Mixed venous oxygen saturation during mobilization after cardiac surgery: are reflectance oximetry catheters reliable?

Background: Oximetry catheters immediately reflect changes in mixed venous oxygen saturation (S(v) over bar O-2). We have used the Baxter 2-SAT system to register changes in S(v) over bar O-2, during early mobilizations after cardiac surgery. To assess catheter reliability, readings were compared to blood gases. Methods: A total of 352 paired catheter and bench haemoximetry measurements were obtained at the expected highest and lowest levels of S(v) over bar O-2, during the mobilization procedures. The agreement between methods was explored by a Bland-Altman plot. The influence of haemoglobin (Hgb), pH, cardiac output (CO), posture, catheter identity and catheter calibration on agreement was assessed through analysis of covariance. Results: Data included a substantial number of low S(v) over bar O-2 values, 95 paired means of S(v) over bar O-2 less than or equal to 50% and 37 paired means less than or equal to 40%. Mean oxygen saturation difference between catheter and haemoximeter readings was -1.6 +/- 5.7% (SD). Agreement between the methods depended upon the level of S(v) over bar O-2. At S(v) over bar O-2 of 65%, the two methods were virtually identical. Below 65%, the catheters increasingly underestimated the corresponding haemoximetric values by 1.5% for every 10% reduction in S(v) over bar O-2 Agreement was to some degree dependent on individual calibrations and catheter identity, but to a lesser extent on Hgb, CO and posture. Conclusion: The two methods are interchangeable for most clinical purposes. Catheter readings are, however, substantially lower than the corresponding haemoximetric measurements at low S(v) over bar O-2 values. Careful interpretation of the absolute values resulting from catheter measurements is recommended, especially when S(v) over bar O-2 readings are low.