*

 
Above the AT, CaO2 increased primarily because of increase in Hb (p < 0.001), and CaCO2 decreased because of bicarbonate buffering of lactic acid (p < 0.001; Tables 2 , 3 ). The CO2 progressively decreased from rest to maximal exercise. In contrast, CCO2 increased from rest to the second minute above AT work rate and then significantly decreased until maximal exercise.


View this table:
[in this window]
[in a new window]

  		Table 3. Comparison of CO Measured From O2 and CO2 Contents*

 
Comparison of CO[O2] With CO[CO2]
Using O2 and CO2 values and arterial and mixed venous blood O2 and CO2 contents, CO[CO2] and CO[O2] were independently calculated. The ratios of CO[CO2]/CO[O2] at rest and all levels of exercise were similar (0.99 ± 0.04 to 1.00 ± 0.13, respectively; Table 3 ). The average percent differences of these two methods for CO calculation were approximately zero (Fig 3 ). Using the Bland–Altman method,23 the mean deviations in exercise CO[CO2] from CO[O2] were 0.01 ± 2.43 L/min (contrasted with 2.46 ± 2.98 L/min before correction of PCO2).



View larger version (28K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. CO during exercise measured with O2 and CO2 as the test gases. The CO[CO2] is plotted against CO[O2] (top) and percent deviation of CO[CO2] from CO[O2] is plotted against CO[O2] (bottom) at rest (open squares) and during unloaded pedaling and ramp exercise (solid squares). In the upper panel, the solid line is the least squares regression line through all the data with a slope of 1.00 and a small intercept (0.14). In the lower panel, the dotted line is zero deviation. On average, the CO[CO2] is similar to the CO[O2] at all levels of exercise.

 
Variability of Exercise SV Using CO[O2] and CO[CO2]
The individual’s minute-by-minute SV and mean SV during exercise were separately calculated and graphed from CO[O2] and CO[CO2] (Fig 4 ). The average of the ratios of individual to mean SV for each test gas was necessarily 1.0. Using CO[O2], the SD of the SV ratio was ± 8% and did not change with increasing work rate. Using CO[CO2], the SD of the SV was ± 19%, although SV increased slightly at higher work rates (p < 0.05). The variability in SV, when calculated with CO2, was significantly larger than that when calculated with O2 (p < 0.01).



View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Comparison of variability of SV using CO2 and O2 as test gases. Variability of SV calculated from CO2 (top) and O2 (bottom) is related to work rate. Variability is expressed as the ratio of each SV to the mean SV for the entire exercise period after unloaded pedaling. Thin lines connect the values for each of the 10 studies. The thick solid lines are the regression lines. The dotted lines have a zero slope, ie, each SV is equal to the mean SV. In the upper panel, using CO2 as the test gas, the ratio of SV to mean SV calculated from CO2 increases slightly with work rate with a slope of 0.1%/W (p = 0.04) with an SD around the mean of 19%. In the lower panel, using O2 as the test gas, the ratio of SV to the mean SV does not differ significantly as work rate increases (p = 0.16) with an SD around the mean of 8%. The deviations using CO2 are significantly larger than those using O2 (p < 0.01).

 

   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Comparison of CO[O2] With CO[CO2]
Confirming the Fick principle,1 the ratio of the mean values of CO[CO2] was similar to that of CO[O2], both at rest (0.99 ± 0.04) and during exercise (1.00 ± 0.13), when accurate measures and correct formulas and calculations were used (Fig 3 and Table 3 ). The overall difference of these two methods for CO calculation was negligible. Inasmuch as CO2 is a simpler measurement than O2, it would be tempting to suggest that CO2 might be the more optimal test gas for CO measurements. However, the comparison of SV values during exercise indicates that measurements using O2 are much less variable than those using CO2 as the test gas (Fig 4) . In this and other studies during incremental exercise, the SV increases quickly (in about two thirds of subjects during the first minute of ramp exercise)2 24 and reaches relatively stable peak levels before O2 reaches 30 to 40% of its peak. Thus, the stability of SV during exercise in normal subjects can be used to compare the reproducibility of CO measurements and the reliability of formulas and calculations using the two test gases. In the 82 paired measures of exercise SV, the variability of the minute-by-minute SV from the mean exercise SV using O2 is significantly less than that using CO2 (SD = 8% vs 19%). Because at any one point SV times the same HR equals the O2 or CO2 at the mouth divided by the arteriovenous or venoarterial content differences in the test gas, the higher variability in measuring SV using CO2 as a test gas strongly suggests that it is more difficult to accurately and precisely quantify C(-a)CO2 than C(a-)O2. As can be inferred from Table 4 , the much lower extraction ratio for CO2 than O2 during both rest and all stages of exercise would require more accuracy and precision in blood CO2 than in blood O2 measurement to find equal variability.


View this table:
[in this window]
[in a new window]

  		Table 4. Comparison of 1% Measurement Errors in Mixed Venous O2 and CO2 Contents on Calculations of Arterial–Mixed Venous Differences*

 
Potential Sources of Errors
Differences of CO[CO2] and CO[O2] During Exercise Before PCO2 Corrections:
Initially (Fig 1) , we were surprised to find such large differences in the calculations of CO during exercise using O2 and CO2. Before PCO2 corrections, CO[CO2] and CO[O2] were similar at rest, but CO[CO2] measurements were on average 14% greater than CO[O2] during exercise. These large and consistent differences during exercise could not be explained on physiologic grounds, such as shunting, lung metabolism, Hb, velocity differences of RBCs and plasma, and gas stores.6 25 26 27 28

Therefore, we reviewed our laboratory experience with our blood gas analyzer models. Using tonometered blood as the standard, our IL-1306 models accurately measured the PCO2 at approximately 40 mm Hg, but overestimated the PCO2 by 2.07 mm Hg at average PCO2 target values of 21 mm Hg and underestimated the PCO2 by 2.44 mm Hg at average PCO2 targeted values of 70 mm Hg. We did not have tonometered blood data at other levels of PCO2 (eg, 23 to 36 mm Hg and 45 to 67 mm Hg) for further direct comparisons. Therefore, to further refine the PCO2 measurement error over the range of PCO2 found, we reviewed the relationship between several analyzer models and the AIM for a perfluorocarbon proficiency testing material as noted in Figure 2 .9

There were no PCO2 measurements > 72 mm Hg (the highest values for proficiency testing material or tonometered blood), except for samples obtained at maximal exercise. We found that the differences of the proficiency testing material from the AIM at low, medium, and high levels of PCO2 reported for the IL-1306 were nearly identical to the PCO2 errors of our IL-1306 and IL-1600 models measuring tonometered blood. Such consistent agreement is unlikely to be fortuitous. First, this finding allowed us to correct the IL-1306 model PCO2 measurements at values > 45 mm Hg to accurate values by assuming that the correction remained linear. Secondly, this finding reinforced the value of proficiency testing in quantifying model biases for PCO2 and confirms the previous suggestion that it is unwise to shift among models or manufacturers without first ascertaining their comparability.9 We suggest that other investigators consider correcting their specific model blood gas data to AIM or tonometered blood values when measuring PCO2 outside the range of 35 to 45 mm Hg.

Correction of PCO2 and CO2 Content Values:
The initial large and consistent differences in CO using O2 and CO2 before PCO2 correction disappeared when we corrected PCO2 with equation 6 . Thus, it is apparent that the main reason for the initial larger exercise CO[CO2] measurements than CO[O2] (Fig 1) was the error in PCO2 measurement at values > 45 mm Hg. Although the difference between PCO2 values of 57.4 vs 58.6 mm Hg or 76 vs 78 mm Hg may not be important in a clinical setting, there are two reasons these differences become crucial when such PCO2 values are converted to blood CO2 content for calculation of CO First, relatively small errors in PCO2 values in mixed venous blood lead to large errors when plasma and blood CO2 content are calculated (Table 5 ). Second, because the CO2 contents of blood are larger than the O2 contents of blood, the extraction ratios for CO2 by the lung are smaller than the extraction ratios for O2 by the tissues of the body (Table 4) . As shown in Table 4 , small measurement errors in CO2 content have a larger effect on CO than errors of the same magnitude in O2. Therefore, smaller errors in measurement of PCO2 and calculation of CCO2 could cause relatively larger errors in the calculation of C(-a)CO2, CO[CO2], and SV[CO2] (Table 5) . Furthermore, fewer factors need to be considered to calculate blood O2 content as compared with blood CO2 content, because almost all the O2 is in the RBC compartment whereas the CO2 is in both plasma and RBC compartments.


View this table:
[in this window]
[in a new window]

  		Table 5. Importance of Correction of PCO2 Measurement Error for Calculations of Arterial–Mixed Venous Difference and CO*

 
Temperature Changes During Exercise: In situations in which either in vivo or in vitro blood temperature is significantly different from 37°C, equation 5 (in common with McHardy’s original equation19 ) may not be valid, because the partitioning of CO2 carriage between plasma and erythrocytes may be temperature dependent.17 We did not measure the rectal, esophageal, or intravascular temperatures during exercise. Considering other publications,29 30 31 32 33 34 the peak rectal, arterial blood, and esophageal temperature would increase, on average, about 0.2 to 0.8°C from rest to 15 min of submaximal exercise. It is likely that the average increase of peak temperature would not be any greater during the 10 min of incremental exercise in our subjects. We assumed that the temperature at rest or during exercise was approximately 37°C. However, if blood temperature at the end of incremental exercise increased 0.5°C, using the proper formulas to correct s and pK' for this temperature change,17 21 22 the mean CCO2, CaCO2, C(-a)CO2, CO[CO2], and SV[CO2] would change < 1% from the values reported herein at maximal exercise. The changes at mild, moderate, and heavy exercise would necessarily be less. These changes from our reported values are not used in the Figures or Tables because we did not have actual temperature measurements and because such small changes would not influence our data interpretations.

Other Potential Errors of Blood CO2 Content Values:
If we had used the McHardy–Visser equation,19 35 to replace equation 5 for calculation of CbCO2 value, the CO[CO2] measurements would have increased only 1 to 3% at all levels of exercise. Consistent underestimation or overestimation of pH both in arterial and mixed venous blood would cause trivial (1 to 2%) differences in the calculation of C(-a)CO2, CO[CO2], and CO[CO2]/CO[O2] during exercise.

Potential Errors of Blood O2 Content Values:
It is reassuring that SO2 measured by co-oximetry was consistently about 1% lower than SO2 calculated from electrode measurements of PO2 and pH. This 1% difference can be attributed to the small amounts of methemoglobin and carboxyhemoglobin normally present and did not appreciably affect the calculation of CO[O2], because methemoglobin and carboxyhemoglobin are found both in arterial and venous blood.20 Because dissolved O2 is trivial in equation 1 , small measurement errors of PO2 would have a very small effect on the calculation of CO[O2]. Therefore the differences in the ratio of CO[CO2]/CO[O2] during exercise could not be explained by the measurement errors of PO2, Hb, and SO2, inasmuch as such errors could only result in a 1 to 3% difference in calculations of O2 content and CO[O2].

The constant 1.34 in equation 1 is the one most commonly used by physiologists and clinical physicians for calculating O2 content. However, some investigators have also used 1.306, 1.36, or 1.39 mL/g Hb for O2 capacity.36 37 38 39 40 41 If we had used any one of these other constants to replace 1.34 in equation 1 , the CO[O2] measures would have changed by 1 to 4% at rest and all levels of exercise. Thus, varying the value for O2 capacity could not explain the differences between CO[CO2] and CO[O2] during exercise noted in Figure 1 .

Potential Errors of O2, CO2, and Matching Samples:
Analytic errors in respiratory gas exchange measurement have been studied and are acceptable for research and clinical purposes with a modern rapid breath-by-breath analyzer in a well-calibrated system.2 6 10 11 25 26 27 The error range of O2 is ± 1.0 to 2.6% and that of CO2 is ± 0.9 to 2.6%.42 43 44 45 The change in O2 to change in work rate relationship of 10.3 mL/min/W found in this study is similar to that previously reported, and also supports the accuracy of the gas exchange measurements.26 46 47

Blood gas samples were drawn for 15 to 20 s in the midportion of the 30-s periods to match the concurrent gas exchange measurements. This duration minimized fluctuations in gas exchange measurements related to breath-by-breath and beat-by-beat volume changes as well as fluctuations in PO2, PCO2, and pH related to the ventilatory cycle. The measurements of C(-a)CO2, C(a-)O2, CO2, O2, and HR were all concurrent. As previously found, matching of blood and gas exchange values is valid during ramp, unsteady-state exercise.48 Also, despite a constant work rate, heavy-intensity exercise is not a steady state.49

Significance
It is evident that Fick principle CO and SV values can be determined in normal subjects using either O2 or CO2 as the test gas. However, using current techniques, the variability of CO and SV based on CO2 is two to three times greater than that based on O2 because of the difficulty in quantifying the venous–arterial differences in blood CO2 contents, even when the mixed venous blood is directly sampled and the accuracy of the PCO2 measurement has been validated with tonometered blood. These findings raise concerns about the validity of estimating CO and SV using CO2 as the test gas.


   Footnotes
 
Abbreviations: AIM = all instrument means; ANOVA = analysis of variance; AT = anaerobic threshold; CaCO2 = arterial CO2 content; CaO2 = arterial O2 content; C(a-)O2 = arterial–mixed venous O2 content difference; CbCO2 = total blood CO2 content; CbO2 = blood O2 content; CO = cardiac output; CO[CO2] = CO using CO2 as the test gas; CO[O2] = CO using O2 as the test gas; CplCO2 = plasma CO2 content; C(-a)CO2 = mixed venous–arterial CO2 content difference; CCO2 = mixed venous CO2 content; CO2 = mixed venous O2 content; Hb = hemoglobin concentration; HR = heart rate; IL = Instrumentation Laboratory; PCO2[cor.] = corrected PCO2; PCO2 = mixed venous partial pressure of CO2; PO2 = mixed venous partial pressure of O2; SaO2 = arterial oxyhemoglobin saturation; SO2 = oxyhemoglobin saturation; SV = stroke volume; SV[CO2] = stroke volume using CO2 as the test gas; SV[O2] = stroke volume using O2 as the test gas; SO2 = mixed venous oxyhemoglobin saturation; CO2 = CO2 output; E = minute ventilation; O2 = O2 uptake; O2max = maximal O2 uptake

{altfoot}*From the Division of Respiratory and Critical Care Physiology and Medicine, Department of Medicine, Harbor-UCLA Medical Center, St. John’s Cardiovascular Research Center, Torrance, CA.

Supported in part by the Milly Liang Liu, MD, and Steve C. K. Liu, MD, Research Fund.

Received for publication November 30, 1999. Accepted for publication April 6, 2000.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Fick, A (1870) Uber die Blutquantums in den Herzventrikeln. Sitzung Phys med Gesell Würzburg July,16
  2. Stringer, WW, Hansen, JE, Wasserman, K (1997) Cardiac output estimated noninvasively from oxygen uptake during exercise. J Appl Physiol 82,908-912[Abstract/Free Full Text]
  3. Davies, G, Hess, D, Jebson, P (1987) Continuous Fick cardiac output compared to continuous pulmonary artery electromagnetic flow measurement in pigs. Anesthesiology 66,805-809[ISI][Medline]
  4. Doi, M, Morita, K, Ikeda, K (1990) Frequently repeated Fick cardiac output measurements during anesthesia. J Clin Monit 6,107-112[ISI][Medline]
  5. Quinn, TJ, Weissman, C, Kemper, M (1991) Continual trending of Fick variables in the critically ill patient. Chest 99,703-707[Abstract/Free Full Text]
  6. Keinanen, O, Takala, J, Kari, A (1992) Continuous measurement of cardiac output by the Fick principle: clinical validation in the intensive care. Crit Care Med 20,360-365[ISI][Medline]
  7. Butler, J (1965) Measurement of cardiac output using soluble gases. Fenn, WO Rahn, H eds. Handbook of physiology, section 3: Respiration ,1489-1503 American Physiological Society Washington, DC.
  8. Mahutte, CK, Jaffe, MB (1995) Effect of measurement errors on cardiac output calculated with and modified Fick methods. J Clin Monit 11,99-108[CrossRef][ISI][Medline]
  9. Hansen, JE, Casaburi, R (1998) Patterns of dissimilarities among instrument models in measuring PO2, PCO2, and pH in blood gas laboratories. Chest 113,780-787[Abstract/Free Full Text]
  10. Sue, DY, Hansen, JE, Wasserman, K (1980) Measurement and analysis of gas exchange during exercise using a programmable calculator. J Appl Physiol 49,456-461[Abstract/Free Full Text]
  11. Beaver, WL, Wasserman, K, Whipp, BJ (1986) A new method for detecting the anaerobic threshold by gas exchange. J Appl Physiol 60,2020-2027[Abstract/Free Full Text]
  12. Hüfner G. Neue versuche zur bestimung der saueratoffkapazität des blutfarbstoffes. Arch Anat Physiol 1894; 130–176
  13. Davies, GG, Jebson, PJR, Glasgow, BM, et al (1986) Continuous Fick cardiac output compared to thermodilution cardiac output. Crit Care Med 14,881-885[ISI][Medline]
  14. Foss, ML, Steven, JK (1998) Fox’s physiological basis for exercise and sport 6th ed. ,202-206 WCB/McGraw-Hill Boston, MA.
  15. Gonzalez, NC, Clancy, RL, Moue, Y, et al (1998) Increasing maximal heart rate increases maximal O2 uptake in rats acclimatized to simulated altitude. J Appl Physiol 84,164-168[Abstract/Free Full Text]
  16. Teboul, JL, Mercat, A, Lenique, F, et al (1998) Value of the venous-arterial PCO2 gradient to reflect the oxygen supply to demand in humans: effects of dobutamine. Crit Care Med 26,1007-1010[CrossRef][ISI][Medline]
  17. Douglas, AR, Jones, NL, Reed, JW (1988) Calculation of whole blood CO2 content. J Appl Physiol 65,473-477[Abstract/Free Full Text]
  18. Hachamovitch, R, Brown, HV, Rubin, SA (1991) Respiratory and circulatory analysis of CO2 output during exercise in chronic heart failure. Circulation 84,605-612[Abstract/Free Full Text]
  19. McHardy, GJR (1967) The relationship between the differences in pressure and content of carbon dioxide in arterial and venous blood. Clin Sci 32,299-309[ISI][Medline]
  20. Nunn, JF (1997) Nunn’s applied respiratory physiology 4th ed. Butterworth-Heinemann Oxford, UK.
  21. Austin, WH, Lacombe, E, Rand, PW, et al (1963) Solubility of CO2 in serum from 15 to 38°C. J Appl Physiol 18,301-304[Abstract/Free Full Text]
  22. Kelman, RG (1967) Digital computer procedure for the conversion of PCO2 into blood content. Respir Physiol 3,111-115[CrossRef][ISI][Medline]
  23. Bland, JM, Altman, DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1,307-310[CrossRef][ISI][Medline]
  24. Åstrand, PO, Cuddy, TE, Saltin, B, et al (1964) Cardiac output during submaximal and maximal work. J Appl Physiol 19,268-274[Abstract/Free Full Text]
  25. Habler, OP, Messmer, KFW (1997) The physiology of oxygen transport. Transfus Sci 18,425-435[CrossRef][ISI][Medline]
  26. Wasserman, K, Hansen, JE, Sue, DY, et al (1999) Principles of the exercise testing and interpretation 3rd ed. Lippincott Williams & Wilkins Baltimore, MD.
  27. Chuang, ML, Ting, H, Sun, XG, et al (1999) Aerobically generated CO2 stored during early exercise. J Appl Physiol 87,1048-1058[Abstract/Free Full Text]
  28. Rapaport, E, Kuida, H, Haynes, FW, et al (1956) Pulmonary red cell and plasma volumes and pulmonary hematocrit in the normal dog. Am J Physiol 185,127-132[Abstract/Free Full Text]
  29. Nielsen, B, Nielsen, M (1962) Body temperature during work at different environmental temperatures. Acta Physiol Scand 56,120-129
  30. Aulick, LH, Robinson, S, Tzankoff, SP (1981) Arm and leg intravascular temperatures of men during submaximal exercise. J Appl Physiol 51,1092-1097[Abstract/Free Full Text]
  31. Holmgren, A, McIlroy, MB (1964) Effect of temperature on arterial blood gas tensions and pH during exercise. J Appl Physiol 19,243-245[Abstract/Free Full Text]
  32. Greenleaf, CJ, Greenleaf, JE, Card, DH, et al (1969) Exercise-temperature regulation in man during acute exposure to simulated altitude. J Appl Physiol 26,290-296[Free Full Text]
  33. Kenny, GP, Reardon, FD, Thoden, JS, et al (1999) Changes in exercise and post-exercise core temperature under different clothing conditions. Int J Biometeorol 43,8-13[CrossRef][ISI][Medline]
  34. Ekblom, B, Greenleaf, CJ, Greenleaf, JE, et al (1970) Temperature regulation during exercise dehydration in man. Acta Physiol Scand 79,475-483[ISI][Medline]
  35. Visser, BF (1960) Pulmonary diffusion of carbon dioxide. Phys Med Biol 5,155-160
  36. Gregory, IC (1974) The oxygen and carbon monoxide capacities of fetal and adult blood. J Physiol 236,625-634[Abstract/Free Full Text]
  37. Farrell, K, Wasser, T (1997) The S factor: a new derived hemodynamic oxygenation parameter; a useful tool for simplified mathematical modeling of global problems of oxygen transport. Adv Exp Med Biol 411,149-155[ISI][Medline]
  38. Yang, SS, Maranhao, V (1986) Cardiac catheterization. Cheng, TO, ed ,194-211 Pergamon Press The international textbook of cardiology. New York, NY.
  39. Braunitzer, G (1963) Molekulare struktur der hämoglobine. Nova Acta Acad Caesar Leopold Card 26,113-125
  40. Wasserman, K (1994) Coupling of external to cellular respiration during exercise: the wisdom of the body revisited. Am J Physiol 266,E519-E539[Abstract/Free Full Text]
  41. Wasserman, K (1988) New concepts in assessing cardiovascular function. Circulation 78,1060-1071[Abstract/Free Full Text]
  42. Eccles, RC, Swinamer, DL, Jones, RL, et al (1986) Validation of a compact system for measuring gas exchange. Crit Care Med 64,808-811
  43. Takala, J, Keinanen, O, Vaisanen, P, et al (1989) Measurement of gas exchange in intensive care: laboratory and clinical validation of a new device. Crit Care Med 17,1041-1047[ISI][Medline]
  44. Westenskow, DR, Cutler, CA, Wallace, WD (1984) Instrumentation for monitoring gas exchange and metabolic rate in critically ill patients. Crit Care Med 12,183-187[ISI][Medline]
  45. Ultman, JS, Bursztein, S (1981) Analysis of error in the determination of respiratory gas exchange at varying FIO2. J Appl Physiol 50,210-216[Abstract/Free Full Text]
  46. Hansen, JE, Sue, DY, Oren, A, et al (1987) Relation of oxygen uptake in low work rate in normal men and men with circulatory disorders. Am J Cardiol 59,669-674[CrossRef][ISI][Medline]
  47. Hansen, JE, Casaburi, R, Cooper, DM, et al (1988) Oxygen uptake as related to work rate increment during cycle ergometer exercise. Eur J Appl Physiol 57,140-145[CrossRef]
  48. Furuike AN, Sue DY, Hansen, et al. Comparison of physiologic dead space/tidal volume ratio and alveolar-arterial PO2 difference during incremental and constant work exercise. Am Rev Respir Dis 1982; 126:579–583
  49. Roston, WL, Whipp, BJ, Davis, JA, et al (1987) Oxygen uptake kinetics and lactate concentration during exercise in man. Am Rev Respir Dis 135,1080-1084[ISI][Medline]



This article has been cited by other articles:


Home page
 Exp PhysiolHome page
A. Crisafulli, C. Carta, F. Melis, F. Tocco, F. Frongia, U. M Santoboni, P. Pagliaro, and A. Concu
Haemodynamic responses following intermittent supramaximal exercise in athletes
Exp Physiol, November 1, 2004; 89(6): 665 - 674.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Physiol. Heart Circ. Physiol.Home page
A. Crisafulli, F. Melis, F. Tocco, U. M. Santoboni, C. Lai, G. Angioy, L. Lorrai, G. Pittau, A. Concu, and P. Pagliaro
Exercise-induced and nitroglycerin-induced myocardial preconditioning improves hemodynamics in patients with angina
Am J Physiol Heart Circ Physiol, July 1, 2004; 287(1): H235 - H242.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
G. Laszlo
Respiratory measurements of cardiac output: from elegant idea to useful test
J Appl Physiol, February 1, 2004; 96(2): 428 - 437.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Respir. Crit. Care Med.Home page
X.-G. Sun, J. E. Hansen, N. Garatachea, T. W. Storer, and K. Wasserman
Ventilatory Efficiency during Exercise in Healthy Subjects
Am. J. Respir. Crit. Care Med., December 1, 2002; 166(11): 1443 - 1448.
[Abstract] [Full Text] [PDF]

Home page
 Anesth. Analg.Home page
R. Uda, M. Onaka, T. Okuno, H. Mori, X.-G. Sun, and C. Lee
The Second Gas Effect is Not Statistically Valid * Response
Anesth. Analg., March 1, 2002; 94(3): 765 - 766.
[Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
S. Godfrey, H. Kazemi, D. Systrom, K. Wasserman, X.-G. Sun, J. E. Hansen, H. Ting, and W. W. Stringer
Carbon Dioxide Pressure-Concentration Relationship in Arterial and Mixed Venous Blood
J Appl Physiol, November 1, 2001; 91(5): 2412 - 2414.
[Full Text] [PDF]

Home page
 CirculationHome page
X.-G. Sun, J. E. Hansen, R. J. Oudiz, and K. Wasserman
Exercise Pathophysiology in Patients With Primary Pulmonary Hypertension
Circulation, July 24, 2001; 104(4): 429 - 435.