Interpretation of Pulmonary Function Tests A Practical Guide, 3. ed

4. Diffusing Capacity of the Lungs

An important step in the transfer of oxygen from ambient air to the arterial blood is the process of diffusion, that is, the transfer of oxygen from the alveolar gas to the hemoglobin within the red cell. The pertinent anatomy is shown in Figure 4-1A. The path taken by oxygen molecules is shown in Figure 4-1B. They must traverse the alveolar wall, capillary wall, plasma, and red cell membrane and then combine with hemoglobin.

The diffusing capacity of the lungs (Dl) estimates the transfer of oxygen from alveolar gas to red cell. The amount of oxygen transferred is largely determined by three factors. One factor is the area (A) of the alveolar-capillary membrane, which consists of the alveolar and capillary walls. The greater the area, the greater the rate of transfer and the higher the Dl. Area is influenced by the number of blood-containing capillaries in the alveolar wall. The second factor is the thickness (T) of the membrane. The thicker the membrane, the lower the Dl. The third factor is the driving pressure, that is, the difference in oxygen tension between the alveolar gas and the venous blood (∆Po2). Alveolar oxygen tension is higher than that in the deoxygenated venous blood of the pulmonary artery. The greater this difference (∆Po2), the more oxygen transferred. These relations can be expressed as


The diffusing capacity of oxygen (Dlo2) can be measured directly, but this is technically extremely difficult. Measuring the diffusing capacity of carbon monoxide (Dlco) is much easier and provides a valid reflection of the diffusion of oxygen. In essence, the difference between alveolar and venous carbon monoxide tension (APco) is substituted for oxygen in Equation 1.

FIG. 4-1. Alveolar-capillary membrane through which oxygen must diffuse to enter the blood. In B, alveolar wall is represented by the black rectangle. Hgb, hemoglobin.

Several techniques for estimating Dlco have been described. The most widely used is the single-breath (SB) method (SBDlco). The subject exhales to residual volume and then inhales a gas mixture containing a very low concentration of carbon monoxide plus an inert gas, usually helium. After a maximal inhalation to total lung capacity (TLC), the patient holds his or her breath for 10 seconds and then exhales completely. A sample of exhaled alveolar gas is collected and analyzed. By measuring the concentration of the exhaled carbon monoxide and helium, the value of the Dlco can be computed. The helium is used to calculate TLC, and the exhaled carbon monoxide is used to calculate the amount of carbon monoxide transferred to the blood.

The technical details of measurement of SBDlco are complex. To improve accuracy and reproducibility of testing among laboratories, the American Thoracic Society has established standards for performance of the test [1,2].


An average normal value is 20 to 30 mL/min per mm Hg; that is, 20 to 30 mL carbon monoxide is transferred per minute per mm Hg difference in the driving pressure of carbon monoxide. The normal values depend on age (decrease with aging), sex (slightly lower in females), and size (taller people have larger lungs and therefore a higher Dlco). The inclusion of helium provides an estimate of total alveolar volume (Va). Dividing Dlco by Va tends to normalize for difference in size and therefore the Dlco/Va ratio (or Krogh constant) tends to be the same in various-sized normal subjects. To an extent in a given subject, the Dlco is also directly related to the volume of the inhaled breath. The smaller the volume, the lower the Dlco. The Dlco/Va in this situation would change little, however, if at all. This fact is useful, especially when repeated tests are obtained overtime. The volume inhaled might vary from year to year, but the Dlco/Va tends to correct for this. Most patients can hold their breath for 10 seconds, but some subjects with very small vital capacities cannot inhale a sufficient quantity of the gas mixture to give a valid test.

PEARL: In the healthy subject, the Va is essentially the same as the TLC and can be used as an estimate of TLC. The Va is also a good estimate of TLC in most restrictive conditions. With obstructive disease, Va underestimates TLC, just as the nitrogen washout and inert gas dilution techniques do (see section 3C, page 31). However, TLC obtained with plethysmography minus the Va provides an estimate of the severity of nonuniform gas distribution, that is, the volume of poorly ventilated lung (see Chapter 3, first Pearl, page 40).


Usually, an increased Dlco is not a concern. However, there are some interesting causes of an increased Dlco, as follows:

1. Supine position: Rarely is the Dlco measured while the subject is supine, but this position produces a higher value because of increased perfusion and blood volume of the upper lobes.

2. Exercise: It is difficult to hold one's breath for 10 seconds during exercise. When this is done just after exercise,

however, Dlco is increased because of increased pulmonary blood volumes.

3. Asthma: Some patients with asthma, especially when symptom-free, have an increased Dlco, possibly because of more uniform distribution of pulmonary blood flow.

4. Obesity: The Dlco can be increased in obese persons, especially those who are massively obese. This increase is thought to be due to an increased pulmonary blood volume.

5. Polycythemia: This is an increase in capillary red cell mass. This essentially amounts to an increase in area (A) in Equation 1.

6. Intra-alveolar hemorrhage: In conditions such as Goodpasture's syndrome, the hemoglobin in the alveoli combines with carbon monoxide to produce an artificially high uptake of carbon monoxide, which causes an increase in the calculated Dlco. Indeed, sequential Dlco measurements have been used to follow increases or decreases in intra-alveolar hemorrhage.

7. Left-to-right intracardiac shunts: These lead to an increased pulmonary capillary volume.


Any process that decreases the surface area available for diffusion or thickens the alveolar-capillary membrane will decrease the Dlco (Fig. 4-1B and Eq. 1). On the basis of these considerations, conditions that reduce the diffusing capacity can be determined (Table 4-1). The major ones are listed here.

Conditions That Decrease Surface Area

1. Emphysema: Although lung volume is increased, alveolar walls and capillaries are destroyed, and thus, the total surface area is reduced. Reduction of the Dlco in a patient with significant airway obstruction strongly suggests underlying emphysema.

2. Lung resection: If only a small portion of lung is resected (such as a lobe in an otherwise healthy patient), capillary recruitment from the remaining normal lung can result in an unchanged Dlco. If sufficient capillary surface area is lost, as with a pneumonectomy, the Dlco is reduced.

TABLE 4-1. Causes of a decreased diffusing capacity

Decreased area for diffusion Emphysema Lung/lobe resection Bronchial obstruction, as by tumor Multiple pulmonary emboli Anemia

Increased thickness of alveolar-capillary membrane Idiopathic pulmonary fibrosis

Congestive heart failure


Sarcoidosis, involving parenchyma Collagen vascular disease—scleroderma, systemic lupus erythematosus

Drug-induced alveolitis or fibrosis—bleomycin, nitrofurantoin, amiodarone, methotrexate

Hypersensitivity pneumonitis, including farmer's lung Histiocytosis X (eosinophilic granuloma)

Alveolar proteinosis Miscellanous

High carbon monoxide back pressure from smoking Pregnancy

Ventilation-perfusion mismatch

3. Bronchial obstruction: A tumor obstructing a bronchus obviously reduces area and lung volume. Again, the Dlco/Va might be normal.

4. Multiple pulmonary emboli: By blocking perfusion to alveolar capillaries, emboli effectively reduce area. Also, primary pulmonary hypertension causes a reduction in capillary area.

5. Anemia: By reducing pulmonary capillary hemoglobin, anemia also effectively reduces area, as does any condition that lowers capillary blood volume. The usual correction for men with anemia is this equation:

Dlco (cor) = Dlco (unc) x [10.22 + Hb]/[1.7 x Hb] (Eq. 2) [1]

where cor is corrected, unc is uncorrected, and Hb is hemoglobin. For women, the factor in the first set of brackets is 9.38 instead of 10.22.

Conditions That Effectively Increase Wall Thickness

As is discussed in Chapter 6, however, much of the reduction in Dlco in these fibrotic conditions is thought to be due to mismatching of ventilation and perfusion.

1. Idiopathic pulmonary fibrosis: This is also called cryptogenic fibrosing alveolitis. It thickens the alveolar-capillary membrane and also decreases lung volume.

2. Congestive heart failure: In this disorder, transudation of fluid into the interstitial space (tissue edema) or into the alveoli lengthens the pathway for diffusion.

3. Asbestosis: This is pulmonary fibrosis caused by exposure to asbestos.

4. Sarcoidosis: The lesions thicken the alveolar walls.

5. Collagen vascular disease: Conditions such as scleroderma and systemic lupus erythematosus probably alter or obliterate capillary walls, a situation that effectively increases the barrier to diffusion. This may be the first pulmonary function test result to become abnormal in these conditions.

6. Drug-induced alveolitis or fibrosis: Bleomycin, nitrofurantoin, amiodarone, and methotrexate can be involved in these conditions.

7. Hypersensitivity pneumonitis: This condition includes farmer's lung.

8. Histiocytosis X: This condition is eosinophilic granuloma of the lung or Langerhans' cell histiocytosis.

9. Alveolar proteinosis: Alveoli are filled with a phospholipid- rich material.

Miscellaneous Causes

1. The high carbon monoxide tension in the blood of a heavy smoker can decrease the APco or driving pressure. This lowers the Dlco (Eq. 1).

2. Pregnancy usually reduces the Dlco by approximately 15%, but the mechanism is not fully understood.

3. With an isolated, unexplained reduction in Dlco (with normal results on spirometry and normal lung volumes), pulmonary vascular disease, such as primary pulmonary hypertension, recurrent pulmonary emboli, or obliterative vasculopathy, should be considered. A recent study in which computed tomography was used to evaluate an isolated reduction in Dlco found mild emphysema or mild fibrosis to explain most of the cases [3].

FIG. 4-2. Case of severe restrictive disease. Total lung capacity (TLC) is markedly reduced, the ratio of forced expiratory volume in 1 second to forced vital capacity (FEV1/FVC) is normal, the carbon monoxide diffusing capacity of the lung (Dlco) is reduced, and the oxygen saturation (SaO2) is decreased with exercise. The maximal voluntary ventilation (MVV) is not as severely reduced as the FEV1; thus, the calculation of FEV1 x 40 does not work in this situation. The steep slope of the flow-volume (FV) curve and the reduced Dlco suggest a pulmonary parenchymal cause of the severe restriction. The diagnosis in this case was idiopathic pulmonary fibrosis. Numbers in parentheses are percent of predicted, except for slope FV curve, in which numbers are a range.

Figures 4-2 through 4-4 present cases in which knowledge of the Dlco is very useful.


The test for Dlco is very sensitive. We have found transient decreases of 3 to 5 mL/min per mm Hg with mild respiratory infections in healthy subjects. It is a useful test for following the course of patients with idiopathic pulmonary fibrosis or sarcoidosis and for monitoring the toxicity of chemotherapy or for evaluating therapeutic interventions. The test has also been used to follow the extent of intra-alveolar hemorrhage in conditions such as Goodpasture's syndrome.

FIG. 4-3. As with Figure 4-2, this pattern is consistent with a restrictive process (reduced TLC, FVC, and FEV1, and a normal FEV1/FVC ratio). However, it differs from the case in Figure 4-2 in that the Dlco is normal, as is the slope of the FV curve. The MVV is also low. Further testing revealed a severe reduction in respiratory muscle strength (see Chapter 9), consistent with the diagnosis of amyotrophic lateral sclerosis. (AbbrEV1ations and numbers in parentheses are defined in the legend to Fig. 4-2.)

One might expect changes in the resting Dlco to be closely correlated with the arterial oxygen tension (Pao2). However, this is not always so. For example, with lung resection Dlco is reduced, but the Pao2 is generally normal. However, a low resting Dlco often correlates with a decrease in Pao2 during exercise.

FIG. 4-4. In this case there is no apparent ventilatory limitation, the area under the FV curve being normal. All test values are normal except for the striking reduction in the Dlco and the desaturation with exercise. Primary pulmonary hypertension was diagnosed. (AbbrEV1ations and numbers in parentheses are defined in the legend to Fig. 4-2.)

PEARL: A Dlco of less than 50% of predicted suggests a pulmonary vascular or parenchymal disorder. In a patient with a normal chest radiograph and no EV1dence of airway obstruction, this should lead to further investigation, such as high-resolution computed tomography, to look for interstitial changes, or a pulmonary vascular study, such as echocardiography, to measure pulmonary artery pressure.


1. American Thoracic Society. Single-breath carbon monoxide diffusing capacity (transfer factor): recommendations for a standard technique—1995 update. Am J Respir Crit Care Med 152:2185-2198,1995.

2. Crapo RO, Jensen RL, Wanger JS. Single-breath carbon monoxide diffusing capacity. Clin Chest Med 22:637-649, 2001.

3. Aduen JF, Zisman DA, Mobin SI, Venegas C, Alvarez F, Biewend M, et al. Retrospective study of pulmonary function tests in patients presenting with isolated reduction in single-breath diffusion capacity: implications for the diagnosis of combined obstructive and restrictive lung disease. Mayo Clin Proc 82:48-54, 2007.

If you find an error or have any questions, please email us at Thank you!