Effect of carboxyhemoglobin on the accuracy of mixed venous oximetry monitors in dogs.

OBJECTIVE

To assess the accuracy of mixed venous hemoglobin oxygen saturation estimated, using in vivo pulmonary artery reflectance oximetry with ranging concentrations of carboxyhemoglobin.

DESIGN

Criterion standard, a comparison of an alternative test to the "gold standard."

SETTING

Laboratory animal facility of a large university.

SUBJECTS

Five mongrel dogs.

INTERVENTIONS

Anesthetized dogs were mechanically ventilated to normocarbia and instrumented with arterial catheters and pulmonary artery oximetry catheters. The dogs were ventilated with increasing inspired concentrations of carbon monoxide, and blood samples were analyzed for fractional concentrations of oxyhemoglobin and carboxyhemoglobin. Carboxyhemoglobin levels ranged from 0% to 70%. At each level of carboxyhemoglobin, FIO2 was varied from 0.21 to 0.09. Co-oximeter readings were compared with mixed venous oxygen saturation measurements of the pulmonary artery oximetry catheter.

MEASUREMENTS AND MAIN RESULTS

Mixed venous oxygen saturation measurements from the pulmonary artery oximetry catheter system progressively overestimated fractional oxyhemoglobin in the presence of carboxyhemoglobin. This catheter's mixed venous oxygen saturation reading could be corrected for the presence of carboxyhemoglobin by a derived formula. Regression analysis for corrected mixed venous oxygen saturation (calculated) vs. fractional oxyhemoglobin yields a slope and intercept of 0.98 and -1.01, respectively, with a correlation coefficient of .98. The bias and precision values for fractional oxyhemoglobin vs. mixed venous oxygen saturation (calculated) are 1.9 and 2.2 (n = 66), with bias representing the degree of systematic error or deviation from the true measurement, and precision representing the confidence limits (or standard deviation) for individual variations from the true measurement.

CONCLUSIONS

Mixed venous oxygen saturation monitoring does not detect the presence of carboxyhemoglobin and progressively overestimates fractional oxyhemoglobin as carboxyhemoglobin increases. Mixed venous oxygen saturation values of the standard pulmonary artery oximetry catheter approximately equal functional hemoglobin saturation. Bench co-oximeter blood analysis is required in patients suspected of having increased carboxyhemoglobin levels.