To the Editor:
In their recent “Perspective”, Hughes and Pride 1 described the carbon monoxide (CO) transfer factor (TL,CO) as the product of two primary measurements: alveolar volume (VA) and CO transfer coefficient KCO (TL/ VA). VA is primary in that it has a physiological identity. However, Forster 2 has pointed out that in the context of the measurement of transfer factor, the volume is merely that of the gas exchanger; it has no more physiological significance than can be attached to the volume (VA+Vbag) that is used for the rebreathing method. Forster 2 also warned against interpreting KCO as a separate index. Its status as a primary index would have been different if Krogh 3 had been right in her belief that TL was proportional to VA; in this case, KCO in healthy persons would have been constant. Unfortunately, she made few observations and many subsequent studies showed that she was wrong 2. The refutation of her hypothesis destroyed the case for regarding KCO as a primary physiological index. Instead, it is merely a stage in the measurement of the transfer factor for the lung as a whole. In this circumstance, the transfer factor per litre of lung volume is designated appropriately by the ratio TL/VA and not, as Hughes and Pride 1 have claimed, by KCO.
Since TL is not proportional to VA, the ratio TL/VA misrepresents the real relationship between the variables 4. In some circumstances, the resulting distortion can contribute to diagnosis. For example, in emphysema, the transfer defect appears to be amplified as the ratio is depressed if the lung volume is increased. In the converse circumstance of a small lung volume due to a restrictive defect, the distortion can lead to a transfer defect being overlooked as the ratio may then be normal or increased. These facts are not disputed by Hughes and Pride 1; instead, they point out that anyone using the ratio should choose the appropriate model for its interpretation. This is fine for lung function specialists, but not for peripheral users. At every European Respiratory Society annual conference that I have attended, speakers have referred to standardizing transfer factor for lung volume by using the ratio. The high profile given by our Editor to the defence of KCO of Hughes and Pride 1 will help to ensure that the practice continues!
The “defence” elegantly demonstrates that the ratio can be rationally interpreted. This involves making allowances for the anatomical and physiological factors that contribute to TL and choosing between four alternative models of lung function. In addition, distortion due to the faulty arithmetic model is minimized by expressing longitudinal change as a ratio.
The schema of Hughes and Pride 1 is of Byzantine complexity and is based on an index carbon monoxide transfer coefficient that is arithmetically flawed. Time will tell if it can provide more information than is contained in gas transfer factor for the lung and alveolar volume. Meanwhile, my colleagues and I have rightly been criticized for not qualifying the circumstances when it is appropriate to make an allowance for alveolar volume in patients based on that in healthy subjects. There are few data and a strong prima facie case for believing otherwise in some circumstances. However, Hughes and Pride 1 give several examples of where it is appropriate to make such an adjustment and cite an instance where this turned out to be so, despite a reasonable expectation to the contrary 5. It is our view that there will be others. This could be a fruitful field for further research. Meanwhile, any laboratory using carbon monoxide transfer coefficient should do so circumspectly. Gas transfer factor for the lung should also be reported and interpretation should take alveolar volume into account. There is no place for carbon monoxide transfer coefficient in respiratory surveys.
- © ERS Journals Ltd