Relationship of arterial carbon dioxide and alveolar ventilation

This chapter is most relevant to Section F2(i) from the 2017 CICM Primary Syllabus, which expects the exam candidates to be able to "describe the control of breathing". The concentration of arterial CO2 is certainly an important component of this control system. So far, only one CICM SAQs have ever asked the trainees to produce a description of how alveolar ventilation changes in relation to changing PaCO2, or how the change in ventilation can affect the arterial CO2 tension (Question 16 from the first paper of 2019).  Extending this slightly, one might envision some scenario where the trainees might one day be expected to not only describe these relationships, but also how they are affected by disease states and interventions. But so far that has not been the case. As such, the time-poor exam candidate is warned against reading any further, and directed instead to the chapter on control of ventilation which actually has exam relevance.

In summary:

  • Increasing PaCO2 causes an increase in minute ventilation.
  • This is mediated by peripehral chemoreceptors over the timescale of seconds, and by central chemoreceptors over minutes.
  • The relationship between PaCOis fairly linear in the range of 45-80 mmHg; the rate of minute volume increases by 2-5L/min per every 1mm Hg of CO2 increase.
    • The CO2/ventilation response curve is shifted to the left by metabolic acidosis and hypoxia
    • Sleep, sedation, anaesthesia and opiates shift the curve to the right and decrease the slope of the curve (i.e. the increase in minute ventilation is reduced per unit rise of CO2)
    • Age decreases the ventilatory response to CO2
    • A high level of physical fitness also diminishes hypercapnic respiiratory drive
  • The response to raised PaCO2 is rapid; about 75% of the maximum minute volume change is achieved over minutes
  • At a stable metabolic rate and with minimal inspired CO2 the relationship between minute volume and PaCO2 is described by a hyperbolic curve.

Influence of PaCO2 on minute ventilation

It will surprise nobody with even some minimum of medical training when we say that increasing CO2 increases one's "respiratory drive", whatever that is. That's a tenet of physiology which is so well embedded that we generally take it for granted. Interestingly, prior to the late nineteenth century, though "carbonic acid" was a well-known component of exhaled gas, nobody paid much attention to it, and the toxic effects of rebreathing expired air were attributed to "poisonous, volatile, and probably inodorous, impurities" which were present in a person's breath "arising from want of cleanliness". 

This was for the first time debunked by J.S Haldane (of the epic moustache) and Lorraine Smith (of the Lorraine Smith effect) in 1892 by means of an elegant experiment, based mainly on "a person simply allowed to remain in a closed chamber till the air became very impure". One of them got into the chamber and sat there until the FiCO2 in the chamber exceeded 5% and they "requested to be released on account of the exhaustion produced by the hyperpnoea". They concluded, "Any one breathing air which contains 5.6 per cent. carbonic acid is obliged to breathe at about 30, and as deeply as possible". Unsatisfied with this demonstration of the respiratory stimulant effects of CO2, the experimenters pushed on into greater concentrations of the gas. This is classic Haldane; "We tried on ourselves the effect of breathing air which contained 18.6 per cent. of carbonic acid".

Haldane and Smith breathing FiCO2 of 18%

Though both heroic and hilarious, this early study only counted the respiratory rate, whereas some might argue that the minute volume might be of greater interest.  J.S Haldane & Priestley J.G (1905) were probably the first to quantify this parameter relative to increasing CO2 concentration.  To do this, they trapped each other in a box. 

Body plethysmograph from Haldane & Priestley (1905)

Each investigator took turns being confined in this small space and made to breathe weird gas mixtures. The rubber neck hole "fitted the neck as tightly as possible without causing congestion", which meant that any movement of gas into and out of the subject's chest was recorded by the piston as a displacement of gas volume from the plethysmograph box. As was usual with J.S.Haldane, the investigators themselves got into the box several times. From their records, it is possible to determine that J.S Haldane weighed 83 kg and his alveolar ventilation at rest was quite stable over multiple measurements (between 8.5 to 9.8 litres) whereas his 69.7kg co-author had a more variable minute volume because "during part of the experiments he was in training for a boat race"

Haldane and Priestley's objective was to test the effects of an increasing CO2 load on minute ventilation.  It being inconvenient to add carbon dioxide directly to the bloodstream of volunteers, these data come from experiments where the arterial CO2 concentration was increased by contributing CO2 to the inspired gas mixture. The scientists sat in the plethysmograph and then enclosed each other's heads in a small box (of approximately 45L capacity) into which they piped gas mixtures with increasing CO2 concentrations. Because the Severinghaus electrode was still over fifty years away, the measurement of arterial CO2 was substituted by the measurement of end-tidal CO2, it being thought of as a reasonable surrogate.  Out of respect to these pioneers, the original data are presented below:

original data from Haldane and Priestley (1905)

As one can plainly see, as the fraction of inspired CO2 increased, so the tidal volume increased dramatically. The younger more physiologically robust Priestley (he was 25 at the time, whereas Haldane was 45) was able to tolerate an FiCO2 up to 7.66%, which would make for an arterial CO2 tension of around  64 mmHg. At this level of hypercapnia, his minute volume increased by about 700%. The experience was not entirely benign, particularly when Haldane turned down the FiO2 to 13%. "The subject became extremely cyanosed, as shown by the leaden blue colour of the face and lips; and consciousness began to fail", he calmly observed.   

 The minute volume increase can then be plotted against the arterial CO2 concentration. It is quite a steep curve at higher CO2 concentrations: minute volume increases dramatically to keep the arterial CO2 stable. In fact, calling it a curve is also somewhat generous. It is essentially a linear relationship at the higher range of values. Here are some curves generated by Mohan & Duffin (1997) from healthy volunteers rebreathing their own CO2 at a stable PaO2 of around 100 mmHg:

ventilatory responses to CO2 from Mohan et al (1997)

As PaCO2 decreases, minute volume also decreases, and at the lower range of PaCOvalues this relationship is no longer as linear as it is at higher values, largely because one cannot sustain a minute volume of 0 L/min for very long. Therefore, a sort of "dog leg" appears, with the minute volume decrease turning into a plateau at an arterial CO2 tension of around 40 mmHg. Physiology textbooks usually focus on this middle point of the ventilation response range, and in those chapters one typically finds a diagram which looks something like this:

ventilatory response to changes in inspired CO2 in mid range of CO2

The shape of this graph was reproduced from Mohan & Duffin (1997), to maintain some relationship to reality. The line is usually extended to intercept the x-axis at some imaginary minute volume of 0.  This intercept has relevance which will be demonstrated below (it is a convenient means of comparing CO2 response characteristics under different physiological conditions). The line also has a slope, as lines generally do, and this is also an important examinable feature, as it describes the sensitivity to CO2. In normal young men, this sensitivity averages about 2-5L/mmHg (i.e. for every 1mm Hg increase in arterial CO2, the minute volume should increase by around 2-5L). 

The timing of this response is relatively rapid. The graphics below come from Schaefer (1958), with some minor modification. These were youngish adult males (19-39 years of age) breathing an FiCO2 of around 7.5%. As one can see, at the initial challenge both the respiratory rate and minute volume doubled over the course of about ten minutes.

timing of ventilator response to CO2

This response appears to be in phases. The times and comments below are from Tansley et al (1998):

  • Initial rapid component: occurs over 8-26 seconds and contributes 12-30% of the response; attributed to the effect of hypercapnia on the carotid bodies
  • Second delayed component: occurs over 65-180 seconds and contributes 70-88% of the total response; attributed to the effect of hypercapnia on central medullary chemoreceptors
  • Third slow component: occurs over hours, contributes the last 15-25% of the ventilatory response to hypercapnia, and is caused probably by some sort of central adaptation, though this is unclear. "We have very little idea of what particular mechanism might underlie the slow component of the ventilatory response to hypercapnia"Tansley et al confessed. 

Obviously, the nice linear increase in minute ventilation in response to rising PaCO2 is not something which could extend indefinitely over the range of possible CO2 values. For example, there no conceivable universe where at a PaCO2 of 300 mmHg you will respire to a minute volume of 600L/min. Clearly the curve has to plateau somewhere, perhaps at around the point where the heart stops from hypercapnic acidosis. 

Influence of minute ventilation on CO2

These abovementioned data were all obtained from cruel experiments where volunteers were breathing an artificially CO2-enriched gas mixture. Obviously, the lowest amount of PaCO2 one can contribute to this inspired gas mixture is zero, which is close enough to the normal atmospheric concentration of CO2. This leads to the next important curve to consider. By increasing minute volume, one can decrease the CO2. When the inspired CO2 is minimal, the relationship of CO2 to changing ventilation is a hyperbolic curve, usually described by a graph such as this one (modified from Bergen et al, 1977). 

Changes in CO2 in response to changing ventilation.jpg

This curve (usually referred to as the "metabolic hyperbola")  describes the effects of increasing or decreasing the minute ventilation on the arterial CO2 concentration where the whole body CO2 production remains stable and constant. The point at which the two curves meet is the ventilatory set point. That is, at rest and with a normal metabolic demand one's minute volume and CO2 will trend towards this point. 

Influence of oxygenation on the ventilatory effects of CO2 

As is hinted by the graph above, the sensitivity to CO2 changes depending on the oxygen concentration, as well as many other factors. Hypoxia tends to make you more sensitive to changes in PaCO2, i.e. if one is hypoxic then one's minute volume will be higher at any given CO2 value.  The graphs below illustrate this concept.

changes in CO2 sensitivity with different oxygen concentrations.jpgThe plot on the left is modified from Mohan & Duffin (1997); the plot on the left is a simplified allegory of the same concept which is usually found in physiology textbooks.  The main point to emphasise here is that the point of intercept remains the same, whereas the slope changes. If one is ever asked to describe the influence of oxygen on CO2 sensitivity by an examiner, one could do worse than produce one of these graphs, with this characteristic fan of iso-oxic lines. 

Other factors which influence the ventilatory effects of CO2 

Oxygen is not the only factor which modifies the ventilatory sensitivity to CO2. There are multiple other possible influences. In discussions of CO2 sensitivity, textbooks like to bring out this diagram:

Factors which change the respiratory sensitivity to CO2 

This specific one is based on Figure 9.5 from the Lange book, Pulmonary Physiology by Levitzky (2007, p. 204). It is reproduced everywhere, it usually looks exactly the same, and there are never any references to point to its origins. Individually, one may be able to hunt down articles which demonstrate the change in the CO2 sensitivity associated with sleep (Douglas et al, 1982), opiates (Gross, 2003), anaesthesia (Munson et al, 1966) and metabolic acidosis (Fencl et al, 1966). With a little effort, one can hunt these down to create something of a collage:

changes in CO2 sensitivity with different factors - a collage of published data

Additionally, other factors influence the ventilatory response to CO2 which are usually not mentioned in physiology textbooks, but which are nonetheless important. For example, one such factor is age. Older people tend to have a blunted response to rising CO2, even though their homeostatic setpoint can be relatively normal. Kronenberg & Drage (1973) captured a group of elderly men aged 64-73 and demonstrated that their respiratory drive was weakened by age as compared to a younger cohort; as the data below demonstrate the elderly subjects had increased their minute volume by 1.8L/min per 1mm Hg of CO2 rise, whereas the younger men increased their minute volume by 3.44L/min on average.

changes in respiratory drive associated with age (Kronenberg & Drage, 1973)

Even though one might be tempted to attribute this age-related change in CO2 sensitivity to some sort of nonspecific enfeeblement of old age, it is clear that this is not the case. In fact trained athletes also demonstrate a similar respiratory modification. Byrne-Quinn et al (1971) managed to convince a group of college jocks from Boulder, Colorado to undergo a series of tests in isocapnic hypoxia and iso-oxic hypercapnia, demonstrating that in both circumstances their ventilatory responses were diminished. The controls were selected from Boulder locals who, if the investigators are to be believed, were complete slobs ("None of these had ever engaged in formal physical conditioning and they were generally considered to be sedentary.")

decreased respiratory drive in athletes

In summary, the factors which affect the respiratory response to raised CO2 are:

  • Hypoxia
  • Metabolic acidosis and alkalosis
  • Opiates
  • Sedatives and anaesthetics
  • Natural sleep
  • Age
  • Physical fitness

These factors generally change both the point of intercept (i.e. the ventilatory set point) and the slope of the line (i.e. the sensitivity to CO2). The entire system responds to CO2- not just any specific receptor network or organ system. The response to these factors is therefore the integrated response of the entire organism. Specific contributions made to this response mechanism by each receptor system and the integrated control of these systems are better discussed in a chapter dedicated to control of ventilation and oxygenation.


Haldane, John Scott, and James Lorrain Smith. "The physiological effects of air vitiated by respiration"  Journal of Pathology and Bacteriology, Volume 1, Issue 2 fmi, 123-255 (1892)

Haldane, J. S.; Priestley, J. G.;  "The regulation of the lung-ventilation." Journ. of Physiol 32 (1905).

Schaefer, Karl Ernst. "Respiratory pattern and respiratory response to CO2." Journal of Applied Physiology 13.1 (1958): 1-14.

Berger, Albert J., Robert A. Mitchell, and John W. Severinghaus. "Regulation of respiration." New England Journal of Medicine 297.4 (1977): 194-201.

Mohan, Ravi, and James Duffin. "The effect of hypoxia on the ventilatory response to carbon dioxide in man." Respiration physiology 108.2 (1997): 101-115.

Douglas, Neil J., et al. "Hypercapnic ventilatory response in sleeping adults." American Review of Respiratory Disease126.5 (1982): 758-762.

Pattinson, Kyle TS. "Opioids and the control of respiration." British journal of anaesthesia 100.6 (2008): 747-758.

Gross, Jeffrey B. "When you breathe in you inspire, when you don't breathe, you… expire: new insights regarding opioid-induced ventilatory depression." Anesthesiology: The Journal of the American Society of Anesthesiologists 99.4 (2003): 767-770.

Romberg, Raymonda, et al. "Pharmacodynamic effect of morphine-6-glucuronide versus morphine on hypoxic and hypercapnic breathing in healthy volunteers." Anesthesiology: The Journal of the American Society of Anesthesiologists 99.4 (2003): 788-798.

Munson, Edwin, et al. "The effects of halothane, fluroxene and cyclopropane on ventilation: a comparative study in man." Anesthesiology 27.6 (1966): 716-728.

Fencl, V., T. B. Miller, and Ja R. Pappenheimer. "Studies on the respiratory response to disturbances of acid-base balance, with deductions concerning the ionic composition of cerebral interstitial fluid.American Journal of Physiology-Legacy Content 210.3 (1966): 459-472.

Tansley, John G., et al. "Human ventilatory response to 8 h of euoxic hypercapnia." Journal of Applied Physiology 84.2 (1998): 431-434.

Kronenberg, Richard S., and Charles W. Drage. "Attenuation of the ventilatory and heart rate responses to hypoxia and hypercapnia with aging in normal men." The Journal of clinical investigation 52.8 (1973): 1812-1819.

Byrne-Quinn, Edward, et al. "Ventilatory control in the athlete." Journal of Applied Physiology 30.1 (1971): 91-98.