Femoral venous oxygen saturation and central venous oxygen saturation in critically ill patients.

OBJECTIVES

To investigate the relationship between central venous oxygen saturation (ScvO(2)) and femoral venous oxygen saturation (SfvO(2)) in a large group of critically ill patients.

DESIGN

Observational study.

PATIENTS

A group of unselected critically ill patients with central line placed into superior vena cava were included.

SETTING

A 26-bed intensive care unit in a tertiary referral hospital.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Venous blood samples of superior vena cava and femoral vein were collected within an interval of 5 to 15 minutes and analyzed with blood gas/electrolyte analyzer immediately. Although SfvO(2) was significantly correlated with ScvO(2) (r = 0.493, P < .001; Pearson correlation, 2 tailed), the limits of agreement were wide (up to 61% to -41%) between the 731 pairs of blood samples collected from 357 patients. The fit line of scatter diagram ScvO(2) vs SfvO(2) had a large intercept (48.68%) and a low slope (0.2978); ScvO(2) was still around 50% while SfvO(2) was nearing 0%. The distribution of blood flow, measured with Doppler ultrasound, had a similar trend in 237 patients and 412 measurements. The ratio of femoral artery flow over common carotid artery flow varied widely (from 0 to 7.13). Blood flow was not distributed in a fixed ratio to the superior vena cava-drained organs and tissues.

CONCLUSIONS

Central venous oxygen saturation was not representative of the whole systemic circulation in critically ill patients. Central venous oxygen saturation alone might be misleading in goal-directed therapy.