Lets face it, we don't really care about venous oxygenation (or arterial, for that matter). The real issue is how much oxygen is in the cells? How are those cells using it?
The intracellular environment is a place of massively heterogenous oxygen demands. Some organelles are involved in enzymatic metabolism of drugs or maintenance of electrolyte gradients - these areas require a higher oxygen tension than, for example, mitochondrial ATP synthesis.
The relationship of PaO2 to cellular PO2 and cellular oxygen utilization
One could come to view all that protein synthesis and molecular manipulation as mere twaddle, and focus squarely on the oxygen demand of ATP synthesis, because, after all, when that breaks down, terrible things begin to happen. Along the way to this tragic level of hypoxia, sacrifices are made. One good review article of this topic identifies four "threshold" stages of adaptive cellular behaviour in the face of worsening hypoxia.
This graph is an adaptation of a much better one from the 1990 article by Connet et al, which I have coloured in like a crazed toddler. In essence, it sends one message: that the "minimum" PO2 at which a cell can still function is defined by how hard it is working. Some lazy chondroblast somewhere may only require a minute oxygen tension to carry on its ATP synthesis, because its demands are not very high. Conversely, exercising muscle may be bathed in highly oxygenated blood and still be relying on anaerobic glycolysis for a little extra ATP.
But, of course, there is a Pcrit- the "critical partial pressure pressure" of PO2 at maximum oxygen extraction, below which the cell will struggle to maintain the same level of function. The article with the nice graph puts this value at 0.5 mmHg PO2, but hastens to add that each tissue has its own Pcrit. Oh's Manual (Ch 14) uses Siggaard-Andersen's value (0.8 mmHg PO2), which equates roughly to an PvO2 of 26mmHg, or an SvO2 of 50%; another author suggests 0.7mmHg as the critical value.
The most descriptive way of viewing this issue is through the lense of the oxygen extraction ratio (ERO2). This value is a relationship of oxygen delivery (DO2 which Ficks equation can give us) and oxygen extraction (VO2, which we can infer from the SvO2).
The relationship of SvO2 to the oxygen extraction ratio (ERO2)
ERO2, the ratio of delivered oxygen to extracted oxygen, is a useful value. A large value suggests that too much oxygen is being extracted, suggesting that tissue perfusion is poor.
This value can be derived from the mixed venous saturation.
Let us revisit Fick. What's that cardiac output equation again?...
Rearranged, we can see a relationship between DO2 and (CO x CvO2)
In fact, if one looks closely at this, on the left of the equation we now have the oxygen extraction value, DO2 - VO2.
Now, one can divide both sides by DO2 (CO x CaO2):
Because the oxygen content of blood is pretty predictable (lets just assume the hemoglobin concentration and its capacity to carry oxygen are not fluctuating wildly), we can reduce the formula to contain just the changing variable: oxygen saturation.
But ... The SaO2 in the ICU is pretty constant. Its not ever going to deviate much beyond 90-100%. So, by consciously disregarding the SaO2, we arrive to the conclusion that SvO2 = 1- ERO2.
Relationship of changes in oxygen extraction ratio to mixed venous saturation
The graph below demonstrates the relationship between VO2 and DO2. Because we now know that SvO2 = 1- ERO2, for every ERO2 value we can calculate the corresponding SvO2.
The flat plateau of the ERO2 curve represents the body's stable, unchanging metabolic demand. Once that demand is reached, the rate of oxygen extraction remains pretty flat.
However, as oxygen delivery decreases, it eventually reaches a critical point. Beyond that ERcrit, the amount of oxygen extracted from the delivered blood is near-maximal. At this point the tissue demands for oxygen are no longer being met.
Is there an empirical critical threshold for oxygen extraction in the ICU population?
Claudio Ronco and friends have performed some work on dying ICU patients, measuring their oxygen extraction ratios and arterial lactate, to work out how low you can go with the venous saturation. Lactates started rising in their population at an oxygen extraction ratio (ERO2) of around 60%, thus giving us a critical SvO2 value of around 40%.