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

To investigate the relationship between central venous oxygen saturation (ScvO2) and femoral venous oxygen saturation (SfvO2) in a large group of critically ill patients.Observational study.A group of unselected critically ill patients with central line placed into superior vena cava were included.A 26-bed intensive care unit in a tertiary referral hospital.None.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 SfvO2 was significantly correlated with ScvO2 (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 ScvO2 vs SfvO2 had a large intercept (48.68%) and a low slope (0.2978); ScvO2 was still around 50% while SfvO2 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.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.