The carbon dioxide dissociation curve

This chapter is most relevant to Section F8(iv) from the 2023 CICM Primary Syllabus, which expects the exam candidates to be able to "explain the carbon dioxide dissociation curve". Though this has not appeared in any past papers, it is regarded as a staple of the ANZCA primary exams, and is therefore likely to appear in the CICM First Part at some stage. 

In summary:

The CO2 dissociation curve describes the change in the total CO2 content of blood which occurs with changing partial pressure of CO2.

  • This curve is more linear and steep than the oxygen-haemoglobin dissociation curve
  • It has no plateau
  • As the result of this, shunt has little effect on CO2 (increasing the ventilation of already well-ventilated regions will improve CO2 exchange, even though it will not improve oxygenation)

There are two main points of interest along this curve:

  • The arterial point corresponds to the CO2 content of arterial blood:
    • PCO2 =  40 mmHg
    • CO2 content is 480 ml/L (or, 48ml/dL)
  • The mixed venous point: CO2 content of mixed venous blood:
    •  PCO2 is 46 mmHg
    • CO2 content is 520 ml/L. Because of the Haldane effect, if this blood were to be "arterialised" by the addition of oxygen while the total CO2 content remained the same,  the extra CO2 liberated by the oxygenation of haemoglobin would produce an increase in PCO2 to something like 55 mmHg.

The physiological CO2 dissociation curve is a line which connects the venous and arterial points, and represents the normal physiological progression of blood on the way through the circulation

The carbon dioxide dissociation curve and its various iterations are encountered in the chapter on transport of carbon dioxide in the blood and the Haldane Effect, but one does not dwell on it in those instances because those sections have their own priorities. The main point of this chapter is to arm the exam candidates with enough of the basics that, when they are confronted with a viva question on this subject in some darkest timeline, they will confidently sketch the curve and hold forth on its properties like it's nothing. 

The carbon dioxide dissociation curve

This is a graphical repesentation of the CO2 content of whole blood, plotted against PCO2. By some bizarre convention, everybody (Brandis, Part OneNunn's) graph the content in ml/dL, which is completely different to the way we measure it in virtually every other situation (this may be another weird feature of local culture, but in Australia everything seems to be in ml/L). 

If one were so inclined, one could easily reproduce the simple version of this curve using an equation. This specific empiric fudge comes from a letter to the editor written by Felix Meade in 1972 ("Sir—some of your reader will be using computers to facilitate the processing of physiological data and will have experienced inconvenience and limitation due to lack of a suitable mathematical expression for the carbon dioxide dissociation curve"). Meade fit the formula to empiric data collected by others, and presented comparison tables which demonstrate its validity over a range of 10-80 mmHg. And yes, the content of whole blood is in ml/L. The author is please to report that he was spared any inconvenience and limitation in using a computer (Google Sheets) to facilitate the processing of this physiological data:

Carbon dioxide dissociation curve - plain

This, in fact, is just a plain version of what you'd have already seen when reading about the different forms of CO2 transport in the blood, as this diagram (below) appears in legitimate publications as well as here:

From the perspective of scoring points and passing exams, it is of course not essential to plot the graph from a PCO2 of 10 to a PCO2 of 80 mmHg, as all the money is really in the 40-46 range:

There are important elements here which one needs to mention:

  • The arterial point:  this corresponds to the CO2 content of arterial blood, where the PCO2 is 40 mmHg and the CO2 content is 480 ml/L (or, 48ml/dL if you are that way inclined). For extra marks, one might even extend the discussion of this to show that on the corresponding venous curve, at 40mmHg PCO2 the total CO2 content would be something closer to 500 ml/L (50ml/dL). One would have to mention that this is due to the Haldane effect, where deoxyhaemoglobin has a higher affinity for CO2
  • The mixed venous point: this corresponds to the CO2 content of mixed venous blood, where the PCO2 is 46 mmHg and the CO2 content is 520 ml/L. Because of the Haldane effect, if this blood were to be "arterialised" by the addition of oxygen while the total CO2 content remained the same,  the extra CO2 liberated by the oxygenation of haemoglobin would produce an increase in PCO2 to something like 55 mmHg.

The mixed venous point and the arterial point are viewed as the most relevant points of interest, but theoretically one could also include a curve with  SpO2=0%. That is how it was originally presented when the phenomenon was first reported in 1914 by Christiansen Douglas and Haldane. The authors has plotted both the 100% and the 0% curves over a broad range of carbon dioxide partial pressures. This is the graph from their original article:

CO2 dissociation curve of JS Haldane's blood

Note how, in a classic Haldane move, the authors drained hundreds of samples from Haldane himself to provide the experiments with enough blood (3-4ml at a time).

The "physiological" CO2 dissociation curve

Occasionally, in textbooks the arterial and venous points are joined by a line. It might not be immediately clear as to what the point of this is, but with a little digging one can find that the resulting relationship is occasionally referred to as the "physiological CO2 dissociation curve" or the "total CO2 dissociation curve".

physiological CO2 dissociation curve

This curve describes what happens to the CO2 content of blood during its natural physiological progression from the venous state to the arterial state, and then back again. It is usually appended to the middle of a chapter that explains the Haldane effect, and is perhaps meant to illustrate it, in the sense that the Haldane effect takes part in the change in the content/pressure relationship between the arterial and venous blood. However, it is not the whole story, and it does not have a lot of illustrative or explanatory power (for example, one would not be able to understand the Haldane effect just from looking at this graph).

A comparison of oxygen and carbon dioxide dissociation curves

Both gases have dissociation curves, and so it probably makes sense to compare them on the same coordinate field:

comparison of oxygen and carbon dioxide dissociation curves

What is the point of doing this? As one can plainly see, the CO2 dissociation curve is more linear and much steeper:

  • When CO2content increases substantially,  partial pressure of CO2 does not change very much
  • There is no plateau (it just keeps going up)

 What are the implications of this? Well. The sigmoid shape (specifically, the plateau) of the oxygen-haemoglobin dissociation curve gives rise to a phenomenon whereby it is impossible to compensate for a shunt. As blood travelling through well-ventilated regions of the lung is already maximally oxygenated (i.e. lays along on the plateau), there is no way you can get any better oxygenation by increasing ventilation. In contrast, because the CO2 dissociation relationship is more linear, increasing ventilation of the already well-ventilated regions will still continue to improve the CO2 clearance from those regions, and increasing the minute ventilation can compensate for even very large shunt (up to 50%, apparently). 


Meade, Felix. "A formula for the carbon dioxide dissociation curve." BJA: British Journal of Anaesthesia 44.6 (1972): 630-630.

Defares, J. G., and B. F. Visser. "On the form of the physiological CO2 dissociation curve, the physiological O2 dissociation curve, and the diffusion curves of O2 and CO2 along the capillary path." Annals of the New York Academy of Sciences 96.4 (1962): 939-955.

Teboul, Jean-Louis, and Thomas Scheeren. "Understanding the Haldane effect.Intensive care medicine 43.1 (2017): 91-93.

Christiansen, Johanne, C. G. Douglas, and J. S. Haldane. "The absorption and dissociation of carbon dioxide by human blood." The Journal of physiology 48.4 (1914): 244-271.

Klocke, ROBERT A. "Mechanism and kinetics of the Haldane effect in human erythrocytes." Journal of applied physiology 35.5 (1973): 673-681.