Diffusion Capacity for Carbon Monoxide

This article is part of a series on USMLE Step 1 Study.

The diffusion capacity of the lung for carbon monoxide (DLCO) is an important test of lung function. I’ve experienced some confusion as I have worked towards understanding this concept, so I wanted to share what I’ve learned.

First, a question:

A 62 year old woman with COPD is evaluated for chronic dyspnea. She has long history of smoking. Respiratory rate is 30; pulse 102; the patient is afebrile. On exam she is cachectic, with distant heart sounds, increased AP diameter of the chest and expiratory wheezing.

What is the most likely finding on pulmonary function test?

  1. Non reversible obstructive pattern with reduced DLCO.
  2. Non reversible obstructive pattern, DLCO within normal.
  3. Reversible obstructive pattern with normal DLCO.
  4. Restrictive pattern.
  5. FEV1/FVC 90% of normal.

The answer is 1. Elements of the history and physical, including the smoking history and increased chest diameter, indicate emphysema. Emphysema is obstructive (as we’re told); it’s irreversible; and DLCO is reduced.

We can use DLCO to distinguish between irreversible obstructive diseases: DLCO is reduced in emphysema and normal in chronic bronchitis. We can also use it to distinguish between restrictive diseases: DLCO is reduced in interstitial lung disease and normal in neuromuscular diseases like MS.

What exactly is DLCO measuring? In short, it measures the extent to which the alveolar-capillary barrier permits diffusion. According to Fick’s law, the volume of gas that diffuses through a tissue barrier, in ml/min, is given by:

V = ΔP * A * D / T

= (Partial pressure difference across the membrane) * (Surface area) * (Diffusivity constant) / (Thickness of the barrier).

So, now we see rationale for why DLCO can distinguish the diseases mentioned above.

  • In emphysema, alveolar walls, including the capillary beds within, are destroyed, creating more V per unit Q. This, in effect, reduces the surface area that alveoli can use to participate in gas exchange. So, with less surface area, we have less diffusion. In asthma, on the other hand, the integrity of the alveolar walls is just fine.
  • In interstitial lung disease, including pulmonary fibrosis, we have increased thickness of the alveolar barrier. So, once again, rate of diffusion falls. However, in MS, once again, the alveolar-capillary interface works just fine.

This is all fine. I started to get confused, though, when I learned that:

  • In anemia, DLCO is lower
  • In polycythemia, DLCO is higher.

What exactly is going on here? Where in Fick’s law would red cell count affect DLCO?

Here’s an illustration of what I think is going on.


Let’s consider polycythemia. Red blood cells are in excess. Thus we have a greater amount of hemoglobin. This increases the amount of reactants in reaction 2. And, with a greater amount of reactants, by the law of mass action, we have greater conversion to product. So, in polycythemia, we see faster conversion of CO away to carboxyhemoglobin.

Thus, CO disappears faster, and so we have lower serum levels of CO. This, then, corresponds to a reduction in products in reaction 1. And, importantly, only reaction 1 contributes to the diffusion gradient. Recall that the diffusion gradient is given by the difference in partial pressures across the barrier. And, only dissolved gases contribute to partial pressure. Protein-bound gases do not contribute to partial pressure. So, since we have a reduction in products of reaction 1, and since only reaction 1 contributes to the diffusion of gas, we see why the rate of reaction 1 would increase in polycythemia. Again, it comes down to the law of mass action: decreased products causes an increased reaction rate. So, we have an increased rate of diffusion in polycythemia.

How does this relate to Fick’s law? Well, what happens in polycythemia is simply that the pressure gradient is increased. When we think of increased pressure gradient, we often think of increased alveolar pressure, but decreased capillary pressure has the same effect. Polycythemia and anemia thus affect DLCO by altering the size of the partial pressure gradient.

Why is carbon monoxide used in the first place?

Firstly, it’s nontoxic in amounts as small as those which are used for the DLCO test. Also: very little CO exists in the blood as is, so we don’t have an appreciable level of pre-existing dissolved CO that could unpredictably alter our diffusion gradient.

And the most interesting reason: CO has an extremely high affinity for hemoglobin. This ensures that reaction 2 is carried out to full completion, and therefore that dissolved CO remains low. This, like we’ve just discussed, ensures a robust concentration gradient for CO. And this gradient, then, ensures that diffusion of CO continues unabated, regardless of how much has already bound CO. So, concentration of CO rises linearly with diffusion time. This makes the DLCO a good measurement tool, because the degree of impairment of diffusion correlates linearly with a decrease in DLCO.


First Aid 2014 p. 599

Oxygen, on the other hand, increases rapidly at first, and then levels off. This is because reaction 2 proceeds slower than it does with CO, so the products of reaction 1 build up with time, and so rate of diffusion slows.

Though CO increases linearly, it still increases slow enough such that equilibrium will not be reached by the time pulmonary blood reaches the end of the capillary bed. Recall that Fick’s law also incorporates the diffusivity constant; CO simply doesn’t diffuse that well. (CO2, in contrast, diffuses much faster.) But the fact that CO diffuses slowly is good for us. We wouldn’t want CO reaching equilibrium too quickly. If it did, CO would equilibrate easily, even in cases with poor diffusion. Thus the test would be useless.

Diffusion and perfusion limitation

Since PaCO does not reach equilibrium with PACO by the end of the pulmonary capillary, CO is said to be diffusion limited. CO2 and N2O, on the other hand, are perfusion limited.

These terms might also produce some confusion. Keep in mind that, if the experimenters were to raise PACO, more diffusion would occur. PACO certainly still matters. So, diffusion rate in “diffusion-limited” gases also depends on perfusion. The best terminology would be to say that CO is diffusion-and-perfusion-limited, while CO2 and N2O are perfusion-limited only. All that matters in terms of this distinction is whether or not the gas equilibrates by the end of the capillary bed.

One last point: in healthy individuals, oxygen is perfusion-limited, but in lung disease, including emphysema, fibrosis, and low V/Q, oxygen may become diffusion-limited. This is simply another way of saying that PaO2 does not equilibrate with PAO2 by the end of the pulmonary capillary bed. And, this should be obvious to us. We know that emphysema patients, for example, can experience lower-than-normal O2 saturation on room air, and might require supplemental oxygen. Well, obviously, their PAO2 is just fine. However, their PaO2 is low and they have an appreciable A-a gradient. The only way this would be possible would be that capillary oxygen isn’t fully equilibrating. In essence, A-a gradient is synonymous with diffusion-limitation.

Please provide thoughts, comments, and corrections in the comments! Hope this helps!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s