This chapter is most relevant to Section F7(iv) from the 2017 CICM Primary Syllabus, which expects the exam candidates to be able to "define diffusing capacity and its measurement".  This has come up at least once in the CICM past papers, as an unexpected element of the answer to Question 20 from the first paper of 2012, which presented itself as  "list the physiological factors affecting the diffusion of oxygen across the alveolar membrane". An earnest (albeit longwinded) attempt to explore those factors and fail that question by interpreting it literally can be seen in the chapter on the diffusion of gases through the alveolar membrane. Here, instead, will focus specifically on diffusing capacity, and how it can be determined.

In summary:

Diffusing capacity = Net rate of gas transfer / Partial pressure gradient

Factors affecting diffusing capacity include:

  • Factors which influence gas properties
    • The density of the gas
    • Size of the molecules
    • The temperature of the medium
  • Factors which influence the gas exchange surface area
    • Age (with increasing age, total available surface area decreases, irrespective of the other factors)
    • Body size: height influences the size of the lungs
    • Lung volume
    • Shunt, dead space and V/Q inequality
  • Factors which influence the membrane characteristics
    • Disease states which increase the thickness of the blood-gas barrier, which include:
      • Pulmonary oedema
      • Interstitial lung disease, eg. pulmonary fibrosis
  • Factors which influence uptake by erythrocytes
    • The affinity of haemoglobin for oxygen
    • Haemoglobin concentration
    • Cardiac output (insofar as it affects capillary transit time)
  • Sources of error in the course of measurement, due to alveolar haemorrhage, carbon monoxide poisoning, anaemia, etc

With exercise, both major elements affecting diffusing capacity are altered:

  • Oxygen uptake in the pulmonary capillaries increases because:
    • Surface area increases (larger tidal volumes)
    • Pulmonary blood flow increases (increased cardiac output)
    • V/Q matching improves
  • Partial pressure gradient in the pulmonary capillaries increases because:
    • Oxygen extraction ratio increases, decreasing the PO2 of mixed venous blood
    • Increased minute ventilation decreases the alveolar PCO2 
    • Increased delivery of haemoglobin to the absorptive surface

It is difficult to recommend any single article regarding this, as most of them focus on one specific aspect and none seem to offer a brief broad overview of the sort that a time-poor exam candidate may be wishing for. A reasonable source is Hsia (2001). which leans somewhat to the side of exercise-induced change in DLCO, and Ayers et al (1975), which is a solid (though dated) discussion of the different ways the DLCO may be pathologically decreased. As with all things related to pulmonary function testing, the excellent PFTBlog is an excellent resource, particularly with regards to the methods of testing DLCO

Definition of diffusing capacity

Though it was not explicitly asked for in the question, an unwritten expectation of the examiners in Question 20 from the first paper of 2012 was that the trainees would define diffusing capacity in the course of listing factors which affect the diffusion of respiratory gases. From the college examiners' own comment, the definition for this concept is:

"The diffusing capacity is defined as the  volume of gas that will diffuse through the membrane each minute for a partial pressure difference of 1mmHg."

Nunn's defines it a little differently:

"[Diffusing capacity is] the propensity of a gas to diffuse as a result of a given pressure gradient"

Diffusing capacity = Net rate of gas transfer / Partial pressure gradient

This property is usually referred to as DL or DL, and it is typically measured in gas volume, per unit pressure, per unit time; for example the SI units are mmol/min/kPa, and the traditional units are ml/min/mmHg. In essence, this parameter describes the ease with which gases can find themselves transported into alveolar capillary blood, and is therefore a handy distillation of all the factors which influence the diffusion of respiratory gases into one numerical representation. 

For oxygen, the equation is:

DLO2 = oxygen uptake / PO2 gradient

The oxygen uptake is sort-of measurable, as it is the difference between mixed venous and arterial oxygen content. However, the PO2 gradient here is the difference between alveolar PO2 and the pulmonary capillary PO2,  of which the latter is basically impossible to ever measure directly. With various assumptions, one can guess as to what the capillary PO2 should be, and perform some back-of-the-envelope calculations.  For some weird reason, the only source which actually lists this value seems to be Gehr et al (1981), which is a book chapter on the comparative respiratory physiology of mammals. There, along with Thompson's gazelle and the dwarf mongoose, one can find the value for humans, which is reported as 2.47 ml/mbar/sec. A more authoritative (but still unreferenced) value reported by conventional notation can be found in ER Weibel's textbook from 1984, who gives 20-30ml/min/mmHg.  

The diffusing capacity of carbon dioxide is even more difficult to track down. Nunn's does not give references or even exact measurements, but rather delivers a line about how it is 20.5 times greater than the diffusing capacity of oxygen. Rabbit data from Heller et al (1998) reports a DLCO2 of 14.0 ml/mmHg/min. 

As will be demonstrated below these values - measured in resting individuals - do not represent the true maximum of the diffusional capacity of the lung.  This can only be revealed by strenuous exercise, where the delivery of blood to the capillary increases significantly.

Measurement of diffusing capacity

Obviously, the gas you're most interested in oxygen, and so it would make some sort of logical sense to measure this gas directly, but in reality, there are several practical barriers. Or at least, there were barriers when the issue of measuring diffusing capacity had first come up. Basically, in order to measure the DLO2, you would need to be able to accurately measure both the oxygen uptake and the partial pressure gradient.  For the gradient, you would need to calculate the alveolar oxygen (easily done), and then to measure the arterial oxygen (as a surrogate of pulmonary end-capillary oxygen). Then, "the tension of O2 and CO2 in the arterial blood have to be measured by the microtonometer technique developed by Riley-a technique which requires considerable practice and dexterity", wrote Dacie in 1957 to whom oxygen-sensing Clark electrodes were not available.  This does not seem like much of an obstacle to the modern-day intensivist who at any given moment has litres of their patient's blood (both venous and arterial) ready for sampling, as well as accurate instruments to measure the gas content thereof. However, historically it was a major problem, and there remains a bit of a reluctance to sample the arterial of the ambulatory outpatient group. One's referrals would quickly dry up after people realise what you're planning to do to them. 

Thus, the use of carbon monoxide has been historically much more popular. Marie and August Krogh had first come up with this in 1915: 

"It is assumed further that when a small proportion of CO is allowed to pass into the blood, the gas will combine practically instantaneously with the haemoglobin and the CO pressure in the blood can be taken as 0. When therefore a mixture of CO with air is enclosed in the lungs during a certain time and the drop in CO percentage is determined the diffusion through the alveolar wall can be calculated."

In short, one gives a patient some non-lethal and known dose of carbon monoxide to inhale. The patient holds that breath for ten seconds, and then exhales it. As carbon monoxide has nowhere else to go but into erythrocytes, any difference between the inhaled and exhaled amounts of CO must have diffused across the blood-gas barrier and become bound to haemoglobin. Thus, in the equation:

DLCO = Carbon monxide uptake / Carbon monoxide gradient 

the carbon monoxide uptake is the "missing" difference of inhaled and exhaled CO, and the gradient is assumed to be between the alveolar partial pressure of CO (which is known, because you gave it) and the arterial partial pressure of CO (which is 0 mmHg, because we know that all of it ends up being bound to haemoglobin). Thus, the measurement of DLCO can be carried out non-invasively.

There are three main methods for the measurement of DLCO: the single breath method, the steady state method and the rebreathing method. The single breath method is described in glorious detail by the excellent ERS/ATS standards statement (Cotes et al, 1993), from which the author has liberally "borrowed" some explanatory images. The rebreathing technique is explored in detail here and the steady-state method here.  An indepth knowledge of this subject is not (cannot possibly be) expected of the CICM exam candidates, and so here, it will suffice to summarise it as follows:

Single breath method of measuring DLCO

single breath DLCO measurement ytechnique from the ERS-ATS statement

  • A period of time of breathing room air should ideally precede any measurement
  • First, the patient exhales maximally (down to RV)
  • The patient then inhales a gas mixture of  0.3% carbon monoxide and 10% helium
    (the helium is for the measurement of alveolar volume)
  • This is a vital capacity breath (i.e. up to TLC), and its volume is measured
  • The patient holds this breath for ten seconds
    • This breath hold is meant to ensure the equal distribution of carbon monoxide to all the lung units,  irrespective of their time constant 
    • It is important to avoid Valsalva-ing at this point, as it can affect the intrathoracic blood volume and falsely decrease the DLCO.
  • The patient then exhales.
    • The first 0.75 litres is completely ignored, as this is considered to be dead space gas, not representative of the rest.
  • A gas sample is then taken
    • Total alveolar volume can be measured from the expiratory helium concentration (this is a classical application of the tracer gas dilution measurement technique for measuring lung volumes) 
    • Carbon monoxide uptake can be determined from the difference between the inhaled and exhaled partial pressure measurements
    • The partial pressure gradient for carbon monoxide can be determined from the exhaled partial pressure measurement

Rebreathing method DLCO

  • This is virtually the same as the single breath method, except there is no breath holding.
  • The patient is made to breathe rapidly (the recommended respiratory rate is 30) while breathing from a reservoir with a known quantity and volume of gas, containing  0.3% carbon monoxide and 10% helium
  • The quantity of gas in the bag is usually adjusted so that it is roughly the same as the tidal volume of the subject, i.e. it empties completely during inspiration
  • After a period of such rapid breathing, the gas is sampled
  • Calculation of alveolar volume and carbon monoxide uptake can then be performed in exactly the same way as for the single breath 
  • For some reason, this technique is virtually unknown in clinical practice, and appears to be mainly used in scenarios where one needs to measure the DLCo without significantly interrupting the breathing pattern of the subject, eg. when they are pedalling madly on an exercise cycle. 

Steady state method of measuring DLCO

  • The subject is made to breathe a controlled gas mixture which contains 0.3% carbon monoxide.
  • Their exhaled gas is collected in a bag
  • After a period of breathing (long enough for a steady state to be established) the exhaled gas is analysed
  • Carbon monoxide delivery and exhaled gas volume is known, and so it is easy to calculate the carbon monoxide uptake.
  • The alveolar concentration of carbon monoxide can be calculated from a modified form of the alveolar gas equation
  • Again, this technique appears to be virtually unknown in routine clinical practice; its major advantage is the complete lack of reliance on any level of patient participation, making it suitable for use in uncooperative or sedated subjects

Factors which influence diffusing capacity

The equation which describes this parameter is quite simple, and the factors which affect it can be divided into gas properties and respiratory system properties. A gas with a higher diffusive capacity will be able to negotiate the blood-gas barrier more easily than a gas with a lower diffusive capacity, at any given pressure gradient. Similarly, the properties of the respiratory system may change in a way which may increase or decrease the diffusive capacity, for the same gas and at the same partial pressure gradient. Of the properties of the respiratory system, three main factors may change: either the surface area changes, or the membrane thickness changes, or the uptake of the gas by red blood cells is somehow altered. One can generate a memorable point-form list to describe these factors, for the purposes of exam preparation. Thus:

  • Factors which influence gas properties
    • All the factors which influence the diffusion coefficient of the gas will play a role in this, including:
      • The density of the gas
      • Size of the molecules
      • The temperature of the medium
  • Factors which influence the gas exchange surface area
    • Age (with increasing age, total available surface area decreases, irrespective of the other factors)
    • Body size: height influences the size of the lungs
    • Lung volume
      • The greater the lung volume, the greater the diffusing capacity, i.e. if one is comparing between individuals, one should use  a metric which is indexed to alveolar volume (eg. diffusing capacity per litre of alveolar volume)
      • Everything which affects lung volume is therefore a potential source of error, eg. lung disease, posture, obesity, pregnancy, etc.
    • Factors which change the ventilation-perfusion characteristics:
      • Shunt: no diffusion takes place
      • Dead space: no diffusion takes place
      • V/Q scatter: inefficient incomplette diffusion takes place
  • Factors which influence the membrane characteristics
    • This is basically the disease states which increase the thickness of the blood-gas barrier, which include:
      • Pulmonary oedema
      • Interstitial lung disease, eg. pulmonary fibrosis
    • Strictly speaking, one should include the viscosity of the medium here (i.e of the cytosol, basement membrane and capillary plasma). However, practically these are stable elements which can be ignored.
  • Factors which influence uptake by erythrocytes
    • The affinity of haemoglobin for oxygen
    • Haemoglobin concentration
    • Cardiac output (insofar as it affects capillary transit time
  • Sources of error
    • Loss of carbon monoxide to extravascular alveolar haemoglobin, eg. in the context of alveolar haemorrhage due to Goodpasture syndrome
    • Presence of "homegrown" carbon monoxide, due to smoking or extensive haemoglobin breakdown (eg. intravascular haemolysis) which could limit CO uptake
    • Competition between CO and oxygen (if the patient had been previously breathing 100% FiO2, for example)
    • Haemoglobin concentration, when low, can falsely decrease the DLCO measurement even though the performance of the alveolar/capillary complex remains completely healthy

If one looks closely enough, one might detect that this list is virtually identical to the list of factors which influence the diffusion of gases across the alveolar membrane, with the notable exception fo the partial pressure gradient (which is incorporated into the definition of diffusive capacity) and the various factors related to measurement error. 

Change of diffusing capacity with exercise

Some might say that the discussion of any resting diffusing capacity is a misnomer because it refers to an unstressed system, which in fact has a much higher capacity for diffusion. Indeed, with vigorous exercise, DLO2 increases from 20-30 ml/min/mmHg to something close to 100-120 ml/min/mmHg, which is the "real" capacity for diffusion. This increase is because the oxygen uptake rate in the equation (DLO2 = oxygen uptake / PO2 gradient) increases significantly. It does not take any great stretch of the imagination to explain why this might be. Consider: the minute volume increases, not only because of the increased respiratory rate but also because of the increase in the tidal volume. With increased lung volume, the total alveolar gas exchange area is increased. Moreover, the cardiac output is increased. With that, the delivery of blood to the pulmonary capillaries is increased. This changes the V/Q distribution, as more capillaries are recruited in lung regions which were previously either "true" dead space or had a V/Q  much greater than 1.0. To summarise this in a palatable form:

With exercise, both major elements affecting diffusing capacity are altered:

  • Oxygen uptake in the pulmonary capillaries increases because:
    • Surface area increases (larger tidal volumes)
    • Pulmonary blood flow increases (increased cardiac output)
    • V/Q matching improves (areas of high ventilation receive greater blood flow, and dormant capillary beds are recruited)
  • Partial pressure gradient in the pulmonary capillaries increases because:
    • Oxygen extraction ratio increases, decreasing the PO2 of mixed venous
    • Increased minute ventilation decreases the alveolar PCO2 (thus increasing the alveolar PO2, all other things remaining equal)
    • Increased delivery of haemoglobin to the absorptive surface acts as an oxygen sink and maintains a low capillary partial pressure

How much of an increase in DLO2 should we expect? The college answer to Question 20 from the first paper of 2012 addresses this in a cryptic remark, "...alveolar ventilation increases and there is better matching of ventilation and perfusion increases from 21ml/min/mmHg up to  65ml/min/mmHg". 
Presumably, the values quoted in the latter half of this obiter dictum are referring to changes in DLCO, and are derived from somewhere reputable, but who knows where that is. One generally expects they come from a textbook, and textbook values usually come from studies performed in the 1960s.  Without knowing specifically which medieval source the examiners had in mind, the search for substantive peer-reviewed references is essentially the same as throwing darts at the literature. For example, a short search turns up a study by Turino et al (1963), whose healthy volunteers got resting DLCO values ranging from 18 to 22, and exercise values ranging from 55 to 64  ml/min/mmHg. This seems approximately correct, and in any case one cannot conceive of a universe where having or not having the exact numbers here would be the deciding factor in one's exam performance. 

References

Ogilvie, C. M., et al. "A standardized breath holding technique for the clinical measurement of the diffusing capacity of the lung for carbon monoxide." The Journal of clinical investigation 36.1 (1957): 1-17.

Cotes, J. E., et al. "Standardization of the measurement of transfer factor (diffusing capacity)." (1993) European Respiratory Journal 6: 41-52

Weibel, Ewald R., et al. "Morphometric model for pulmonary diffusing capacity I. Membrane diffusing capacity." Respiration physiology 93.2 (1993): 125-149.

Krogh, Marie. "The diffusion of gases through the lungs of man." The Journal of physiology 49.4 (1915): 271-300.

Turino, G. M., et al. "Effect of exercise on pulmonary diffusing capacity." Journal of Applied Physiology 18.3 (1963): 447-456.

Hsia, Connie CW. "Recruitment of lung diffusing capacity: update of concept and application." Chest 122.5 (2002): 1774-1783.

Ayers, Larry N., et al. "Diffusing capacity, specific diffusing capacity and interpretation of diffusion defects." Western Journal of Medicine 123.4 (1975): 255.