Accuracy of continuous central venous oxygen saturation monitoring in patients undergoing cardiac surgery

Abstract

Objective: Continuous assessment of central venous oxygen saturation (S(cevox)O(2)) with the CeVOX device (Pulsion Medical Systems, Munich, Germany) was evaluated against central venous oxygen saturation (S(cv)O(2)) determined by co-oximetry.

Methods: In 20 cardiac surgical patients, a CeVOX fiberoptic probe was introduced into a standard central venous catheter placed in the right internal jugular vein and advanced 2-3 cm beyond the catheter tip. After in vivo calibration of the probe, S(cevox)O(2), S(cv)O(2), mixed venous oxygen saturation (S(mv)O(2)) haemoglobin (Hb), body temperature, heart rate, central venous and mean arterial pressure, and cardiac index were assessed simultaneously at 30 min intervals during surgery and at 60 min intervals during recovery in the intensive care unit. Agreement between S(cevox)O(2), and S(cv)O(2) was determined by Bland-Altman analysis. Simple regression analysis was used to assess the correlation of S(cevox)O(2), and S(cv)O(2) to Hb, body temperature and haemodynamic parameters.

Results: Values of S(cevox)O(2) and S(cv)O(2) (84 data pairs during surgery and 106 in the intensive care unit) ranged between 45-89% and 43-90%, respectively. Mean bias and limits of agreement of S(cevox)O(2) and S(cv)O(2) were -0.9 (-7.9/+6.1)% during surgery and -1.2 (-10.5/+8.1)% in the intensive care unit. In 37.9% of all measured data pairs, the difference between S(cevox)O(2) and S(cv)O(2) was beyond clinically acceptable limits (> or =1 s.d.). Mean bias was significantly influenced by cardiac index. Sensitivity and specificity of S(cevox)O(2) to detect substantial (> or =1 s.d.) changes in S(cv)O(2) were 89 and 82%, respectively.

Conclusions: In adult patients during and after cardiac surgery, the current version of the CeVOX device might not be the tool to replace S(cv)O(2) determined by co-oxymetry, although sensitivity and specificity of S(cevox)O(2 )to predict substantial changes in S(cv)O(2) were acceptable.