[Forced expiration index, signal and noise]

Although forced expiration measurements are extensively used, there is no general agreement concerning the best way to quantify the data. In this context, it may be of interest to examine indices from the point of view of their signal/noise ratio. The signal depends upon the actual sensitivity of the index to the kind of abnormality which is to be detected. In general, the relationship between signal and degree of abnormality is strongly non-linear, so that sensitivity is not a constant. The noise depends upon the kind of investigation which is made. When a subject is compared to himself (bronchomotor challenge, follow-up studies, etc.), it is mainly due to intraindividual variability. FEV1 is an example of index with a low sensitivity to peripheral airway abnormality, but a high reproducibility, so that its signal/noise ratio for paired measurements is comparatively good. When two groups are to be compared, the relevant noise is interindividual variability. Coefficient of variation of transit times and slope ratios are quite effective in detecting abnormalities in young smokers, probably because their sensitivity to mild peripheral airway disease is large compared to their interindividual variability. It follows that, according to the type of study, as well as to the nature and extent of abnormality, many indices may qualify as being the best.