American Journal of Respiratory and Critical Care Medicine

The single-breath carbon monoxide diffusing capacity (DlCO) is the product of two measurements during breath holding at full inflation: (1) the rate constant for carbon monoxide uptake from alveolar gas (kco [minute−1]) and (2) the “accessible” alveolar volume (Va). kco expressed per mm Hg alveolar dry gas pressure (Pb*) as kco/Pb*, and then multiplied by Va, equals DlCO; thus, DlCO divided by Va (DlCO/Va, also called Kco) is only kco/Pb* in different units, remaining, essentially, a rate constant. The notion that DlCO/Va “corrects” DlCO for reduced Va is physiologically incorrect, because DlCO/Va is not constant as Va changes; thus, the term Kco reflects the physiology more appropriately. Crucially, the same DlCO may occur with various combinations of Kco and Va, each suggesting different pathologies. Decreased Kco occurs in alveolar–capillary damage, microvascular pathology, or anemia. Increased Kco occurs with (1) failure to expand normal lungs to predicted full inflation (extrapulmonary restriction); or (2) increased capillary volume and flow, either globally (left-to-right intracardiac shunting) or from flow and volume diversion from lost or damaged units to surviving normal units (e.g., pneumonectomy). Decreased Va occurs in (1) reduced alveolar expansion, (2) alveolar damage or loss, or (3) maldistribution of inspired gases with airflow obstruction. Kco will be greater than 120% predicted in case 1, 100–120% in case 2, and 40–120% in case 3, depending on pathology. Kco and Va values should be available to clinicians, as fundamental to understanding the clinical implications of DlCO. The diffusing capacity for nitric oxide (DlNO), and the DlNO/DlCO ratio, provide additional insights.

The single-breath diffusing capacity for carbon monoxide (DlCO) (known in Europe as the transfer factor, TlCO) is, after spirometry and lung volumes, the most clinically useful routine pulmonary function test. The DlCO, as pointed out by its originator, Marie Krogh (1), is the product of two separate but simultaneous measurements (Figure 1): the rate constant kco (the rate of uptake of CO from alveolar gas), and the alveolar volume (Va). The important point is that Kco (kco reexpressed per mm Hg alveolar Pco) is linearly related to the alveolar uptake efficiency for carbon monoxide (2, 3). Because of the special properties of carbon monoxide, Kco directly reflects the quality of alveolar–capillary gas uptake. Many articles and pulmonary function testing (PFT) laboratories do not quote Va and Kco from which the DlCO is derived; this may result in significant loss of clinical information.

Rate of Uptake of Alveolar Carbon Monoxide (kco)

During breath holding in the single-breath DlCO, CO is removed from alveolar gas at an exponential rate [loge(CO0/COt)/BHT], where CO0 and COt are the alveolar concentrations at the start and finish of the breath-holding time (BHT). This expression is a rate constant with units of minute−1 or second−1; in Figure 1 it is represented by the slope, kco.

Alveolar Volume (Va)

The DlCO is measured during breath holding at full inflation; in absolute terms, this represents total lung capacity (TLC). The lung volume during breath holding is measured simultaneously by dilution of any nonabsorbable gas, most commonly helium (He) (Figure 1), at the same time as the kco is measured (4). The alveolar volume (Va) is an “accessible” volume, that is, that seen by the gas-exchanging surface, derived from the single-breath helium dilution volume after subtracting an “estimated” anatomic dead space (Vdanat) from the inspired volume (Vi) (Figure 1). The Vi starts from residual volume and finishes at maximal inflation (∼TLC); the inspiration should be made as rapidly as possible. In normal subjects, Va is within 10% of TLC, with a mean Va/TLC ratio (combining men and women) of 93.5% ± 6.6 (1 SD) (5); the Va/TLC ratio has no significant dependence on age, sex, height, or weight (5), but decreases substantially when there is intrapulmonary airflow obstruction and maldistribution of ventilation. Vdanat represents 2–3% of the TLC in normal subjects, the remaining 4% of the Va/TLC difference occurring because gas mixing in the 10-second breath hold is incomplete. In disease, the difference between the single-breath Va and the multibreath or plethysmographic TLC, and the Va/TLC ratio, deserves more study (5) as an index of gas mixing efficiency.

Combining Va and kco

Equation 1 is the first step in the calculation of the DlCO:

where kco is the fractional change in CO concentration, expressed in minute−1, and ¢CO is the uptake of CO from alveolar gas during breath holding at TLC. Equation 1 gives a large value for ¢CO because, as pointed out by Marie Krogh (1), the calculation implies that all alveolar gas is pure CO. For the second step, to obtain DlCO, both sides of the equation are divided by Pb*, where Pb* is barometric pressure, usually approximately 760 mm Hg, minus water vapor pressure at 37°C in alveolar gas (Pb – Ph2o), Thus:
where ¢CO/Pb* is the alveolar CO uptake per minute per mm Hg Pco, which, as a conductance, defines the DlCO. In pulmonary function reports DlCO is divided by Va with ml STPD exchanged for L BTPS:
where DlCO/Va and Kco have units of ml minute−1 mm Hg−1 L−1, 1,000 converts milliliters to liters, and 1.2 is the STPD-to-BTPS factor. These units, as reported in pulmonary function laboratories, give the impression of a volume “adjustment,” leading to much confusion, whereas it is obvious from equation 3 that kco (the rate constant) only differs from DlCO/Va (= Kco) by three constant factors (Pb*, 1,000, and 1.2) and in its units. The ratios DlCO/Va (= Kco) to kco and to kco/Pb* (both constant except for minor variations in Pb) are given in the legend to Figure 1. Therefore, DlCO/Va (= Kco) is effectively the rate constant, representing alveolar carbon monoxide uptake efficiency. Unless required by the context, this review uses the term Kco in preference to DlCO/Va.

The previous section has shown that kco (second−1 or minute−1), kco/Pb* (minute−1 ⋅ mm Hg−1), and DlCO/Va (= Kco) (ml minute−1 mm Hg−1 L−1 BTPS) are physiologically equivalent, except in their units, to the rate of removal of CO from alveolar gas, that is, the slope (on a semilogarithmic plot) of carbon monoxide uptake in Figure 1, labeled kco. Kco, expressed as kco, is the rate constant for alveolar CO uptake; Kco, expressed as DlCO/Va, is the carbon monoxide diffusing capacity per unit alveolar volume, at the alveolar volume (Va) at which the measurement is made; it remains, in essence, a pressure-adjusted rate constant for alveolar carbon monoxide uptake. The difficulty, or confusion, stems from the notion that “per unit volume” implies DlCO corrected for lung volume, a concept that is wrong because DlCO measured at a different volume, at a different level of Va/VaTLC, would yield a different value for DlCO/Va (= Kco) (see Figures 2 and 3). Paradoxically, DlCO/Va contains no information about the value of Va, being a weighted mean value of the rate of CO uptake in the “accessible” Va. Therefore, it would be prudent to replace the misleading (although physiologically correct) “diffusing capacity per unit alveolar volume” by Kco, which, unlike the earlier term kco, is numerically the same as DlCO/Va.

How should the Kco be defined? Krogh (1) called k (= kco) “permeability,” and Kco has been referred to as the Krogh factor (6). Cotes (7) and others (8) refer to TlCO/Va (∼DlCO/Va) as the “transfer coefficient.” Hughes and Pride (2) referred to Kco as “essentially the rate constant for alveolar CO uptake.” No clarifying definition has emerged from the American Thoracic Society/European Respiratory Society (ATS/ERS) Task Force on Standardization of Lung Function Testing (9, 10), who still refer to Kco or DlCO/Va, like most authors, as “diffusing capacity per unit alveolar volume.” We regard the Kco as an index of the efficiency of alveolar transfer of carbon monoxide (approximately the rate of CO uptake); “transfer” is a better term than “diffusion” because of the importance of the reaction rate of carbon monoxide with pulmonary capillary blood (see Equation 4). Nevertheless, “rate constant for carbon monoxide uptake” is probably the best operational definition for the Kco.

Effect of Lung Volume

As the lung volume decreases from TLC to FRC, the DlCO falls and Kco rises (8, 11) (Figure 2). Expressed as a percentage of the value at predicted TLC (∼Vamax), DlCO at 50% Vamax is 79%, and Kco is 158% (8). This increased efficiency of alveolar uptake of carbon monoxide (Kco) at resting breathing volumes protects the DlCO against undue volume dependence, that is, DlCO is 80% of its TLC value at 50% Vamax rather than the expected 50%. The physiological reason for the increase in Kco with decreasing alveolar expansion is given in the Roughton–Forster (12) equation (1/Dl = 1/Dm + 1/θ⋅Vc), normalized to Va:

where Dm is the membrane diffusing capacity; θblCO is the reaction rate of carbon monoxide with blood (minute−1 mm Hg−1), adjusted to a standard hemoglobin (Hb) concentration; and Vc is the pulmonary capillary volume. With a decrease in alveolar expansion, the ratio Va/Dm remains almost constant (13), so the fall in Va/DlCO (= rise in Kco) is caused by the decrease in Va/Vc (= rise in Vc/Va), with Vc remaining constant as Va decreases (13). The change in Vc/Va is consistent with the stability of pulmonary blood flow (approximately the cardiac output) during lung volume changes.

Changes on Exercise

During exercise, DlCO (and Kco) rises at constant Va (14). This was first shown by M. Krogh in 1915 (1). The reason is that the rise of pulmonary artery (and, to a lesser extent, pulmonary venous) pressure, which accompanies the increase in pulmonary blood flow, distends the pulmonary capillary bed and recruits additional alveolar septal vessels (15). This increases capillary volume (Vc) and the membrane diffusing capacity (Dm) (14). On exercise at constant Va, Vc/Va increases; Dm/Va also increases because vascular distension expands the alveolar surface available for gas exchange. Thus, DlCO/Va (Kco) increases. With the rebreathing technique, usually used in exercise studies for measuring DlCO (14), mean Va does not change from rest to exercise (14, 16), being mostly constrained by the volume of the rebreathing bag, but Va did increase on exercise according to the open-circuit DlCO method (16); in this case, the increase in Va would itself contribute to the increase in DlCO, although its effect would be reduced by a fall in Kco accompanying the rise in Va.

Variables That Can Be Controlled

Other factors that influence Kco (but not Va) are anemia and alveolar Po2 because θblCO (see Equation 4) decreases as [Hb] falls or as alveolar Po2 rises. A rise in the steady state Pco in plasma (called the “back-pressure”), due to recent cigarette smoking or multiple preceding measurements of DlCO, also lowers the DlCO. Standard corrections for these three factors are available (9). The test gas includes 21–25% oxygen (depending on the helium concentration), so Kco is usually measured at a normal alveolar PaO2. Kco is greater supine than erect, but clinical measurements are always made in the seated upright posture.

Reference Values

The predictions for DlCO depend on age, sex, and height (17). Of the components of the DlCO, Va depends on sex and height but not on age, and, in adults, Kco depends inversely on age and height but, in a review of the literature, hardly at all on sex (18). The highest values for Kco have been found in boys and girls before the age of puberty (6), suggesting that the pulmonary capillary bed has developed earlier than alveolar volume. The decline in Kco in adults with age may be related to changes in the microvasculature, secondary to the loss of lung elasticity with aging. The inverse relationship with height for Kco may be because the apices of the lungs are less well perfused in the upright position in taller people for gravitational reasons. There is considerable scatter in the predicted values for different reference equations for DlCO and Kco, and there is no consensus on the “best choice” (10). Thus, there is a need to acquire new reference values for DlCO and for its components. The European Standardization Working Party (17) recommends that Kco (predicted) be calculated as DlCO (predicted)/TLC (predicted), from measurements made at different times and often in different places. Predicted values for Kco would be better based on the two simultaneous measurements, that is, from DlCO divided by single-breath “accessible” Va rather than from two separate procedures (DlCO and TLC).

Nomenclature and Units

This review refers to the DlCO as the carbon monoxide diffusing capacity, and uses traditional units (ml and mm Hg). In Europe, the DlCO is termed the “carbon monoxide transfer factor” (TlCO) and SI units are used for gas uptake (mmol) and pressure (kPa). Divide by 3.0 to convert traditional to SI units.

Clinical Causes of Decreases or Increases in Kco

Alveolar and/or microvascular damage and destruction, leading to loss of alveolar or capillary surface area, affecting both Dm and Vc, reduce the rate of carbon monoxide uptake per unit volume, leading to a low Kco as a percentage of the predicted value; in some circumstances, Kco may exceed the upper limit of normal at predicted TLC, and this has clinical significance Table 1 [1932]).


Low Kco
High Kco
MechanismClinical ExamplesMechanismClinical Examples
With Normal or Near Normal Va
Microvascular destructionIdiopathic pulmonary hypertension (19)Increased pulmonary blood flow or redistributionLeft-to-right intracardiac shunts (26)
Pulmonary vasculitis (20)Asthma (27)
Microvascular remodeling and dilationHepatopulmonary syndrome (21, 22)
Pulmonary arteriovenous malformations (23)
With Reduced Va
Alveolar destructionEmphysema (low “accessible” Va)Incomplete alveolar expansion to TLCInspiratory muscle weakness (28)
Chest wall restriction (29)
Poor cooperation or comprehension
Alveolar destructionDiffuse interstitial lung disease with fibrosisIncreased pulmonary blood flowPneumonectomy (30)
Microvascular destructionBronchiolitis obliterans (24)Microvascular congestion/dilationObesity (31)
Microvascular destructionChronic heart failure (severe) (25)Alveolar hemorrhageAnti-GBM disease (32), SLE

Definition of abbreviations: GBM = glomerular basement membrane; SLE = systemic lupus erythematosus.

Clinical examples are not an exhaustive list.

In relation to increases in Kco, incomplete alveolar expansion, without compromise of alveolar structure, elevates Kco by increasing Vc/Va; a lesser increase in Vc/Va is also largely responsible for the increase in Kco with increases in pulmonary blood flow, either through the whole lung, as in a left-to-right shunt, or through part of the lung, as after a pneumonectomy. The increase in Kco (and also DlCO) in asthma is probably linked to better perfusion of the apices of the lungs (27), and this may explain, in part, the increase in Kco in some obese patients, although a raised capillary volume and low Dm have been found (33), suggesting an element of pulmonary vascular congestion as in chronic heart failure (34, 35).

Diversion of blood flow from a resected lung, for example, pneumonectomy, increases perfusion per unit volume in the remaining lung by 80–100%, depending on the preoperative partitioning of flow between the two lungs, and assuming total pulmonary blood flow (∼cardiac output) remains the same postpneumonectomy. This will increase the Kco in the lung that remains. Corris and colleagues (30) established an empirical relationship in 28 patients for the increase in Kco that occurred postpneumonectomy:

where x was the percentage flow (%) to the resected lung, based on a preoperative radioisotope lung perfusion scan. Kco postpneumonectomy was 110–131% predicted (mean Kco preoperatively for both lungs averaged 98%); in the case in which flow to both lungs was equal preoperatively (x = 50%), Equation 5 predicts the Kco in the remaining nonresected lung to increase by +22%, that is, in an average case to 120% predicted (98 + 22%). The reason for this increase in Kco is the expected doubling of blood flow per unit volume in the remaining lung. This ΔKco is consistent with the 20% increase in Kco when pulmonary blood flow in normal lungs increases from 5 L minute−1 at rest to 10 L minute−1 on moderate exercise (14). Va after pneumonectomy averaged 50% of the preoperative value; thus, the remaining lung was expanded to its predicted TLC when the Kco was measured postoperatively. Note that for a similar reduction of overall Va to 50% predicted, but applied to both lungs by underexpansion (e.g., neuromuscular disease), the increase in Kco at 50% Va/VaTLC is considerably greater (+58%: see Figure 2) than the +22% occurring postpneumonectomy.

The effect of 50% volume loss from two different causes, (1) reduced alveolar expansion and (2) “loss of units” (pneumonectomy), is illustrated for DlCO in Figure 3A and for Kco in Figure 3B. The difference for Kco in Figure 3B arises from different changes in the two components of the Kco from the Roughton–Forster formula (Equation 4), Va/Dm and Va/Vc. With restricted alveolar expansion, Dm/Va (inverse of Va/Dm) and Vc (13) remain relatively constant; hence halving lung volume (to 50% Va/VaTLC) will increase Vc/Va to 200% and increase Kco to 158% (Figure 3B). After pneumonectomy, the whole cardiac output must be distributed to the remaining lung whose blood flow, per unit volume, probably doubles. A doubling of pulmonary blood flow during moderate exercise in normal subjects increases the Kco to 120%; this arises from changes in both the Dm and Vc components of the Roughton–Forster equation: Dm/Va increases to 133% and Vc/Va to 141% of their resting values (14). The larger increase in Vc/Va at 50% Va/VaTLC with underexpansion (200%) compared with exercise (141%), and, by implication, postpneumonectomy may arise because the number of alveoli and alveolar capillaries in two lungs is twice the number postpneumonectomy.

The increase in Kco postpneumonectomy (30) (increased blood flow per unit volume) is a general phenomenon in many lung diseases in which blood flow is redistributed to less diseased areas with an increase in local flow and blood volume per unit alveolar volume; this redistribution may be the explanation for increases in Kco seen occasionally in other conditions in which interstitial or vascular disease, in its early stages, is patchy, leading to blood flow diversion to the remaining normal lung. Thus, a normal or mildly elevated Kco is seen in a proportion of cases with sickle cell disease (36), interstitial lung disease, and sarcoidosis (37).

Pulmonary hemorrhage (32), in which blood recently shed from capillaries takes up carbon monoxide, is the one example of a raised Kco that is not linked to an increased rate of alveolar–capillary uptake. Kco is more sensitive than DlCO in detecting pulmonary hemorrhage (38) because of a small accompanying fall in Va. In 39 patients, the maximal increase above baseline averaged 219% for Kco but only 182% for DlCO. In nine patients the peak rise in DlCO was less than 50%, but the rise in Kco above baseline always exceeded 50%.

Clinical Causes of a Low “Accessible” Alveolar Volume (Va)

In the single-breath DlCO, there are three distinct causes of a low Va (as a percentage of Vamax predicted, ∼93.5% ± 6.6 [1.0 SD] TLC) (Table 2) resulting in different values for the Kco (see Table 1):

  1. Incomplete alveolar expansion (Kco > 120% predicted).

  2. Loss of lung units (Kco 100–120% predicted). Besides pneumonectomy, localized destruction of lung ± fibrosis, infiltration with granulomas or inflammatory exudates, atelectasis, alveolar edema, and pneumonic consolidation are other causes.

  3. Poor mixing with maldistribution of inspired gas. This is most obvious in the case of a bulla. But, intrapulmonary airflow obstruction from any of the major causes (emphysema, bronchitis, bronchiolitis, bronchiectasis, asthma) generally lowers the Va/TLC ratio, when Va is measured with 10-second helium dilution and TLC with body plethysmography or multibreath inert gas wash-in or washout (4). Va, even in normal subjects, is an “accessible” rather than an absolute volume. The Kco is variable and depends on the pathology (Table 2). But, clearly there is a continuum in the sense of different values of Va and Kco within a single diagnostic category.

These three causes may coexist: causes 1 and 2 in interstitial lung disease, and causes 2 and 3 in COPD or bronchiectasis.


PathophysiologyClinical ExamplesKco as % Kco at Predicted TLC
Restrictive (Reduced TLC)
Incomplete alveolar expansionInspiratory muscle weakness120–140 (28)
Chest wall, pleural restriction
Inadequate inspiration to TLC
Loss of units “localized”Pneumonectomy, local destructive or infiltrative pathology100–120 (30)
Loss of units “diffuse”Interstitial lung disease with fibrosis<80
Obstructive (Normal/High TLC)
Poor mixing + normal alveolar functionAsthma100–120 (27)
Poor mixing + localized loss of unitsBronchiectasis90–100
Poor mixing + some alveolar loss/disorganizationBronchiolitis obliterans70–100 (24)
Poor mixing + diffuse alveolar disorganizationCOPD (chronic bronchitis and emphysema)40–90

Definition of abbreviations: COPD = chronic obstructive pulmonary disease; Kco = rate constant for carbon monoxide uptake per unit barometric pressure (kco/Pb* ∼ DlCO/Va); Va = alveolar volume.

D. Cramer, Royal Brompton Hospital, London, UK, personal communication.

The DlCO is the product of its two components, Kco and Va (Equation 1). The most compelling argument in favor of the Kco (unadjusted) is set out in Table 3, where the same value of DlCO (as a percentage of the predicted value) may occur from different combinations of its components (Kco and Va). The combination of low Va and high Kco has a different clinical significance (extrapulmonary restriction) compared with the combination of low Kco and normal Va (microvascular injury), although the DlCO is practically the same.


Diagnosis% Predicted% Predicted% PredictedComment
A. Inspiratory muscle weakness5912050Lack of alveolar expansion
B. Pneumonectomy5811151Localized loss of lung units
C. Diffuse interstitial lung disease548466Alveolar capillary damage (±loss of units)
D. Emphysema545991Alveolar capillary damage
E. Idiopathic pulmonary hypertension565896Microvascular damage

Definition of abbreviations: DlCO = single-breath diffusing capacity for carbon monoxide; Kco = rate constant for carbon monoxide uptake per unit barometric pressure (kco/Pb* ∼ DlCO/Va); Va = alveolar volume.

In chronic inspiratory muscle weakness (28, 39), the Kco is usually less (120–130%) (Table 3, diagnosis A) than that predicted from the decrease of Va (Kco predicted would be 150%; Figure 3B), presumably due to secondary changes stemming from microatelectasis, retention of secretions, and infection. In interstitial lung disease (Table 3, diagnosis C), especially preceding the overt fibrotic phase, the Kco may be within the “normal” range (say 80–100%), but in the presence of a low Va, this could be interpreted as “abnormal” because the expected compensation via the “loss of units” model is lacking. In emphysema (in this example) (Table 3, diagnosis D) there is relatively little gas mixing deficit after inspiration to TLC, and Kco predicted is less than Va predicted, suggesting disorganization of peripheral airspaces, which remain (mostly) ventilated. This contrasts with Table 3, diagnosis C, in which the DlCO is similar, but Kco is higher than the Va. This suggests that the disease is more localized with up to 30% of alveolar units destroyed or infiltrated with inflammatory exudate (gas mixing from the Va/TLC ratio [data not shown] is normal), and that the remaining alveolar units are functioning well, even if not entirely normally, as gas exchange units. The analysis adds less in Table 3, diagnosis E, in which a low DlCO in the presence of normal lung volumes without airflow obstruction suggests straightaway some pulmonary vascular pathology.

In an earlier section (Measurement of Kco and Va: Combining Va and Kco) we pointed out that current practice reports DlCO/Va (= Kco) literally as DlCO divided by Va with units ml minute−1 mm Hg−1 L−1; this redundancy of units (the units of DlCO/Va and Kco are essentially minute−1 mm Hg−1, that is, kco/Pb*; see Equation 2) has led to the idea that DlCO/Va “adjusts” or “corrects” the DlCO when the Va is lower than predicted. Because DlCO/Va (= Kco) is not a constant function versus Va (Figures 2 and 3), several authors (4042) have claimed that DlCO/Va has no clinical value, and even that the Kco is an “arithmetically flawed” index (7) (if this were the case, we would expect Kco × Va [= DlCO] to share this flaw). The confusion arises from the substitution for Kco of its equivalent (DlCO/Va), which gives the impression of a “volume correction.” The ATS/ERS Task Force (9, 10) counsels caution in the use of the DlCO/Va ratio, but nowhere is the connection made that the DlCO/Va is essentially a rate constant, similar to kco and kco/Pb* except in its units. It is clear that the nonlinear relationship between Kco and lung volume (Figure 2) precludes DlCO/Va from being a “volume correction” for the DlCO when Va is reduced, but Kco remains a true reflection of alveolar CO uptake efficiency at a given volume. In our opinion, the emphasis on DlCO/Va as a correction factor for lung volume is misconceived, and reflects a misapprehension of the physiology. Hence, we believe the term DlCO/Va should be replaced by the more informative term, Kco.

Corrections have been proposed on the basis of the relationship in normal subjects between change of lung volume and the change in DlCO/Va (Kco). A typical relationship (data from 24 subjects) is as follows (8):

where DlCOTLC and VaTLC are expected values for DlCO and Va at a normal predicted TLC. For a Va/VaTLC ratio of 0.5, DlCO would be multiplied by 1.26 to adjust for the volume reduction. The relationship for Kco was
Thus, Kco would be adjusted down at 0.5 Va/VaTLC by multiplying by 0.64 (1/1.57). Johnson (8) studied retrospectively the pulmonary function records of 2,313 patients, and analyzed subgroups of patients with asthma, emphysema, extrapulmonary restriction, interstitial lung disease, and lung resection. Before adjustment, there was wide dispersion between DlCO and Kco (as a percentage of the predicted value), with Kco (%) exceeding DlCO (%) by up to 50%. After adjustment of Kco for a low Va, DlCO and Kco tended to converge. Using a similar approach, Frans and colleagues (43) reported convergence of DlCO and Kco values after adjustment of both for the low Va in diffuse interstitial lung disease. Basically, these “corrections” return a high Kco in extrapulmonary restriction to values in the 90–100% predicted range, and in interstitial lung disease and sarcoidosis they adjust the Kco down from 75–105% predicted to 50–75% predicted, more in line with the uncorrected DlCO. The authors (8, 43) argue that “normal” values for DlCO/Va (unadjusted) in interstitial lung disease give the clinician a “false” impression that the gas-exchanging part of the lung is “healthy.” Stam and colleagues (44) studied patients who developed alveolar injury after bleomycin treatment. After bleomycin, the Kco reduction was greater (by up to 11%) if its value was referenced to the Kco at the actual lung volume rather than the predicted TLC. They concluded, like others (8, 43), that the volume restriction had “misleadingly” increased the Kco.

The flaw in the argument is that alveolar restriction by underexpansion is only one of at least three mechanisms causing a low Va (Table 2). For example, it is unlikely that the majority of the alveolar units contributing to the Kco in interstitial lung disease, pneumonectomy, or airflow obstruction from various causes are “underexpanded.” Hughes and Pride (2) presented corrections for a high Kco and low Va using two models (alveolar underexpansion, and increased pulmonary blood flow; based on Equation 5; see their Table 3), but this is hardly a practical solution for the clinician, and does not address the question of low Va caused by poor gas mixing.

It is not unreasonable to seek an interpretation of, or correction for, the DlCO when Va is reduced. For example, it would be legitimate to correct DlCO and Kco, using Equations 6 and 7, for underexpansion of the lung during breath holding (due to extrapulmonary restriction or technical artifact) provided that alveolar deflation was the sole cause of the low Va. Our contention is that any “correction” of the DlCO for volume (Va) must take into account the reason for the volume deficit—for 50% volume loss (Va = 0.5 VaTLC) DlCO/Va (= Kco) will be significantly greater in extrapulmonary restriction than after a pneumonectomy or maldistribution of inspired gas as in bullous emphysema. There is no easy solution to this problem. There is no “correct” way in which the rate constant (kco ∼ kco/Pb* ∼ Kco) can be properly adjusted for all the causes of low alveolar volume. A grasp of physiological principles (see Determinants of Kco in Normal Subjects) is the best way to understand the clinical significance of DlCO, Kco, and Va.

In the last two decades, the measurement of pulmonary diffusing capacity using nitric oxide (DlNO) has been introduced (45, 46). DlNO is 4 to 4.5 times greater than DlCO, partly because the physical diffusivity of nitric oxide is about twice that of carbon monoxide, and partly because red cell resistance to nitric oxide uptake is less than that to carbon monoxide (47) owing mostly to the much faster combination (by 280-fold) of nitric oxide with hemoglobin (Hb). Unlike DlCO, DlNO is Po2 independent (48). The low red cell resistance suggests that DlNO is measuring mostly the diffusive component of the alveolar to red cell transfer pathway, related to the surface area/thickness ratio of the blood gas barrier. Since the work of Roughton and Forster (12) this has been referred to as the membrane diffusing capacity (Dm). DmNO is related to the better known DmCO by α (= 1.97), the ratio of the physical diffusivities of nitric oxide and carbon monoxide in plasma, that is, DmNO/α = DmCO.

Guenard and colleagues (45) measured DlNO and DlCO simultaneously by the classical single-breath technique. Assuming DmNO/α = DmCO, they showed that the Roughton–Forster formulation (1/DlCO = 1/DmCO + 1/θblCOVc) could be rearranged:

Reasonable values of DmCO and Vc were obtained in normal subjects (45).

Although, for clinical interpretation, DlNO may be regarded as a surrogate for the membrane diffusing capacity (Dm), the notion that θblNO is infinite has been called into question. Measurements of DlNO before and after experimentally induced hemolysis (49) and after blood substitution, in anesthetized dogs, with cell-free heme-based oxyglobin (50), suggest that DlNO is not entirely “red cell independent.” After oxyglobin exchange transfusion, in the red cell–free state, DlNO increased 1.5 times (DlCO did not change), which suggests that DmNO is 1.5 times DlNO rather than its equivalent. It was suggested previously that DlNO might be a surrogate for DlO2 (51).

Because of reservations about the relevance of in vitro measurements of θblCO and θblNO to the in vivo situation (49, 50), interest is shifting from estimates of Dm and Vc toward the DlNO/DlCO ratio. Assuming, for clinical purposes, that θblNO is infinite so that DlNO = DmNO = DmCO⋅α, and from the Roughton–Forster equation for carbon monoxide (12):

Thus, the DlNO/DlCO ratio is weighted toward the Dm/Vc ratio and α (the NO/CO physical solubility ratio). It is also equivalent to the Kno/Kco ratio because Dl = K × Va, and Va is common to DlNO and DlCO when measured simultaneously by the standard single-breath technique with inhalation of nitric oxide and carbon monoxide. Measurements of the DlNO/DlCO ratio have been performed in normal subjects, at rest and during exercise (5355), and over a range of lung volumes (56, 57). The DlNO/DlCO ratio has been studied in several clinical situations. For example, the DlNO/DlCO ratio is increased in heavy smokers (58), otherwise healthy, and in diffuse parenchymal disease (59) and in chronic thromboembolic pulmonary hypertension (59), possibly because Vc/Va is reduced more than Dm/Va. In contrast, the DlNO/DlCO ratio is decreased at FRC versus TLC (56), the explanation being that DlNO is more sensitive to alveolar underexpansion than DlCO. For example, from VaTLC to Va50% TLC the DlNO declines by 43% versus 29% for DlCO (56). The reason is that the fall in DlCO is buffered by an increase in Kco (+35%) whereas Kno (being less influenced by the rise in Vc/Va) increases by only 10% (56). Thus, the DlNO/DlCO ratio could become a marker for extrapulmonary restriction.

The DlNO/DlCO ratio gives some insights into the components (Dm and Vc) of the Roughton–Forster equation in a single maneuver without the two-step approach with carbon monoxide at different alveolar Po2 as well as by-passing θblCO, the value of which is somewhat controversial (51, 52). Experience to date with Dm and Vc partitioning has been disappointing because commonly both change equally (the only notable example of discordance [Dm↓, Vc↑] being chronic heart failure [34, 35]); thus, we would expect a low DlNO/DlCO ratio in chronic heart failure, at least in the early stages. It is also possible that by factoring out Va, the DlNO/DlCO ratio (= Kno/Kco) may provide additional insight into other respiratory diseases.

The single-breath DlCO is, physiologically, the product of two simultaneous measurements: the rate of carbon monoxide uptake from alveolar gas to pulmonary capillary blood (kco), reexpressed per mm Hg alveolar dry gas pressure (Pb*) as kco/Pb*, and the “accessible” alveolar volume (Va), which approaches, in normal subjects, TLC. kco/Pb* is linked mathematically to DlCO/Va (= Kco). The term DlCO/Va is misleading because, as kco/Pb*, it reflects the rate of alveolar uptake of CO.

The common causes of a low Va are (1) underexpansion of alveoli in relation to their predicted TLC, (2) loss of alveolar units by destruction or infiltration with exudates or transudates, (3) poor gas mixing and penetration during the 10-second single-breath maneuver, and (4) some combination of cases 1, 2, and 3. Thus, there is no single factor or equation with which the pulmonary function laboratory can “correct” or “adjust” DlCO for all the causes of a low Va, and the use of the term DlCO/Va should be replaced by its alternative, Kco.

Clinical interpretation of a low DlCO (as a percentage of the predicted value) stems from inspection of the components of the DlCO (Kco and Va) and knowledge of physiological principles. From a consideration of the Kco and Va (as a percentage of the predicted value), together with spirometry and lung volume measurements, it should be possible to distinguish emphysema from bronchiectasis, bronchiectasis from asthma, diffuse interstitial lung disease from extrapulmonary restriction, and both from pulmonary microvascular disease.

In the future, the diffusing capacity for nitric oxide (DlNO) may enable us to focus on alveolar structure differently from the DlCO. The DlNO/DlCO ratio may be a surrogate for the Dm/Vc ratio and DlNO may provide information on total barrier (tissue and blood) thickness, largely independent of any chemical resistance introduced by the presence of hemoglobin.

1. Krogh M. The diffusion of gases through the lungs of man. J Physiol 1915;49:271296.
2. Hughes JMB, Pride NB. In defence of the carbon monoxide transfer coefficient KCO (TL/VA). Eur Respir J 2001;17:168174.
3. Fitting J-W. Transfer factor for carbon monoxide: a glance behind the scene. Swiss Med Wkly 2004;134:413418.
4. Ogilvie CM, Forster RE, Blakemore WS, Morton JW. A standardized breath holding technique for the clinical measurement of the diffusing capacity of the lung for carbon monoxide. J Clin Invest 1957;36:117.
5. Roberts CM, MacRae KD, Seed WA. Multi-breath and helium single breath dilution lung volumes as a test of airway obstruction. Eur Respir J 1990;3:515520.
6. Rosenthal M, Cramer D, Bain SH, Denison D, Bush A, Warner JO. Lung function in white children aged 4 to 19 years. II. Single breath analysis and plethysmography. Thorax 1993;48:803808.
7. Cotes JE. Carbon monoxide transfer coefficient KCO (TL/VA): a flawed index. Eur Respir J 2001;18:893894.
8. Johnson DC. Importance of adjusting carbon monoxide diffusing capacity (DlCO) and carbon monoxide transfer coefficient (K CO) for alveolar volume. Respir Med 2000;94:2837.
9. MacIntyre N, Crapo RO, Viegi G, Johnson DC, van der Grinten CPM, Brusasco V, Burgos F, Casaburi R, Coates A, Enright P, et al.. Standardization of the single breath determination of carbon monoxide uptake in the lung (ATS/ERS Task Force: Standardisation of Lung Function Testing). Eur Respir J 2005;26:720735.
10. Pellegrino R, Viegi G, Brusasco V, Crapo RO, Burgos F, Casaburi R, Coates A, van der Grinten CPM, Gustaffson P, Hankinson J, et al.. Interpretive strategies for lung function tests (ATS/ERS Task Force: Standardisation of Lung Function Testing). Eur Respir J 2005;26:948968.
11. McGrath M, Thomson ML. The effect of age, body size and lung volume change on alveolar–capillary permeability and diffusing capacity in man. J Physiol 1959;146:572582.
12. Roughton FJW, Forster RE. Relative importance of diffusion and chemical reaction in determining rate of exchange of gases in the human lung. J Appl Physiol 1957;11:290302.
13. Stam H, Kreuzer FJA, Versprille A. Effect of lung volume and positional changes on pulmonary diffusing capacity and its components. J Appl Physiol 1991;71:14771488.
14. Hsia CCW, McBrayer DG, Ramanathan M. Reference values of pulmonary diffusing capacity during exercise by a rebreathing technique. Am J Respir Crit Care Med 1995;152:658665.
15. Glazier JB, Hughes JMB, Maloney JE, West JB. Measurements of capillary dimensions and blood vlolume in rapidly frozen lungs. J Appl Physiol 1969;26:6576.
16. Snyder EM, Johnson BD, Beck KC. An open-circuit method for determining lung diffusing capacity during exercise: comparison to rebreathe. J Appl Physiol 2005;99:19851991.
17. Cotes JE, Chinn DJ, Quanjer PH, Roca J, Yernault J-C. Standardization of the measurement of transfer factor (diffusing capacity). Eur Respir J 1993;6:4152.
18. Hughes JMB, Pride NB. Carbon monoxide transfer coefficient (transfer factor/alveolar volume) in females versus males. Eur Respir J 2003;22:186189. [Published erratum appears in Eur Respir J 22:570.]
19. Steenhuis LH, Groen HJM, Koëter GH, van der Mark Th W. Diffusion capacity and haemodynamics in primary and chronic thromboembolic pulmonary hypertension. Eur Respir J 2000;16:276281.
20. Steen V, Medsger TA. Predictors of isolated pulmonary hypertension in patients with systemic sclerosis and limited cutaneous involvement. Arthritis Rheum 2003;48:516522.
21. Whyte MKB, Hughes JMB, Peters AM, Ussov W, Patel S, Burroughs AK. Analysis of intrapulmonary right to left shunt in the hepatopulmonary syndrome. J Hepatol 1998;29:8593.
22. Rodriguez-Roisin R, Krowka MJ. Hepatopulmonary syndrome—a liver-induced lung vascular disorder. N Engl J Med 2008;358:23782387.
23. Dutton JAE, Jackson JE, Hughes JMB, Whyte MKB, Peters AM, Ussov W, Allison DJ. Pulmonary arteriovenous malformations: results of treatment with coil embolization in 53 patients. AJR Am J Roentgenol 1995;165:11191125.
24. Schwarer AP, Hughes JMB, Trotman–Dickenson B, Krausz T, Goldman JM. A chronic pulmonary syndrome associated with graft-versus-host disease after allogeneic marrow transplantation. Transplantation 1992;54:10021008.
25. Niset G, Ninane V, Antoine M, Yernault J-C. Respiratory dysfunction in congestive heart failure: correction after heart transplantation. Eur Respir J 1993;6:11971201.
26. De Troyer A, Yernault J-C, Englert M. Mechanics of breathing in patients with atrial septal defect. Am Rev Respir Dis 1977;115:413421.
27. Collard P, Njinou B, Nejadnik B, Keyeux A, Frans A. Single breath diffusing capacity for carbon monoxide in stable asthma. Chest 1994;105:14261429.
28. Hart N, Cramer D, Nickol AH, Moxham J, Polkey MI, Pride NB. Effect of pattern and severity of respiratory muscle weakness on carbon monoxide transfer and lung volumes. Eur Respir J 2002;20:9961002.
29. Siegler D, Zorab PA. The influence of lung volume on gas transfer in scoliosis. Br J Dis Chest 1982;76:4450.
30. Corris PA, Ellis DA, Hawkins T, Gibson GJ. Use of radionuclide screening in the preoperative estimation of pulmonary function after pneumonectomy. Thorax 1987;42:285291.
31. Saydain G, Beck KC, Decker PA, Cowl CT, Scanlon PD. Clinical significance of elevated diffusing capacity. Chest 2004;125:446452.
32. Ewan PW, Jones HA, Rhodes CG, Hughes JMB. Detection of intrapulmonary haemorrhage with carbon monoxide uptake: application in Goodpasture’s syndrome. N Engl J Med 1976;295:13911396.
33. Oppenheimer BW, Berger KI, Rennert DA, Pierson RN, Norman RG, Rapoport DM, Kral JG, Goldring RM. Effect of circulatory congestion on the components of pulmonary diffusing capacity. Obesity (Silver Spring) 2006;14:11721180.
34. Puri S, Baker BL, Dutka DP, Oakley CM, Hughes JMB, Cleland JGF. Reduced alveolar–capillary membrane diffusing capacity in chronic heart failure: its pathophysiological relevance and relationship to exercise performance. Circulation 1995;91:27692774.
35. Guazzi M. Alveolar–capillary membrane dysfunction in heart failure: evidence of a pathophysiological role. Chest 2003;124:10901102.
36. Santoli F, Zerah F, Vasile N, Bachir D, Galacteros F, Atlan G. Pulmonary function in sickle cell disease with or without acute chest syndrome. Eur Respir J 1998;12:11241129.
37. Hughes JMB. Physiology and practice of pulmonary function. Lichfield, UK: Association for Respiratory Technology and Physiology (ARTP); 2011. p. 241.
38. Greening AP, Hughes JMB. Serial estimations of carbon monoxide diffusing capacity in intrapulmonary haemorrhage. Clin Sci 1981;60:507512.
39. Laroche CM, Mulvey DA, Hawkins PN, Walport MJ, Strickland B, Moxham J, Green M. Diaphragm strength in the shrinking lung syndrome of systemic lupus erythematosus. Q J Med 1989;71:429439.
40. Chinn DJ, Cotes JE, Flowers R, Marks A-M, Reed J. Transfer factor (diffusing capacity) standardized for alveolar volume: validation, reference values and applications of a new linear model to replace KCO (TL/VA). Eur Respir J 1996;9:12691277.
41. Plummer AL. The carbon monoxide diffusing capacity: clinical implications, coding, and documentation. Chest 2008;134:663667.
42. van der Lee I, Zanen P, van den Bosch JMM, Lammers JJ-W. Pattern of diffusion disturbance related to clinical diagnosis: the KCO has no diagnostic value next to the DlCO. Respir Med 2006;100:101109.
43. Frans A, Nemery B, Veriter C, Lacquet L, Francis C. Effect of alveolar volume on the interpretation of the single breath DlCO. Respir Med 1997;91:263273.
44. Stam H, Splinter TAW, Versprille A. Evaluation of pulmonary diffusing capacity in patients with a restrictive lung disease. Chest 2000;117:752757.
45. Guenard H, Varenne N, Vaida P. Determination of lung capillary blood volume and membrane diffusing capacity by measurement of NO and CO transfer. Respir Physiol 1987;70:113120.
46. Borland CD, Higenbottam TW. A simultaneous single breath measurement of pulmonary diffusing capacity with nitric oxide and carbon monoxide. Eur Respir J 1989;2:5663.
47. Carlsen E, Comroe JH. The rate of uptake of carbon monoxide and of nitric oxide by normal human erythrocytes and experimentally produced spherocytes. J Gen Physiol 1958;42:83107.
48. Borland CD, Cox Y. Effect of varying alveolar oxygen partial pressure on diffusing capacity for nitric oxide and carbon monoxide, membrane diffusing capacity and lung capillary volume. Clin Sci 1991;81:759765.
49. Borland CD, Dunningham H, Bottril F, Vuylsteke A. Can a membrane oxygenator be a model for NO and CO transfer? J Appl Physiol 2006;100:15271538.
50. Borland CD, Dunningham H, Bottril F, Vuylsteke A, Yilmaz C, Dane DM, Hsia CC. Significant blood resistance to nitric oxide transfer in the lung. J Appl Physiol 2010;108:10521060.
51. Hughes JMB, Bates DV. Historical review: the carbon monoxide diffusing capacity (DlCO) and its membrane (Dm) and red cell (θ⋅Vc) components. Respir Physiol Neurobiol 2003;138:115142.
52. Ceridon ML, Beck KC, Olson TP, Bilezikian JA, Johnson BD. Calculating alveolar capillary conductance and pulmonary capillary blood volume: comparing the multiple- and single-inspired oxygen tension methods. J Appl Physiol 2010;109:643653.
53. Zavorsky GS, Cao J, Murias JM. Reference values of pulmonary diffusing capacity for nitric oxide in an adult population. Nitric Oxide 2008;18:7079.
54. Aguilaniu B, Maitre J, Glénet S, Gegout–Petit A, Guénard H. European reference equations for CO and NO lung transfer. Eur Respir J 2008;31:10911097.
55. Zavorsky GS, Quiron KB, Massarelli PS, Lands LC. The relationship between single-breath diffusion capacity of the lung for nitric oxide and carbon monoxide during various exercise intensities. Chest 2004;125:10191027.
56. van der Lee I, Zanen P, Stigter N, van den Bosch JMM, Lammers JJ-W. Diffusing capacity for nitric oxide: reference values and dependence on alveolar volume. Respir Med 2007;101:15791584.
57. Glénet SN, Bisschop CD, Vargas F, Guénard HJP. Deciphering the nitric oxide to carbon monoxide lung transfer ratio: physiological implications. J Physiol 2007;582:767775.
58. van der Lee I, Gietema HA, Zanen P, van Klaveren RJ, Prokop M, Lammers J-WJ, van den Bosch JMM. Nitric oxide diffusing capacity versus spirometry in the early diagnosis of emphysema in smokers. Respir Med 2009;103:18921897.
59. van der Lee I, Zanen P, Grutters JC, Snijder RJ, van den Bosch JMM. Diffusing capacity for nitric oxide and carbon monoxide in patients with diffuse parenchymal disease and pulmonary arterial hypertension. Chest 2006;129:378383.
Correspondence and requests for reprints should be addressed to J. Michael B. Hughes, D.M., 4 Cedars Road, London, SW13 0HP, UK. E-mail:

Author Contributions: This review was conceived by J.M.B.H., who drafted the text, illustrations, and tables. N.B.P. contributed to discussion and modification of the concept, design, and interpretation; both authors have approved the final version.

CME will be available for this article at or at

Originally Published in Press as DOI: 10.1164/rccm.201112-2160CI on April 26, 2012

Author disclosures


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record