To the Editors:
We refer to the recent study by Oostveen et al. 1 on bronchodilator response in 4-yr-old children with different wheezing phenotypes. The forced oscillation technique is particularly suited to 2–7-yr-old children, and its suitability to assess bronchodilator responsiveness is still being investigated. Thus, the study is of great clinical relevance and, though of a narrow age range, is a welcome step forward from previous studies characterising bronchodilator responses assessed using measures of airway resistance in preschool children 2–4. However, we would like to raise a point we feel to be of importance.
The authors argued for the case of using the absolute change pre- and post-bronchodilator to express a bronchodilator response in children assessed using forced oscillation technique (FOT) variables. They reasoned that expressing the response as a relative change from baseline homogenises the different responses of subjects with different baseline bronchomotor tone, reflected by their different baseline lung function values.
The reason we advocate the use of the relative over the absolute response assessed using airway resistance from oscillometric methods is that the absolute response is generally dependent on lung function at baseline 2, 4. Statistically, assessment of magnitude of a treatment effect should be adjusted for any differences between the groups or variability in the pretreatment values 5. For instance, for the forced expiratory volume in 1 s (FEV1) it has been pointed out that since the absolute change between repeated measurements is constant across patients with varying FEV1 levels, it is the absolute change that should be used when expressing any treatment differences 6. More accurately, expressing the bronchodilator response using the absolute change is valid provided this change is independent of the baseline value, otherwise the relative change should be used 7. Dependence on age or height can be accounted for using % predicted values. Note that expressing the bronchodilator response in terms of absolute changes in z-scores is statistically identical to using absolute changes when the data are homoscedastic, though it may still be useful to use z-scores when lung function is already expressed in this manner or if the data are heteroscedastic.
In more physiological terms, for airway resistance measures, the smaller the airway calibre (hence the larger the baseline resistance), the more profound the effect of any increases to airway calibre. Thus, the same change in airway calibre represents a larger effective change in a small airway than in a large airway. Bronchomotor tone should be reflected by changes to airway calibre upon the administration of a bronchodilator, rather than by baseline lung function per se. A larger baseline bronchomotor tone would thus result in a larger bronchodilator response and represents our signal of interest, but this signal needs to normalised by the effect of a smaller initial airway calibre. Therefore, we would argue that while expressing the response in terms of relative changes “homogenises” the difference between subjects, it is a difference attributable to a factor that is not the main effect of interest but rather can confound it, and therefore needs to be corrected for.
Until more data on how to express a bronchodilator response assessed using airway resistance become available, we would recommend that authors of individual studies first ascertain whether the change with bronchodilator is related to factors such as age, height and baseline lung function. If yes, then these factors should be taken into account, especially when subjects of a wide range of ages, height or baseline lung function are studied.
Footnotes
Statement of Interest
None declared.
- ©ERS 2010