[Comparison of bench central and mixed pulmonary venous oxygen saturation in critically ill postsurgical patients]

To investigate if there were differences between bench central oxygen saturation (ScvO2) and mixed venous oxygen saturation (SvO2) in a group of acutely ill postsurgical patients.

A prospective comparative study of two sampling sites.

Postsurgical ICU at a University Hospital.

39 acutely ill postsurgical patients, requiring perioperative invasive hemodynamic monitoring, studied during their stay in a postsurgical ICU.

Routine care for acutely ill postsurgical patients.

Blood was simultaneously sampled, according to the clinical course, both from the distal and central port of a pulmonary artery catheter (n = 296 each). Oxygen saturation and blood gas analysis were immediately measured with a CO-oxymeter and a blood gas analyzer. We investigated the relationship and the agreement between the two measures. To assess if bench ScVO2 could give information on time related variations of bench SvO2 we evaluated the absolute sequential changes of the two measures during the period of observation, by analyzing the relationship and the agreement of their absolute changes (n = 1817). There was a significant difference between bench SvO2 and bench ScvO2 (71.8 +/- 8.2 vs 72.7 +/- 8.6% mean +/- SD, p < 0.001). The relationship between the two measures showed a significant correlation (r = 0.90, p < 0.001, and SEE 3.8%). The bias was -0.93 +/- 3.8%, and the limits of agreement were +6.6 and -8.5%. The changes in bench ScvO2 correlated with the respective changes in bench SvO2 (r = 0.86, p < 0.001, and SEE 4.4%). The bias was 0.79 +/- 4.7% and the limits of agreement were +9.4 and -7.8%.

SvO2 cannot be predicted well from bench ScvO2, nor changes in ScO2 can be predicted wel from changes in bench ScvO2. Therefore, in this category of patients, the clinical usefulness of monitoring bench ScvO2 is strongly limited and we must still rely on the SvO2.