Arterial blood gas measurements often show that pO2 is 'normal' even though haemoglobin is bound to carbon monoxide with high affinity.
Is this because there is still oxygen bound to some subunits of the haemoglobin? Surely the CO will displace at least some oxygen, reducing the pO2? As oxygen solubility in plasma is very low, where would oxygen be in the blood if not bound to haemoglobin if there was still high pO2?
First, let's think about what pO2 means:
pO2 is not the same as "oxygen concentration." pO2 is the partial pressure of oxygen, you can think of it as proportional to the "number of oxygen molecules per unit area leaving the solution" (or colliding with the side of the container). There is also pO2 for the atmosphere, which is going to be approximately 21% * 760mmHg = about 160 mmHg. Partial pressures are often a convenient way to measure gas concentrations, but that only works when there aren't any other interactions happening, so it's not valid for blood.
When blood comes in contact with inspired air, there are >100 mmHg of oxygen molecules hitting the solution and <100 mmHg of oxygen leaving, so there is a net diffusion of oxygen from the inspired air to the blood. If the gas in the alveoli and the blood in the lung capillaries have the same partial pressure of oxygen, there is no more net transfer of oxygen, they are at equilibrium.
Hemoglobin can't do anything to cause the blood to carry more "partial pressure" of oxygen. What hemoglobin does is to "hold on" to oxygen molecules. These "held" molecules can't leave the solution, so they don't "count" as much towards the partial pressure. Hemoglobin instead allows for a higher concentration of oxygen at the same partial pressure of oxygen. Another special thing hemoglobin does is to change its affinity for oxygen in the acidic environment of metabolically active tissue, releasing O2 at a higher partial pressure than it was initially dissolved at.
If you pumped plain saline through the lungs, it would get to about the same partial pressure of oxygen, but it wouldn't have as many oxygen molecules.
If not enough fresh air is entering the lungs, or if the air entering the lungs is low in oxygen, you would expect the blood pO2 to drop, but this isn't what happens in CO poisoning.
There is also the issue that what actually happens in CO poisoning is that affinity of hemoglobin for oxygen increases, but this is actually irrelevant to the pO2 question. The oxygen tightly bound to carboxyhemoglobin (CO-poisoned hemoglobin) doesn't "count" towards the pO2.
@Bryan's answer is good. I'd like to add a little structure around what seems to be the primary confusion in your question.
There are three important measurements here:
Arterial $P_{O2}$ ($P_{aO_2}$) is typically expressed in mm Hg. It is very specifically a measurement of the concentration of oxygen dissolved in arterial blood and NOT bound to hemoglobin. Low $P_{aO_2}$ is called hypoxemia.
$O_2$ saturation is a measurement of the percentage of Hemoglobin binding sites occupied by oxygen. It is expressed as a percentage, e.g., 98%.
Hemoglobin concentration is a measurement of the concentration of hemoglobin protein in blood. It is typically expressed in the U.S. in g/dL.
Together, these three measurements ($P_{aO_2}$, $O_2$ saturation, and Hb concentration) determine the total oxygen content of arterial blood. Any one or any combination of these three measurements can result in hypoxia (inadequate oxygenation of tissue).
Carbon monoxide poisoning in an otherwise healthy individual is a specific cause of hypoxia that has a characteristic pattern: normal $P_{aO_2}$, normal Hemoglobin concentration, and low $O_2$ saturation. Because there is no problem breathing in oxygen or having it cross the blood gas barrier, there is a normal level of oxygen dissolved in arterial blood. $P_{aO_2}$ is normal. There is no anemia, so Hb concentration is normal. The problem is strictly a matter of $O_2$ saturation.
In many cases of hypoxia, the picture is mixed, but $CO$ poisoning is a classic example where the problem is exclusively one of $O_2$ binding. Anemia is another example. In anemia, $P_{aO_2}$ is normal, $O_2$ saturation is normal, but total oxygen content remains low because there is not enough Hemoglobin.
These are fairly complicated concepts, but covered quite well in West Respiratory Physiology, Ch. 6. The following table may be helpful. Note the footnote for $O_2$ saturation in the row for CO poisoning. Saturation is normal when calculated for Hemoglobin not bound to CO, in other words, there is nothing wrong with the Hb molecule itself. It is simply bound to something else. This is why administration of 100% $O_2$ is a useful treatment.
