Oxygen Toxicity

Utility of PaO2

PaO2 is not a useful measurement for hypo-oxygenation compared to SaO2. It does, however, calculate % shunt. Physiologic shunting increases with age, so expected PaO2=100-0.325 ×age. PaO2 is very useful for hyper-oxygenation. There is little reason for keeping PaO2 >100 except in CO poisoning.

Oxygen Carrying Capacity

There is a marginal increase in oxygen carrying capacity (about 8mL O2 difference between RA and 100% FiO2). However, this oxygen does not reach the capillaries because it reacts with the surrounding tissues to produce ROS before its destination.

Total Volume of Oxygen in Tissues

Oxygen Delivery

Hyperoxygenation produces vasoconstriction in every tissue but lung and placenta. The systemic result is increased SVR with variable increases in blood pressure and variable decreases in HR. CO, in net, is reduced. Hence, hyperoxygenation improves oxygen content but decreases oxygen delivery. (Paul Marino’s ICU book).


Oxygen also produces two time-dependent effects on ventilation: an early short-lasting decrease in RR f/b longer period of increased RR. The early phase is presumed to occur from decreased cerebral blood flow. The late phase occurs from reverse Haldane effect, resulting in increased PaCO2 and compensatory increase in RR.

This is the best explanation I have found for carbon dioxide retention. In patients who are already hypercarbic, the initial phase decreases RR and worsens hypercarbia; the late phase displaces CO2 and additionally worsens hypercarbia. Chronic COPD patients have adapted to tune out their PaCO2 sensors and do not make appropriate compensation to blow off CO2, resulting in hypercarbic delirium.


Degree of FiO2 directly corresponds to amount of atelectasis as measured on CT 10 minutes after exposure!


And lastly, a 1971 article quotes dramatically shortened lifespans of animals subjected to 100% FiO2: fruit-fly 168 h; mouse 77–306 h; cat 83 h; and monkey 180–390h.