Factors which influence the affinity of haemoglobin for oxygen


This chapter is most relevant to Section F8(i) from the 2023 CICM Primary Syllabus, which expects the exam candidates to be able to "describe the carriage of oxygen in blood". It covers a part of the syllabus which was encountered in the second half of Question 5 from the second paper of 2015. There are several major physiological factors which influence the affinity of haemoglobin for oxygen. Some of these are under our control. The p50 value as reported by the arterial blood gas analyser presents us with a short-hand way of determining whether the curve has shifted to the right or to the left. Simply put, p50 is the partial pressure of oxygen which is required to saturate 50% of haemoglobin. Thus, an increasing p50 would suggest that the affinity of haemoglobin for oxygen is decreasing - it takes more oxygen to achieve 50% saturation. More on this elsewhere.

In summary,

The following physiological factors influence the affinity of haemoglobin for oxygen:

  • The partial pressure of CO2
    • Increasing CO2 shifts the curve to the right
    • Hyperventilation and hypocapnia shifts the curve to the left
  • pH, independent of CO2
    • Decreasing pH (acidosis) shifts the curve to the right
    • Alkalosis shifts the curve to the left
  • The concentration of 2,3-DPG inside the erythrocytes
    • Increased 2,3-DPG (eg. in response to hypoxia or erythropoietin) shifts the curve to the right
    • Decreased 2,3-DPG (eg. as a red cell storage lesion ) shifts the curve to the left
  • The presence of unusual haemoglobin species
    • Methaemoglobin, carboxyhaemoglobin and foetal haemoglobin shift the curve to the left
    • Sulfhaemoglobin shifts the curve to the right
  • Temperature
    • Hyperthermia shifts the curve right
    • Hypothermia shifts it left

Here is that diagram again, as ubiquitous as the anus:

causes of left and right shift  of the oxyhaemoglobin dissociation curve

Yes, everybody who has gone through the process of marinading their brain in physiology in preparation for some sort of viva exam will be able to reproduce the above-depicted curves, and regurgitate the causes of right and left shift. Those overseeing things from the pedestal of seniority will at least recall the names "Bohr" and "Haldane", with some vague recollection of what those people did.

The safely forgotten principles behind this behaviour of haemoglobin are the topic of discussion for this chapter. The author acknowledges a near-total absence of clinical bedside applicability and intends for this resource to act as a reference for that inevitable point in the future when he, too, will forget everything.

Oxygen-haemoglobin dissociation curve in the absence of allosteric effectors

It is interesting to consider what might happen if one were to de-Bohr the immediate environment of the haemoglobin molecule. The curves below resemble those described in Figure 2.2 in Chapter 2 of Chemistry and Biochemistry of Oxygen Therapeutics by Mozzarelli et al.

oxygen-haemoglobin dissociation curve in the absence of allosteric effectors

Yes, the sigmoid shape is dependent on the presence of allosteric effectors, and that in their absence the oxygen-haemoglobin dissociation curve takes on a boring hyperbolic form, with a p50 of around 10mmHg.

In fact, we can see that its shape then begins to resemble the dissociation curve of myoglobin, the p50 of which is around 2mmHg. This is interesting. Myoglobin is a monomer, and for it there is no allosteric interaction between subunits; it has no quarternary structure from which to derive a sigmoid dissociation curve (nor does it need one - what would it do with it?). In contrast, in haemoglobin there are four subunits which exhibit positive cooperativity in oxygen binding.

Thus, the presence of allosteric effectors is the major influence behind positive cooperativity, and in their absence this effect is largely lost.

The influence of 2,3-DPG concentration on the oxygen-haemoglobin dissociation curve

Firstly, what the hell is 2,3-DPG and where does it come from?

Excellent question.

2,3-diphosphoglycerate is a byproduct of the pathway of glycolysis.

Below, the diagram depicts the pathway of glycolysis, albeit slightly neutered ( boring bits removed or de-emphasised), with shameful jpeg burglary of Wikipedia molecule structures.

involvement of 2,3-DPG in the pathway of glycolysis

So, the pathway of glycolysis would normally take a step from 1,3-DPG to 3-phosphoglycerate, producing a molecule of ATP; or it could bypass that ATP-generating step and produce 2,3-DPG instead. An acidic environment inhibits the activity of diphosphoglycerate mutase and promotes a more "normal" glycolytic pathway, thus inhibiting the production of 2,3-DPG and favouring the production of ATP; this is exploited as a homeostatic mechanism which is discussed later.

The effect of 2,3-DPG on haemoglobin is profound. It is probably the most important allosteric effector of positive cooperativity. In brief, the presence of 2,3-DPG stabilises the T state of deoxyhaemoglobin, decreasing its affinity for oxygen. This was explored in a seminal paper by The Benesches of Columbia University (1967). (All these T and R state changes are discussed elsewhere, and one will not digress here into discussing the inaccuracy of the R-T model and the existence of numerous structural haemoglobin variants.)

Binding of 2,3-DPG to haemoglobin

In the human erythrocyte, 2,3-DPG normally exists in a 5mmol/L concentration, which is approximately the same as the intracellular conceration of haemoglobin. Thus, one 2,3-DPG molecule is all that is required to change the affinity of an entire haemoglobin tetramer. It binds to the central cavity of deoxyhaemoglobin, in a space between the H-helices of the beta-chains. (it can also bind to haemoglobin in its oxygenated R-state, but its affinity for this form is very low).

2,3-DPG binding to haemoglobin

The 2,3-DPG molecule fits snugly into the central cavity between haemoglobin subunits when the haemoglobin is in its deoxygenated state. It happens to have several negative charges on its surface, and it interacts with three positively charged amino acid residues on each β chain, forming an ionic bond. This bond must be broken as the haemoglobin molecule transitions to an oxygenated R state, and the 2,3-DPG molecule ends up being expelled from its cozy little cavity.

The absence of one of the histidines (replaced with serine) in the foetal haemoglobin molecule results in a decreased affinity for 2,3-DPG (the cavity is slightly less cozy). The result is an increased affinity for oxygen, and this accounts for the different shape of the oxygen-HbF dissociation curve:

p50 of foetal haemoglobin

The influence of the Bohr effect on oxygen-haemoglobin binding

Christian Bohr portrait

The effect of decreasing pH (more hydrogen ion activity) on haemoglobin is to stabilise the deoxygenated form, decreasing its affinity for oxygen. This (together with the effects of pCO2) is the principle named after Christian Bohr, a man with a piercing gaze, awesome moustache and multiple Nobel laureate offspring.

The chemical basis for the effect of pH on the oxygen affinity of haemoglobin lies in the amino termini and side chains of two histidine molecules, histidine 146 on the β-subunit and histidine 122 on the α-subunit.

So as not to recapitulate the contents of the chapter on the Bohr effect, it will suffice to say that:

  • An acidic environment facilitates the formation of a bond between one β subunit and the  α subunit of the other αβ dimer.
  • The formation of this bond stabilises the deoxygenated T-state (and oxygenated haemoglobin is unable to form this sort of bond)
  • The formation of this bond is contingent on an acidic pH, which is required to protonate the histidine residue (its pKa is around 7.0).
  • Similarly, the binding of PaCO2 has the effect of stabilising the deoxygenated T-state 

In summary, the effect of low pH (and high PaCO2) is to decrease the affinity of haemoglobin for oxygen.  Most of the time, lecturers will at this stage produce a slide demonstrating the changes in the dissociation curve which occur with "normal" acid-base derangement, in the range of 7.6-7.2. This is sensible, because most people never see a pH outside of this range. However, in the ICU, all sorts of bizarre acid-base disturbances crop up.

Thankfully, Dash and Bassingthwaite have published an excellent article in 2010, presenting a set of oxygen-haemoglobin dissociation curves plotted using a mathematical model which they have developed. Their original diagrams can be found here. They have created a Java applet which runs this model, available from www.physiome.org. With the aid of this software, a crazy person can explore hypothetical dissociation curves for a vast range of completely unsurvivable pH values.

Hypothetical oxygen-haemoglobin dissociation curves for a range of absurd pH values

Yes, you will never see a patient with a pH of 6.0. But if you did, their p50 would be around 67mmHg. By comparison, changes in PaCO2 (within a sensible, survivable range) produce relatively minor changes in oxygen-haemoglobin binding. The combined influence of CO2 and CO2-associated changes in pH on the shape of the oxygen-haemoglobin dissociation curve can be seen in the original 1904 paper by Bohr Hasselbalch and Krogh:

Bohrs original diagram fo oxygen hemoglobin dissociation

The influence of CO2 alone, however, is substantially smaller. Changes in pH produced by CO2 dissociation account for much of the above "curvature".

The combined influence of pH and 2,3-DPG on oxygen-haemoglobin binding

So, a low pH by itself decreases the affinity of haemoglobin for oxygen. However, by inhibiting the production of 2,3-DPG, low pH increases the affinity of haemoglobin for oxygen. The interaction of these two competing actions results in a useful homeostatic mechanism.

homeostatic interaction of 2,3-DPG and pH

In essence, 2,3-DPG opposes the Bohr effect. If acidosis shifts the curve to the right, decreasing the oxygen affinity of haemoglobin, the resulting decrease in 2,3-DPG synthesis shifts the curve to the left again, compensating for the change and maintaining the normal position of the curve.

This was observed in vivo by Bellingham, Detter and Lenfant who fed acetazolamide to four healthy volunteers over the course of a week, measuring their 2,3-DPG levels and oxygen affinity. With a rapid initiation of acidosis, there was no 2,3-DPG concentration change, but over the course of hours and days it began to decrease. The investigators concluded that " these mechanisms shift the hemoglobin oxygen dissociation curve opposite to the direct pH (Bohr) effect, and providing the rate of pH change is neither too rapid nor too large, they counteract the direct pH effect and the in vivo hemoglobin oxygen affinity remains unchanged."

The combined influence of pCO2, pH and 2,3-DPG on oxygen release in the tissues

A well-respected biochemistry textbook reports that

  • A change of oxygen tension from 100mmHg to 20mmHg results in the release of 66% of total carried oxygen
  • When combined with a change from pH 7.4 to 7.2, the released fraction rises to 77%
  • When combined with the presence of 40mmHg CO2, the released fraction rises to 90%.

It is just as well that it is a well-respected textbook, because no reference for these numbers is given, other than the authority of the writers, and a series of calculations comparing haemoglobin to a hypothetical non-cooperative oxygen delivery protein.

The influence of temperature on oxygen-haemoglobin binding

It is known that extreme hypothermia increase the affinity of haemoglobin for oxygen by a massive amount - at 0°C, the affinity is 22 times greater than at 37°C. The mechanism of this is discussed with gusto in an excellent 2011 paper (which encompasses not only man, but "heterothermic vertebrates" in general). The key issue appears to be the thermodynamic properties of the interaction of oxygen with haemoglobin: the reaction appears to be exothermic. Bound by Le Chatelier's principle, this system trends to favouring exothermic reaction as the temperature drops - and the affinity for oxygen increases.

This concept is also one established in the "classical period" of physiology; the first paper discussing it seems to have been published by Barcroft and King in 1909. They analysed the dissociation curves of purified haemoglobin solutions at 14°, 26°, 32° and 38° C. An abridged Figure I from the Barcroft-King paper is reproduced below, and the curves generated by them are extrapolated to fit a curve plotted with Severinghaus' data.

changes of oxygen-haemoglobin affinity in hypothermia

The change appears profound (look, one might say - none of the other curves end up with a p50 of 1mmHg). However, the only situation where one might have a haemoglobin molecule behaving in this way in a living patient is deep hypothermic circulatory arrest, where you cool them down to 20°C. In most other situations, over a "normal" physiological range of temperatures, one only encounters trivial fluctuations in p50 due to temperature.


Bohr, C., K. Hasselbalch, and A. Krogh. "Concerning a biologically important relationship–the influence of the carbon dioxide content of blood on its oxygen binding." Skand. Arch. Physiol 16 (1904): 402.

- this is an English translation of the original article, which was beautifully titled "Über einen in biologischer Beziehung wichtigen Einfluss, den die Kohlensäurespannung  des Blutes auf dessen Sauerstoffbindung übt".

Severinghaus, John W. "Simple, accurate equations for human blood O2 dissociation computations."Journal of Applied Physiology 46.3 (1979): 599-602.

Mills, Frederick C., and Gary K. Ackers. "Quaternary enhancement in binding of oxygen by human hemoglobin." Proceedings of the National Academy of Sciences 76.1 (1979): 273-277.

Eaton, William A., et al. "Evolution of allosteric models for hemoglobin."IUBMB life 59.8‐9 (2007): 586-599.

Mozzarelli, Andrea, and Stefano Bettati, eds. Chemistry and biochemistry of oxygen therapeutics: from transfusion to artificial blood. John Wiley & Sons, 2011.

Benesch, Reinhold, and Ruth E. Benesch. "The effect of organic phosphates from the human erythrocyte on the allosteric properties of hemoglobin."Biochemical and biophysical research communications 26.2 (1967): 162-167.

Bellingham, A. J., J. C. Detter, and C. Lenfant. "Regulatory mechanisms of hemoglobin oxygen affinity in acidosis and alkalosis." Journal of Clinical Investigation 50.3 (1971): 700.

Berg, Jeremy M., John L. Tymoczko, and Lubert Stryer. "Hemoglobin Transports Oxygen Efficiently by Binding Oxygen Cooperatively." in Biochemistry, 5th edition (Jeremy M Berg, John L Tymoczko, and Lubert Stryer (2002).

Benesch, Reinhold, and Ruth E. Benesch. "The Chemistry of the Bohr Effect."The Journal of Biological Chemistry 236.2 (1961).

Dash, Ranjan K., and James B. Bassingthwaighte. "Erratum to: Blood HbO2 and HbCO2 dissociation curves at varied O2, CO2, pH, 2, 3-DPG and temperature levels." Annals of biomedical engineering 38.4 (2010): 1683-1701.

Rossi-Bernardi, L., and F. J. W. Roughton. "The specific influence of carbon dioxide and carbamate compounds on the buffer power and Bohr effects in human haemoglobin solutions." The Journal of physiology 189.1 (1967): 1-29.

Arnone, A. "X-ray studies of the interaction of CO2 with human deoxyhaemoglobin."Nature 247, 143 - 145 (18 January 1974)

Perrella, M., D. Bresciani, and L. Rossi-Bernardi. "The binding of CO2 to human hemoglobin." Journal of Biological Chemistry 250.14 (1975): 5413-5418.

Weber, Roy E., and Kevin L. Campbell. "Temperature dependence of haemoglobin–oxygen affinity in heterothermic vertebrates: mechanisms and biological significance." Acta Physiologica 202.3 (2011): 549-562.

Barcroft, Joseph, and W. O. R. King. "The effect of temperature on the dissociation curve of blood." The Journal of physiology 39.5 (1909): 374-384.

Leach, R. M., and D. F. Treacher. "The pulmonary physician in critical care• 2: Oxygen delivery and consumption in the critically ill." Thorax 57.2 (2002): 170-177.