What is the oxyhemoglobin dissociation curve and why is it important?

  • The Oxyhemoglobin dissociation curve describes the non-linear tendency for oxygen to bind to hemoglobin: below a SaO2 of 90%, small differences in hemoglobin saturation reflect large changes in PaO2

The oxyhemoglobin dissociation curve mathematically equates the percentage saturation of hemoglobin to the partial pressure of oxygen in the blood. The strange sigmoid shape of the curve relates to peculiar properties of the hemoglobin molecule itself:

Hemoglobin and oxygen act a little like parents and children. When all are living at home (i.e. hemoglobin is fully saturated) then the parents don’t want any to leave: but once one has flown the nest (i.e. dissociated from hemoglobin) – parents find it progressively easier to let go. What this means that the conformation of the hemoglobin molecule depends on the number of molecules bound: as one molecule of oxygen becomes unbound, the affinity for the others falls [and vice-versa]. This is represented by the oxyhemoglobin dissociation curve.

The lack of linearity of the curve makes interpretation of the oxygen content of blood difficult. At higher saturation levels, above 90%, the curve is flat, but below this level the PaO2 declines sharply, such that at 75% saturation the PaO2 is about 47mmHg (mixed venous blood), at 50% saturation the PaO2 is 26.6mmHg, and at 25% saturation the PaO2 is a miserable 15mmHg.


The oxyhemoglobin dissociation curve

The position of this curve may shift rightwards (lower saturation for given PaO2) or leftwards (higher saturation for a given PaO2). Certain circumstances make the blood more likely to dump oxygen into the tissues, and others make it more likely to cling on to oxygen. Active muscle needs more oxygen, so heat, exercise, acidosis, hypercarbia and increased 2,3-DPG all cause a shift in the curve rightwards – releasing oxygen. Conversely, when activity is minimal – such as in cold weather or during rest, when the tissues are cold, alkalotic, hypocarbic and low 2,3-DPG, then hemoglobin holds onto oxygen. The curve also shifts leftwards in carbon monoxide poisoning.

The oxygen dissociation curve is an essential component in understanding critical care medicine. Everything we do is about optimizing the delivery of blood to the tissues as a means of maintaining homeostasis and promoting healing, and in the end it is the oxygen content of blood that is more important than the partial pressure of oxygen (which we commonly measure). The oxygen content relates specifically to the amount of hemoglobin present and how saturated it is. A reduction in the hemoglobin concentration from 15 to 10g/dl reduces the arterial oxygen content (CaO2) by as much as a reduction in PaO2 from 100mmHg to 40mmHg. Moreover a small drop in SaO2 may represent a large drop in PaO2, due to the shape of the oxyhemoglobin dissociation curve: when hemoglobin is 50% saturated the PaO2 is 28mmHg, at 75% the PaO2 is about 40mmHg (mixed venous blood).

Although many pages of critical care textbooks are often devoted to discussions about oxygen delivery, there is no clear indication what the optimal hemoglobin actually is. We know from the TRICC (transfusion requirements in critical care) study (1) that transfusing patients with blood above a hemoglobin of 7.0g is probably harmful. This probably relates to problems associated with the actual process of storing and transfusing products, than the effect of a “normal” hemoglobin itself. The availability of therapeutic erythropoietin has allowed intensivists induce red cell production, and replace blood mass without external transfusion.

In any case, there is a large physiologic reserve between oxygen delivery and oxygen consumption.  The cardiac output is probably a bigger player in the delivery of O2 to the tissues that the O2 content. This is because the cardiac output can almost instantaneously respond to a fall in PaO2 saturation of Hb. Moderate hypoxemia leads to an increase in the cardiac output and a reduction in peripheral vascular resistance. On the other hand, compensation for a fall in cardiac output is slow and weak – that is because it takes time to increase Hb production and the oxyhemoglobin dissociation curve is flat – it can’t become anymore saturated. Nevertheless, in the clinical setting, it is often easier to increase the Hb or the FiO2 than to increase the cardiac output


   (1)    Hebert PC. Transfusion requirements in critical care (TRICC): a multicentre, randomized, controlled clinical study. Transfusion Requirements in Critical Care Investigators and the Canadian Critical care Trials Group. Br J Anaesth 1998; 81 Suppl 1:25-33.


Copyright 2002 All rights reserved