Variation over a 1-hour period in 26 ventilator-dependent trauma victims who were clinically stable. (Hess D, Agarwal NN. Variability of blood gases, pulse oximeter saturation, and end-tidal carbon dioxide pressure in stable, mechanically ventilated trauma patients. J Clin Monit 1992;8:111.)
And it correlates well at all pH.
PvCO2 < 45 has a 100% sens and 100% spec for PaCO2 <50. It can be used as a rule out if normal. Agreement is poorer when PvCO2 is abnormal.
And it correlates well at all bicarb levels.
Normal venous lactate correlates well with arterial lactate and has a 0.1 LR for elevated arterial lactate.
SvO2 is measured in the pulmonary artery. ScvO2 is measured in the SVC. There is an erroneous belief that arterial gases are always superior to venous gases. This may be true when assessing for pulmonary status because arterial blood comes from pulmonary vascular bed. But the above meta suggests that there is good correlation at near normal levels.
In states of shock, venous gases from ScvO2 or SvO2 are likely more accurate. They originate from systemic tissue beds and more accurately reflects systemic processes during impaired tissue perfusion. An older study addressed this difference in cardiac arrest and found the arterial blood pH averaged 7.41, whereas the average mixed venous blood pH was 7.15; mean arterial PCO2 was 32, whereas the mixed venous PCO2 was 74 (Weil MH, Rackow EC, Trevino R, Grundler W, Falk JL, Griffel MI. Difference in acid-base state between venous and arterial blood during cardiopulmonary resuscitation. N Engl J Med. 1986;315(3):153-6).
To me, this means that shock and cardiac arrest patients need VBG (preferably from central access) and ABG. VBG to tell you about perfusion. ABG to tell you about ventilation. After-all, you can only titrate vent settings to the ABG in these circumstances.