Respiration is often dichotomized into oxygenation and ventilation because the two functions are independent of each other despite sharing the same organ. In fact, under certain conditions, the two functions are inversely related.


This is best exemplified in experiments on breath holding. There is very little change to SaO2 during breath holding (0.1-1%). The degree of change depends on the SaO2 starting point (A). When plotted against time, there is no correlation between degree of desaturation and apnea time (B). This is because oxygenation, except in the extremes, is not dependent on ventilation. The steady drop in SaO2 is due to progressive mixing of deoxygenated blood from the oxygen depleted portions of the lungs with oxygenated blood from areas of the lung with oxygen reserve. This also explains why patients with lower SpO2 prior to intubation become hypoxic during apnea so much faster than expected: it takes ~10s to fall from 99% to 98% and ~1s to fall from 88% to 87%.

Desaturation is due to shunt. Pathology that increases shunting revolve around localized inflammatory responses that increase blood flow despite decreased ventilation (atelectasis, small airway plugging, PE, or PNA). It can also result from diffuse lung processes that cause severe diffusion impairment (end stage COPD, emphysema, pulmonary edema) and overwhelm the lung vasculature’s natural mechanisms for matching perfusion to ventilation.

The best example for thinking about shunt is to imagine an aspiration that blocks the left main bronchus. 50% of the cardiac output enters the left lung with a saturation of 70% and a PaO2 of 30mmHg and leaves the left lung unchanged. The other 50% of the CO enters the right lung and leaves oxygenated (saturation 100% and PaO2 100). The resulting admixture within the left atrium has a PaO2 averaged to 65mmHg, corresponding to a saturation of 85% by the oxygen-hemoglobin dissociation curve. The patient will have a SpO2 of 85%.

If the patient then developed a PE that completely obstructs the left pulmonary artery, the SpO2 returns to 100% because now 100% of the cardiac output enters the right lung where it is fully oxygenated; there is no admixture. Mismatch causes hypoxemia. That being said, PE in and of itself can cause hypoxemia. The mechanism is thought to be through 1) R>L shunt from elevated right heart pressure (foramen ovale re-opens), and 2) inflammation causing hyperemia through thrombophlebitis physiology; this is why CXR findings are so variable in PE and Westermark’s sign (regional oligemia) is so rare.

Shunt does not cause hypercarbia except when severe. And it does so likely through creating regional oligemia and thus deadspace.


The effect of breath holding on PaCO2 and pH is the exact opposite. This explains why the apneustic centers are far more sensitive to PaCO2 (hypercapnic respiratory ventilation HCRV) than PaO2 (hypoxic respiratory ventilation HRV). The body knows that hyperventilating corrects hypercarbia but will have no effect on hypoxemia except in extreme circumstances like prolonged apnea.

Hypoxic respiratory ventilation kicks in at PaO2 < 50 according to some physiologic studies but hyperventilation is of little utility on PaO2 and can actually pose harm by increasing O2 consumption (imagine high altitude climber gasping for breath in vain vs indigenous high altitude populations who simply learn to live with low SpO2).

Ventilation has a profound effect on pCO2 but a very limited effect on oxygenation. In room air, only 8 breaths/min are needed to maintain oxygenation. Only 2 breaths/min are needed at FiO2 of 50%. This is why hyperventilating a hypoxic patient is an exercise in futility, and possibly introduces harm.

Apneic Oxygenation in Man (1959) studied fully sedated and paralyzed patients receiving only passive oxygenation via ETT and reservoir bag containing 100% FiO2. The bag would deflate over time and was refilled with 2-3L every 15min. Nothing explains the independence of ventilation from oxygenation better:

Inverse Relationship – Supplemental O2 causes Hypercarbia

Hyperventilation does not produce improved oxygenation. On the other hand, hyperoxygenation can worsen ventilation. This emphasizes the independence of the two respiratory functions.

Supplemental oxygen, through the Haldane effect, causes hemoglobin to lose its affinity for CO2 in preference for O2 and results in hypercarbia. Hence, it makes no mechanistic sense to treat shortness of breath with supplemental oxygen if they are saturating well on RA. Doing so may actually worsen their dyspnea.

Thankfully, the Haldane effect is usually transient because the organism increases respiratory drive and blows off the excess CO2. But in patients with blunted hypercapnic drive from chronic CO2 retention, the CO2 stays high and can cause delirium/coma. Furthermore, hyperoxgenation causes cerebral vasoconstriction causing hypoventilation and hypercarbia through this separate mechanism.

Inverse Relationship – PEEP causes Hypercarbia

Treating hypoxia with PEEP can worsen PaCO2 by converting West lung zone 2 turn into zone 1, increasing deadspace and worsening hypercarbia. IPAP is the setting to titrate in hypercarbic respiratory failure. IPAP increases tidal volumes and decreases work of breathing. PEEP should be left at minimum settings in hypercarbic respiratory failure (typically from COPD or asthma) unless you really know what you are doing.