Relationship of thermodilution cardiac output to metabolic measurements and mixed venous oxygen saturation

To determine the individual contributions of variables in the Fick equation to cardiac output, we simultaneously measured oxygen uptake (VO2), carbon dioxide production (VCO2), venous oxygen saturation (SvO2) and thermodilution cardiac output (Qth) in 28 medical and surgical ICU patients. Patients were intubated and ventilated with the intermittent mandatory ventilation mode. VO2 and VCO2 (averaged over 3 min) were obtained from a metabolic cart. SvO2 was measured with fiberoptic reflectance oximetry (and COoximetry). Thirty-nine studies (average duration, 4.3 h) with 151 Qth measurements were performed. The relationships between Qth and VO2, Qth and VCO2, Qth and SvO2, and 1/Qth and SvO2, as well as between the sequential changes in these variables were analyzed by least squares linear regression. The ability of changes in the variables VO2, VCO2, and SvO2 to predict changes in Qth were analyzed by receiver operating characteristic (ROC) curves. Qth was weakly related to VO2 (r = 0.45), VCO2 (r = 0.45), or SvO2 (r = 0.36). Changes in Qth were weakly related to changes in VCO2 (r = 0.40), and even less to changes in VO2 (r = 0.18) and SvO2 (r = 0.13). The areas under the ROC curves for increases in Qth > 10 percent were as follows: 0.66 for VCO2, 0.50 for VO2, and 0.55 for SvO2. The areas for decreases in Qth < 10 percent were as follows: 0.78 for VCO2, 0.65 for VO2, and 0.49 for SvO2. None of the above oximetry relationships were substantially altered by use of COoximetry venous oxygen saturations. We conclude that Qth cannot be predicted well solely from VO2, VCO2, or SvO2 nor can changes in Qth be predicted well solely from changes in VO2, VCO2, or SvO2. Of the metabolic variables, changes in VCO2 best predicted changes in Qth.