Estimation of Mixed Venous Pco2 for Determination of Cardiac Output in Children (Results)A total of 14 children (11 boys and 3 girls), mean age 11.7±1.6 years, range 8.2 to 13.9 years, successfully underwent steady-state exercise tests in which PvC02 was measured by both the exponential and equilibrium methods. Eight were healthy children of hospital staff, four were healthy prematurely born children, and two were otherwise healthy children who had undergone testing for growth hormone secretion due to short stature. Nine other children (five boys and four girls) underwent steady-state exercise tests in which PvC02 was obtained by the equilibrium method alone. Five were children of hospital staff and four were prematurely born. The mean age of the 23 children was 11.0±1.9 years, range 7.1 to 13.9 years. Twelve adults (seven men and five women), mean age 33.6±7.2 years, range 24 to 48 years, underwent steady-state exercise tests in which PvC02 was measured by the equilibrium method.


In exercising children, there was a closer agreement between the downstream corrected equilibrium PvC02 and exponential values than between the uncorrected equilibrium and exponential measures. As is seen in Figure 1, the equilibrium downstream corrected values of PvC02 plotted against the exponential values, fell closer to the line of identity than did the uncorrected equilibrium values of PvC02. Mean PvC02 by the exponential method was 62.0±5.1 mm Hg. By the equilibrium method, absolute PvC02 was 68.4±4.9 mm Hg, and corrected PvC02 was 63.0±3.7 mm Hg. By analysis of variance, there was a significant difference between the mean uncorrected equilibrium value and the mean values obtained by the other two methods (p=0.012).

When the uncorrected equilibrium PvC02 was plotted against exponential PvC02 in mm Hg (Fig 1, filled squares), we found a slope of 0.89 (SE, 0.1) and an intercept of 14.1 (SE, 6.8). For corrected equilibrium PvC02 (PvC02 eq,corr) vs exponential PvC02 (PvC02 exp) (Fig 1, open squares), we found a slope of 0.67 (SE, 0.08) and an intercept of 21.7 (SE, 5.2). In adults, Heigenhauser and Jones found the following relationship between the PvC02 eq,corr and the PvC02 exp:

PvC02 eq,corr (kpa)=2.38+0.7 • PvC02 exp (kpa)

When our regression equation is expressed in Idlo-pascals,

PvC02 eq,corr (kpa)

=2.89 (SE, 0.69) + 0.67 (SE, 0.08)

• PvC02 exp (kpa), r2=0.83

the slope and intercept are not significantly different from those obtained by Heigenhauser and Jones.

In a later study from the same group, values fell closer to the line of identity, with a slope of 0.975 and an intercept of 1.46. Their data encompassed a much broader range than did ours, and our values can be superimposed on their concordance plot.

Cardiac Output 

Downstream corrected equilibrium Q in liters per minute plotted against V02 in liters per minute for all adults and children gave us the following regression equation:

Q (L/min) = 1.79 (SE, 0.40) + 7.89 (SE, 0.31)

• Vo2 (L/min), r2=0.95

for submaximal exercise with a range of Vo2 from 0.5 to 2.5 L/min. This equation is very similar to that obtained using the combined values from two previously published dye dilution studies, one involving pubertal boys and the other involving young men:

Q (L/min) = l.75+7.95 • Vo2 (L/min), r2=0.48

This relationship can be seen in Figure 2.

A small difference in the Q—Vo2 relationship between children and adults was determined using stepwise linear regression analysis and a grouping factor, gp (0 for children and 1 for adults):

Q (L/min) = 2.40 (SE, 0.39) + 7.05 (SE, 0.36)

• Vo2 (L/min) + 1.21 (SE, 0.35) • gp,


This difference is eliminated when weight (wt) is factored into the equation

Q (L/min) = 1.42 (SE, 0.33)+5.80 (SE, 0.53)

• Vo2 (L/min)+0.06 (SE, 0.01) • wt (kg),


When calculated without an intercept for comparison with the equation suggested by Jones,

Q (L/min)=5.5 * Vo2 (L/min) + 0.06 • wt (kg)

the equation becomes

Q (L/min)=6.25 (SE, 0.64)

• Vo2 (L/min)+0.075 (SE, 0.016)

# wt (kg)

which has coefficients that are not significantly different from those proposed by Jones. Sex was not a contributing factor to the regression equation, nor was subject category (term, prematurely born, or undergoing testing for growth hormone secretion) a contributing factor among children.

When Q and Vo2 were expressed in mL/min/kg, the following equation was obtained:

Q (mL/min/kg)

=83.5 (SE, 20.0)+6.2 (SE, 0.8)

• Vo2 (mL/min/kg), r2=0.66

In our study, for men alone, the following equation was obtained:

Q (mL/min/kg)=80.3 (SE, 22.8)+5.8 (SE, 0.9)

• Vo2 (mL/min/kg), r2=0.88

which has a slope and intercept not significantly different from the equation described by Faulkner and coworkers in their study of 50 normal adult male subjects at various steady-state work loads using the exponential method (Fig 3):

Q (mL/min/kg)=66+5.2 * Vo2 (mL/min/kg)

Values in children were slightly higher:

Q (mL/min/kg)=97.1 (SE, 27.5)+5.9 (SE, 1.0)

• Vo2 (mL/min/kg), r2=0.60

and, when results for adults and children were combined, the grouping factor (adults vs children) contributed significantly to the regression equation (r2 increased from 0.66 to 0.71).


Figure 1. Mixed venous Pco2 (mm Hg) obtained using the equilibrium method (PvC02,eq) vs mixed venous Pco2 obtained using the exponential method (PvC02,exp). Solid squares represent uncorrected values and open squares represent downstream corrected values for PvC02,eq. The lines represent the line of identity for PvC02,exp, and 10% above and below the line of identity. The downstream corrected values of PvC02,eq fall closer to the line of identity of than do the absolute values of PvC02,eq.


Figure 2. Q (L/min) vs Vo2 (L/min). The relationship between Q, calculated using the downstream corrected values of PvC02 for children and adults (“X” and asterisks) and Vo2 is not significantly different from that of other studies in which Q was determined by dye dilution for children and adults (solid squares and open squares) (Adapted from Eriksson et al24 and Ekblom et al23).


Figure 3. Q (mL/min/kg) vs Vo2 (mL/min/kg). Most of the values for adults (“X” = absolute and asterisk = corrected) fall within the 95% confidence interval for values of Faulkner et al22 using the exponential method. The values for children (solid squares = absolute and open squares = corrected) are somewhat higher.