How much oxygen is in the blood?

The Differences Between PaO2, SaO2 and Oxygen C

ontent.
In the field of blood gas interpretation, confusion about PaO2, SaO2 and oxygen
content is second only to confusion about mixed acid-base disturbances.
Arterial PO2 (little 'a')gives us valuable information about adequacy of gas exc
hange within the lungs, when (and only when) it is subtracted from the calculate
d alveolar PO2 (big A). We use the Alveolar Gas Equation to calculate PAO2. The
difference between measured PaO2 and calculated PAO2 is called the Alveolar-arte
rial PO2 difference or 'A-a Gradient' for short. The A-a gradient answers the im
portant question: Are the lungs transferring oxygen properly from the atmosphere
to the pulmonary circulation? If the A-a gradient is elevated, the answer is NO
. If the A-a gradient is normal is YES. (The A-a gradient is discussed in detail
in Chapter 4).
There is a second, equally important question concerning oxygen and gas exchange
, which is the subject of this section:
How much oxygen is in the blood, and is it adequate for the patient?
The answer here must obviously be based on some oxygen value, but which one? Aft
er all, blood gases give us three different oxygen values:
PaO2
SaO2
Oxygen content (CaO2)
Of these three values, PaO2, or oxygen pressure, is the least helpful to answer
the question about oxygen adequacy in the blood. The other two values --oxygen s
aturation and oxygen content -- are more useful for this purpose. I will briefly
define these three terms and then present a more detailed discussion of each, w
ith emphasis on their inter-relationships.
OXYGEN PRESSURE: PaO2.
Oxygen molecules dissolved in plasma (i.e., not bound to hemoglobin) are free to
impinge on the measuring oxygen electrode. This "impingement" of free O2 molecu
les is reflected as the partial pressure of oxygen; if the sample being tested i
s arterial blood, then it is the PaO2. Although the number of O2 molecules disso
lved in plasma determines, along with other factors, how many molecules will bin
d to hemoglobin, once bound the oxygen molecules no longer exert any pressure (b
ound oxygen molecules are no longer free to impinge on the measuring electrode).
Since PaO2 reflects only free oxygen molecules dissolved in plasma and not thos
e bound to hemoglobin, PaO2 cannot tell us "how much" oxygen is in the blood; fo
r that you need to know how much oxygen is also bound to hemoglobin, information
given by the SaO2 and hemoglobin content.
OXYGEN SATURATION: SaO2.
Binding sites for oxygen are the heme groups, the Fe++-porphyrin portions of the
hemoglobin molecule. There are four heme sites, and hence four oxygen binding s
ites, per hemoglobin molecule. Heme sites occupied by oxygen molecules are said
to be "saturated" with oxygen. The percentage of all the available heme binding
sites saturated with oxygen is the hemoglobin oxygen saturation (in arterial blo
od, the SaO2). Note that SaO2 alone doesn't reveal how much oxygen is in the blo
od; for that we also need to know the hemoglobin content.
OXYGEN CONTENT: CaO2.
Tissues need a requisite amount of O2 molecules for metabolism. Neither the PaO2

nor the SaO2 provide information on the number of oxygen molecules, i.e., of ho
w much oxygen is in the blood. (Note that neither PaO2 nor SaO2 have units that
denote any quantity.) Of the three values used for assessing blood oxygen levels
, how much is provided only by the oxygen content, CaO2 (units ml O2/dl). This i
s because CaO2 is the only value that incorporates the hemoglobin content. Oxyge
n content can be measured directly or calculated by the oxygen content equation
(introduced in Chapter 2):
CaO2 = Hb (gm/dl) x 1.34 ml O2/gm Hb x SaO2 + PaO2 x (.003 ml O2/mm Hg/dl).
* * *
More on the definitions and distinctions of PaO2, SaO2 and CaO2
You wish it was this simple, huh? I have shown the 3 short paragraphs above to d
ozens of students, interns, residents; almost all will say they understand the d
ifferences, no problem. But, when given questions to test their understanding, t
hey flub. So more instruction is needed (and, yes, a few problems along the way)
. Understanding the differences between PaO2, SaO2 and CaO2 is essential to prop
er blood gas interpretation. By the end of this and the next chapter -- if you w
ork on all the problems -- you should be able to teach the subject!
PaO2, the partial pressure of oxygen in the plasma phase of arterial blood, is r
egistered by an electrode that senses randomly-moving, dissolved oxygen molecule
s. The amount of dissolved oxygen in the plasma phase -- and hence the PaO2 -- i
s determined by alveolar PO2 and lung architecture only, and is unrelated to any
thing about hemoglobin. (With one exception: when there is both anemia and a siz
able right to left shunt of blood through the lungs. In this situation a suffici
ent amount of blood with low venous O2 content can enter the arterial circulatio
n and lead to a reduced PaO2. However, with a normal amount of shunting, anemia
and hemoglobin variables do not affect PaO2.)
Oxygen molecules that pass through the thin alveolar-capillary membrane enter th
e plasma phase as dissolved (free) molecules; most of these molecules quickly en
ter the red blood cell and bind with hemoglobin (Figure 5-1). There is a dynamic
equilibrium between the freely dissolved and the hemoglobin-bound oxygen molecu
les. However, the more dissolved molecules there are (i.e., the greater the PaO2
) the more will bind to available hemoglobin; thus SaO2 always depends, to a lar
ge degree, on the concentration of dissolved oxygen molecules (i.e., on the PaO2
).
Figure 5-1. Oxygen pressure, saturation and content. Schematic shows cross secti
on of lungs and pulmonary circulation. (CO2, nitrogen and other gas molecules ar
e omitted for clarity.) PaO2 is always slightly lower than PAO2 because of norma
l venous admixture, here represented by a connection between the venous and pulm
onary circulations. See text for discussion. Click on figure to obtain larger im
age.
In this figure:
Hemoglobin content = 15 gm/dl
Alveolar partial pressure of oxygen (PAO2) = 102 mm Hg
Venous partial pressure of oxygen (PvO2) = 40 mm Hg
Venous hemoglobin oxygen saturation (SvO2) = 75%
Arterial partial pressure of oxygen (PaO2) = 95 mm Hg
Arterial hemoglobin oxygen saturation (SaO2) = 97%
Because there is a virtually unlimited supply of oxygen molecules in the atmosph
ere, the dissolved O2 molecules that leave the plasma to bind with hemoglobin ar
e quickly replaced by others; once bound, oxygen no longer exerts a gas pressure

. Thus hemoglobin is like an efficient sponge that soaks up oxygen so more can e
nter the blood. Hemoglobin continues to soak up oxygen molecules until it become
s saturated with the maximum amount it can hold - an amount that is largely dete
rmined by the PaO2. Of course this whole process is near instantaneous and dynam
ic; at any given moment a given O2 molecule could be bound or dissolved. However
, depending on the PaO2 and other factors, a certain percentage of all O2 molecu
les will be dissolved and a certain percentage will be bound (Figure 5-1). In Fi
gure 5-1, the free or dissolved oxygen molecules register a partial pressure of
95 mm Hg and the red blood cells contain a total hemoglobin content of 15 gm/dl.
Each hemoglobin molecule has four Fe++heme sites for binding oxygen. If there is
no interference (as from carbon monoxide, for example), the free O2 molecules b
ind to these sites with great avidity. The total percentage of sites actually bo
und with O2 is constant for a given set of conditions, and is the 'saturation of
blood with oxygen'. This is called SvO2 and SaO2 in the venous and arterial cir
culations, respectively; in Figure 5-1, the respective values are 75% and 97%. A
n SaO2 of 97% simply means that of every 100 hemoglobin binding sites, 97 are oc
cupied with an oxygen molecule and the other three are either bound to something
else or are unbound.
In summary, PaO2 is determined by alveolar PO2 and the state of the alveolar-cap
illary interface, not by the amount of hemoglobin available to soak them up. PaO
2, in turn, determines the oxygen saturation of hemoglobin (along with other fac
tors that affect the position of the O2-dissociation curve, discussed below). Th
e SaO2, plus the concentration of hemoglobin (15 gm/dl in this example), determi
ne the total amount of oxygen in the blood or CaO2 (see equation for CaO2). For
the variables shown in Figure 5-1, the CaO2 is 20 ml O2/dl.
Clinical Problem 5-1. At 10 a.m. a patient has a PaO2 of 85 mm Hg, an SaO2 of 98
%, and a hemoglobin of 14 gm/dl. At 10:05 a.m. she suffers a severe hemolytic re
action that suddenly leaves her with a hemoglobin of only 7 gm/dl. Assuming no l
ung disease occurs from the hemolytic reaction, what will be her new PaO2, SaO2,
and CaO2?
(NOTE: Answers to all Clinical Problems are provided at end of this section.)
a)

PaO2 unchanged, SaO2 unchanged, CaO2 unchanged

b)

PaO2 unchanged, SaO2 unchanged, CaO2 reduced

c)

PaO2 reduced, SaO2 unchanged, CaO2 reduced

d)

PaO2 reduced, SaO2 reduced, CaO2 reduced

From the forgoing discussion the following observations should now be apparent.
the less hemoglobin available to bind the dissolved oxygen molecules, the fewer
total number of oxygen molecules will the blood contain; and
the more hemoglobin available to bind the dissolved oxygen molecules, the greate
r total number of oxygen molecules will the blood contain.
Neither the amount of hemoglobin, nor the binding characteristics of hemoglobin,
should affect the amount of dissolved oxygen, and hence should not affect the P
aO2). Stated another way, the number of dissolved oxygen molecules is independen
t of the amount of hemoglobin or what is bound to it. To repeat one more time (b
ecause it is so important), PaO2 is not a function of hemoglobin content or of i
ts characteristics, but only of the alveolar PO2 and the lung architecture (alve
olar-capillary interface). This explains why, for example, patients with severe
anemia or carbon monoxide poisoning or methemoglobinemia can (and often do) have
a normal PaO2.

The most common physiologic disturbance of lung architecture, and hence of a red
uced PaO2, is ventilation-perfusion (V-Q) imbalance. Less common causes are redu
ced alveolar ventilation, diffusion block, and anatomic right to left shunting o
f blood.
Clinical Problem 5.2 State which of the following situations would be expected t
o lower PaO2.
a)

anemia.

b)

carbon monoxide toxicity.

c)
an abnormal hemoglobin that holds oxygen with half the affinity of norma
l hemoglobin.
d)
an abnormal hemoglobin that holds oxygen with twice the affinity of norm
al hemoglobin.
e)

lung disease with intra-pulmonary shunting.

SaO2 is determined mainly by PaO2. The relationship between the two variables is
the familiar oxygen dissociation curve (Figure 5-2A). The dissociation curve is
experimentally determined from in vitro titration of blood with increasing part
ial pressures of oxygen. At low oxygen pressures there is relatively little incr
ease in SaO2 for a given change in PaO2. Above a PaO2 of 20 mm Hg, the rate of c
hange of SaO2 increases markedly, then slows again beyond a PaO2 of 60 mm Hg.
PaO2 is the most important (but not the only) determinant of SaO2. Other determi
nants of SaO2 for a given PaO2 are conditions that shift the position of the oxy
gen dissociation curve left or right, such as temperature, pH, PaCO2 and level o
f 2,3-DPG in the blood. Shifts of the O2-dissociation curve will be discussed fu
rther in the next chapter.
For now, consolidate your understanding of the difference between PaO2 and SaO2.
Think of PaO2 as the driving pressure for oxygen molecules entering the red blo
od cell and chemically binding to hemoglobin; the higher the PaO2, the higher th
e SaO2. Whatever the SaO2, its value is simply the percentage of total binding s
ites on arterial hemoglobin that are bound with oxygen, and can never be more th
an 100%.
Figure 5-2.
The oxygen dissociation curve, showing PaO2 vs. SaO2 and PaO2 vs. oxygen content
for two different hemoglobin values. P50 is the PaO2 at which hemoglobin is 50%
saturated with oxygen; normal value is 27 mm Hg. (X represents blood gas values
of a case presented in Chapter 6).
The Arterial oxygen content is shown for two hemoglobin vaues, 15 gm/dl and 10 g
m/dl. The relationship between SaO2 and CaO2 for any given hemoglobin content is
linear (excluding the minor influence of dissolved oxygen with normal PO2 value
s).
Clinical Problem 5-3. Using Figure 5-2 to determine SaO2, calculate O2 content o
f a patient with hemoglobin 12 gms/dl, PaO2 50 mm Hg, pH 7.40.
The so-called "steep part" of the O2 dissociation curve is between 20 and 60 mm
Hg PaO2. Compared with the flatter portions, small increases in PaO2 in this reg
ion have a much greater effect on improving SaO2 and therefore O2 content. Figur
e 5-2A shows the oxygen dissociation curve for PaO2 plotted against oxygen conte

nt for two hemoglobin concentrations, 15 gm% and 10 gm%. Note that the shape and
position of the curve are the same irrespective of the hemoglobin content.
SaO2 is unaffected by the hemoglobin content, so anemia does not lower SaO2. The
more hemoglobin, the more oxygen molecules will be bound in a given volume of b
lood, but the percentage of available hemoglobin sites bound to oxygen (the SaO2
) depends only on the PaO2 and curve-shifting factors. Thus, a patient can have
a normal PaO2 and SaO2, but still have a low CaO2 (e.g., with anemia).
CaO2, unlike either PaO2 or SaO2, directly reflects the total number of oxygen m
olecules in arterial blood, both bound and unbound to hemoglobin. In contrast to
the other two variables, CaO2 depends on the hemoglobin content and is directly
related to it; other determinants of CaO2 are the SaO2 (in turn dependent on Pa
O2 and position of the oxygen dissociation curve), and the amount of dissolved o
xygen (the PaO2). Since the dissolved oxygen contributes minimally to CaO2 under
physiologic conditions, CaO2 is determined almost entirely by hemoglobin conten
t and SaO2, and is related linearly to either variable (Figure 5-2B).
Normal CaO2 ranges from 16 to 22 ml O2/dl. Because PaO2 and/or SaO2 can be norma
l in certain conditions associated with hypoxemia, one should always make sure C
aO2 is adequate when assessing oxygenation. About 98% of the normal O2 content i
s carried bound to hemoglobin.
The CaO2 component bound to hemoglobin can be calculated by (Hb x 1.34 x SaO2) a
nd the dissolved component by (.003 x PaO2). This equation can be used to calcul
ate oxygen content of any blood or plasma sample.
Figure 5-3 shows two beakers containing liquid open to the atmosphere. Beaker 1
contains blood with a Hb content of 15 grams%. Beaker 2 contains only plasma (no
hemoglobin). Assuming a barometric pressure of 760 mm Hg (and no water vapor pr
essure), calculate the oxygen content in each beaker.
Figure 5-3.
Beaker 1 contains blood with a hemoglobin content of 15 grams%; beaker 2 contain
s pure plasma, no hemoglobin. Both beakers are open to the atmosphere (dry air;
barometric pressure 760 mm Hg).
Beaker 1 contains hemoglobin that will combine chemically with oxygen; hence the
oxygen content in beaker 1 consists of bound and unbound (dissolved) oxygen mol
ecules. In beaker 2 there is no hemoglobin, just pure plasma; all of its oxygen
content must come from dissolved oxygen.
Dissolved oxygen in both beakers is determined by the PO2 to which the liquid is
exposed and the solubility of oxygen in plasma. The solubility is .003 ml O2/dl
plasma/mm Hg. But what is the PO2? Because there is no CO2 exchange taking plac
e in either beaker (as there is in our lungs) and the surface of the liquid is i
n free contact with the atmosphere, the PO2 in solution is simply the PO2 above
the solution. Given a barometric pressure of 760 mm Hg (dry air), the PO2 in bot
h beakers is
FIO2 x PB = .21 x 760 mm Hg = 160 mm Hg.
Since the PO2 is equal in both beakers, the O2 content represented by dissolved
oxygen is also the same in both beakers; this content is
?
a) .48 ml O2/dl
b) 2.0 ml O2/dl

c) 4.8 ml O2/dl
To calculate content from dissolved oxygen, substitute the values for oxygen sol
ubility and PO2:
O2 content of dissolved O2 = .003 ml O2/dl/mm Hg x 160 mm Hg
= .48 ml O2/dl
There is no hemoglobin in beaker 2 so the entirety of its O2 content comes from
dissolved oxygen and = .48 ml O2/dl. There is far more oxygen content in beaker
1 because oxygen molecules combine chemically with hemoglobin. Once combined, O2
molecules no longer exert any pressure. As O2 molecules are taken up by hemoglo
bin, additional molecules enter the plasma portion of the blood from the atmosph
ere. (Remember: Hemoglobin is like a sponge that soaks up free oxygen molecules
and allows many more to enter the surrounding plasma.) Thus the difference in ox
ygen content between the two beakers is the amount of oxygen bound to hemoglobin
.
The oxygen content represented by hemoglobin-bound oxygen in beaker 1 is
?
a) .48 ml O2/dl
b) 15 ml O2/dl
c) 19.9 ml O2/dl
O2 content is calculated by the oxygen content equation, which in turn requires
knowledge of SaO2, the saturation of hemoglobin with oxygen. SaO2 is determined
by the PO2 to which the blood is exposed in the lungs (in this case 160 mm Hg) a
nd the position of the oxygen dissociation curve. With a normally-positioned cur
ve, the SaO2 at this level of PO2 is approximately 99%. Thus,
Oxygen content (Hb-bound)

= Hb x 1.34 x SaO2

= 15 x 1.34 x .99
= 19.9 ml O2/dl
?
What is the total oxygen content of beaker 1? By what factor is this content gre
ater than that in beaker 2?
The total oxygen content of beaker 1 is of course the sum of the dissolved and b
ound fractions, or .48 + 19.9 = 20.38 ml O2/dl. The total oxygen content of beak
er 2 (.48 ml O2/dl) is thus only about 2.4% of that contained in beaker 1. Put a
nother way, beaker 1 contains about 42 times more oxygen than beaker 2.
Clinical Problem 5-4. A healthy man is in the same room as the two beakers shown
in Figure 5-3. If his PaO2 = 100 mm Hg and Hb content = 15 gms%, what percent o
f his oxygen content is carried in dissolved form?
To summarize much of the forgoing discussion:
Although almost all of the oxygen content is chemically bound to hemoglobin, thi
s quantity is unrevealed by knowing only the PaO2.

Without knowledge of the hemoglobin content, the PaO2 does not even give a hint
of the total oxygen content.
We need to calculate CaO2 to know the amount of oxygen in the blood.
Because the body needs a requisite oxygen content for survival, and PaO2 alone d
oes not indicate oxygen content, a patient can have normal PaO2 and be starved f
or oxygen.
Clinical Problem 5-5. For each of the four conditions below, give the expected c
hanges (increased, decreased, or normal) for PaO2, SaO2 and CaO2. Assume the sub
ject is breathing ambient air, that each situation occurred acutely (in less tha
n 24 hours), and that there is no other abnormal condition.
CONDITION
PaO2
SaO2
CaO2
Severe Anemia
CO Poisoning
Severe V-Q Imbalance

High Altitude

Clinical Problem 5-6. Which patient is more hypoxemic?

Patient A: PaO2 85 mm Hg, SaO2 95%, Hb 7 gm%
Patient B: PaO2 55 mm Hg, SaO2 85%, Hb 15 gm%
Clinical Problem 5-7. Test your understanding by answering the following stateme
nts a-h as either True or False.
a.
If the lungs and heart are normal, then PaO2 is affected only by the alv
eolar PO2.
b.
2.

In a person with normal heart and lungs, anemia should not lower the PaO

c.
PaO2 will go up in a patient with hemolysis of red blood cells, as disso
lved oxygen is given off when the cells lyse.
d.
As the oxygen dissociation curve shifts to the right, PaO2 rises since l
ess oxygen is bound to hemoglobin.
e.
An anemic patient who receives a blood transfusion should experience a r
ise in both SaO2 and CaO2.
f.
The PaO2 in a cup of water is zero since there is no blood perfusing the
water.
g.

The SaO2 in a cup of water is zero since there is no hemoglobin present.

h.

The CaO2 in a cup of water is zero since there is no hemoglobin present.