Objective: To determine the accuracy of continuous (in vivo) measurement of mixed venous oxygen saturation (SvO(2)), using a fibreoptic catheter, in patients having had cardiopulmonary bypass (CPB).

Methods: Using a pulmonary arterial catheter, we prospectively studied 14 patients (age 64 +/- 8) having had cardiopulmonary bypass. Mean hematocrit was 30 +/- 4%. The catheter was calibrated in vitro and in vivo, according to the manufacturer's instructions. Fifty-six simultaneous measurements of continuous SvO(2) (CSvO(2)) and measured SvO(2) (MSvO(2)) were taken with a co-oxymeter and the paired values were analyzed by the linear regression method. To make the two sets of measurements interchangeable, we established, a priori, a maximum limit of 3% (approximately 5% of the measurement), as being an acceptable difference between the two types of measurements.

Results: All the measurements were obtained within four hours of the placement of the catheter. CSvO(2) was weakly correlated with MSvO(2), with a correlation coefficient of r(2) = 0.49 (P < 0.001). The Bland-Altman analysis demonstrates an objective mean bias of 0.8 +/- 3%, with 36% of the values measured falling outside clinically acceptable limits. For values of CSvO(2) <or= 65%, the measurements were not correlated (P = NS). The bias was -2 +/- 7%, and 56% of the values measured fell outside clinically acceptable limits.

Conclusions: Our results demonstrate that, in the immediate aftermath of cardiopulmonary bypass, the continuous measurement of SvO(2) and the co-oxymetry determination of mixed blood is not inter-changeable. Other studies will be necessary to confirm these results and enable us to understand the pathophysiological phenomena accounting for the lack of agreement between the two methods.