Carbon monoxide transfer coefficient (transfer factor/alveolar volume) in females versus males

To the Editors:

The single breath transfer factor of the lung for carbon monoxide (T_L,CO) is derived from the multiple of the carbon monoxide transfer coefficient (K_CO) and the alveolar volume (V_A) during breath holding 1; consequently, the T_L,CO is highly dependent on lung size (∼V_A), and thus is smaller in females for a given height and age. Essentially, the K_CO is the rate constant for alveolar carbon monoxide uptake (its units are per second or per minute) and may be expected to be the same in males as in females, as it was in the original description of the technique 2. The published guidelines 3, 4 are not definitive on this point, so we reviewed all studies of reference values for T_L,CO and K_CO, which have presented data for both males and females.

The European reference values for K_CO 3 are derived from the ratio T_L,CO predicted/total lung capacity (TLC) predicted. This unusual recommendation arose from correspondence in the ERJ 5, which pointed out that the reference equations previously published in 1983 6 produced values for K_CO that were incompatible with those for T_L,CO. These European Respiratory Society reference values have been quoted in 118 articles published since 1993 (23 in the ERJ) and have been adopted elsewhere 7. The T_L,CO values for females were based on a summary equation 6 derived from only five small studies, the most recent being published in 1979. Fortunately, the available database for T_L,CO and K_CO has now expanded considerably; seven studies that examined both males and females in their chosen populations have been published between 1980 and 1992 (table 1⇓).

In addition, we believe the use of the T_L,CO pred/TLC pred ratio is, itself, unjustified. First, because in general it is undesirable to ratio two reference values obtained from different populations; this is compounded in the present instance, when K_CO is inevitably measured in every individual in the actual manoeuvre used to obtain T_L,CO. Secondly, although the original recommendation 18 was that the lung volume during breath-holding (∼V_A) should be measured from the inspired volume preceding the breath hold plus the residual volume (from closed circuit helium dilution), virtually all subsequent population studies (and all those in table 1⇑) have substituted the V_A derived from the dilution of the inhaled marker gas. This single breath V_A, both in the derivation of T_L,CO and in the subsequent calculation of T_L/V_A, is always less than TLC, because the anatomic dead space is subtracted from the V_A but not from the TLC, and because gas mixing is often less efficient with a single breath measurement. Hence, using TLC when a single breath V_A has been used to calculate T_L,CO will lead to a systematic underestimate of predicted K_CO.

Table 1⇑ shows predicted values for K_CO at a standard age and height for males and females from 10 studies. Following American Thoracic Society Guidelines 19, we have compared males of 1.75 m stature with females of 1.65 m. When these studies are averaged (without weighting for the different numbers in each study) there is no sex difference for K_CO at age 45 yrs, although there is a small (nonsignificant) difference between males and females at 65 yrs. In contrast, from the T_L,CO pred/TLC pred ratio 3, the predicted K_CO in females at a given age is 5–10% higher than for males, irrespective of whether they are at identical TLC (approximate height for males 1.6 m versus 1.8 m for females) or at identical height (approximate TLC for males 6.1 L versus 5.1 L for females). Table 1⇑ also shows considerable between-study differences both in mean values and in the age and height coefficients, as is found for most other predicted values for tests of lung function 6. From first principles, a dependence on height would not be expected, but it has been suggested 16 that in taller people the apices of the lungs may be more poorly perfused relative to the mid and lower zones for gravitational reasons; the resulting inhomogeneity in blood flow and blood volume would reduce the measured K_CO. All published data on changes with age are derived from cross-sectional studies. The only cohort study of >8 yrs of T_L,CO and K_CO is that of Watson et al. 20 who followed up, among others, 29 male never-smokers (mean age 37 yrs at start) over a 22-yr period. They found no change in the K_CO.

We understand that a joint working party of the American Thoracic Society and the European Respiratory Society is currently reviewing reference values for spirometry, lung volumes and the transfer factor. We hope the carbon monoxide transfer coefficient will not be neglected, since the current recommendations are unsatisfactory. Ideally, values for total lung capacity should be obtained from the same individuals used to obtain reference values for the single breath transfer factor of the lung for carbon monoxide and the carbon monoxide transfer coefficient, so that the effects of poor inflation and/or true differences in total lung capacity at a given height and age can be allowed for 17.