Clinically interesting and yet not well understood, hypoxic-drive theory holds that people who chronically retain carbon dioxide lose their hypercarbic drive to breathe.

By William A. French, MA, RRT

One of the most clinically interesting and least understood theories in respiratory medicine is the hypoxic-drive theory. This holds that people who chronically retain carbon dioxide lose their hypercarbic drive to breathe. Thus, according to the theory, since the brain no longer responds to hypercarbia, the only remaining autonomic drive is hypoxemia. It then follows that, should patients in this condition be given enough supplemental oxygen to drive their PaO2 levels much higher than 60 mm Hg, they will also lose their hypoxemic drive to breathe.

Periodically, this theory is challenged, with the challenges based primarily on clinical observations that patients who exhibit the typical arterial blood gas (ABG) pattern suggestive of carbon dioxide retention do not simply stop breathing when their PaO2 levels climb.

A recent challenge1 approaches the question primarily from the standpoint of gas transport and the Haldane effect, as well as clinical observation. The neurophysiology of ventilatory control itself, however, can account for the theory as well as seemingly contradictory clinical observations.

Central Control of Ventilation

Neurological control of ventilation begins in the medulla, with the central controller. As part of the central nervous system, the central controller is isolated from the rest of the body by the cerebrospinal fluid (CSF). Although the exact mechanism involved is not fully understood,2 the central controller is primarily responsive to changes in the pH of CSF. At the same time, the blood/brain barrier is selectively permeable to carbon dioxide. Thus, when the carbon dioxide content of the blood increases, more carbon dioxide diffuses into the CSF. The pH of the CSF decreases due to the hydration of the carbon dioxide and the subsequent creation and release of hydrogen ions.

If the increase in carbon dioxide becomes chronic (lasts more than 24 hours), bicarbonate will begin to diffuse into the CSF and restore the CSF pH to its baseline level (7.326). At this point, the medulla is receiving a signal that the blood PCO2 is normal. Whether this restoration of the CSF pH at a higher blood PCO2 blunts the sensitivity of the central controller, or merely shifts the baseline for response upward, is not clear.

Although the central controller does not respond directly to hypoxia, it is known that medullary hypoxia can trigger respiratory depression. Likewise, it is known that high concentrations of carbon dioxide can cause narcosis; however, the PaCO2 must generally reach a level above 90 mm Hg for this to occur.

Peripheral Chemoreceptors

In addition to the central chemoreceptors in the medulla, the body also has peripheral chemoreceptors.2 Two are located near the bifurcation of the common carotid artery. Another is located in the ascending aorta. The carotid chemoreceptors are most active in ventilatory control. Each is a complex mass of tissue having a volume of about 6 mm. The principal cell type is the glomus, which can secrete dopamine as well as (possibly) noradrenaline, serotonin, acetylcholine, and polypeptides. The carotid bodies are innervated by both afferent and efferent branches of the glossopharyngeal (or ninth cranial) nerve.

The primary function of the carotid bodies is to sense and respond to changes in Pao2 levels. In the presence of normal PaCO2 and pH, the response of the carotid body begins a dramatic increase when the Pao2 decreases to less than 60 mm Hg. A number of factors can, however, shift the sensitivity (and, thus, the point of maximal response) of the carotid body. Among these are pH (in particular, acidemia), PaCO2 (principally hypercarbia), hypoperfusion, and increased body temperature. For example, if a person experiences a sudden metabolic acidosis (due to lactic or ketonic acid release), the sensitivity of the peripheral chemoreceptors would shift to a higher hypoxemic threshold, thus stimulating an increase in ventilation at higher Pao2 levels. The same shift would occur at increased PaCO2 levels, although whether there is an upper limit to this response is unclear.

Clinical Implications

It should be clear that patients who experience chronic hypercapnia will eventually readjust the pH of their CSF. Whether this blunts the central chemoreceptor response to pH change or simply shifts the CSF’s carbon dioxide baseline is unclear. Thus, patients who have been labeled carbon dioxide retainers or hypoxic-drive breathers may still be operating with some amount of hypercarbic drive.

Likewise, it should be clear that, under normal circumstances, the peripheral chemoreceptors become most active when the PaO2 drops to less than 60 mm Hg. Patients who have chronic hypercapnia are not in normal circumstances, however, and it may be that their peripheral chemoreceptors become most active at higher PaO2 levels.

From clinical observations, it is apparent that patients who exhibit the ABG pattern of compensated respiratory acidosis will not suddenly become apneic once their PaO2 rises to more than 60 mm Hg. Thus, over-oxygenating a patient under these conditions carries little risk.

A particular pattern, however, has been observed many times; its components are a pH of 7.29, a PaCO2 of 76 mm Hg, a PaO2 of 84 mm Hg, a bicarbonate level of 36 mEq, and a fraction of inspired oxygen (FiO2) of 0.3. Clinically, patients exhibiting this pattern often can be aroused, but are sleepy; they are also observed to be breathing more shallowly than normal. Given these conditions, simply lowering the FiO2 usually results in an increase in ventilation and a subsequent decrease in PaCO2.

Whether this phenomenon is caused by a blunting of the ventilatory drive or some other mechanism is not known, but this pattern is usually observed in patients who are relaxed and unstimulated. Certainly, most respiratory clinicians have observed that similar patients who experience transient increases in PaO2 (for example, through aerosol treatments powered by oxygen or through the use of an FiO2 of 1.0 during pulmonary function testing) do not demonstrate a similar decrease in ventilatory drive or level of consciousness.

In addition, the foregoing addresses only the role of the chemoreceptors in driving ventilation. In order to complete the picture, other potential ventilatory stimuli such as joint and muscle receptors and exogenous chemicals (for example, theophylline) should also be considered.

Case Report

A 63-year-old male was admitted to a hospital in Cleveland, with an acute exacerbation of chronic obstructive pulmonary disease (COPD). His admitting ABG levels (based on sampling while the patient breathed room air) were: pH, 7.41; PaCO2, 66 mm Hg; bicarbonate, 40 mEq; and PaO2, 49 mm Hg. His oxygen saturation level was 84%.

After admission and review by the attending physician, the patient began to use supplemental oxygen, delivered via nasal cannula, at a flow rate of 4 LPM. Approximately 8 hours later, an RCP drew arterial blood for routine ABG analysis. The results showed a pH of 7.36, a PaCO2 of 77 mm Hg, a bicarbonate level of 41 mEq, and a PaO2 of 74 mm Hg. The patient’s oxygen saturation level was 94%.

Clinically, the patient was reported to be alert, but drowsy. He was breathing shallowly, but was in no respiratory distress. Subsequently, the RCP recommended lowering the flow from the nasal cannula to 2 LPM, which was the flow rate that the patient had been using at home. The patient’s ABG levels stabilized, and he became less drowsy.


The neurological drive to breathe is complicated and is not fully understood. From a clinical standpoint, the injudicious and uncontrolled administration of supplemental oxygen to patients with compensated respiratory acidosis, although probably less risky than implied in the clinical literature, is not a good idea. The deliberate under-oxygenation of a patient with compensated respiratory acidosis (or a diagnosis of COPD) because of fear of hypoventilation or apnea, however, creates the greater risk of inducing prolonged tissue hypoxia.


William A. French, MA, RRT, is clinical director and assistant professor, Respiratory Therapy Program, Lakeland Community College, Kirtland, Ohio.


  1. Whitnack J. Why the hypoxic drive theory sucks wind. Available at: Accessed: December 1, 1999.
  2. Nunn JF. Applied Respiratory Physiology. 4th ed. Oxford, UK: Butterworth-Heinemann; 1993.