Abstract
Physiological and clinical studies have shown that the standard method of measuring respiratory impedance by forced oscillation leads to less efficient control of the upper airway shunt effect than the head generator method. To test the effects of these two techniques in epidemiological studies, we compared, in a sample of 73 French agricultural workers, the values obtained with each method for five forced oscillation parameters: resistance, frequency dependence of resistance, inertance, compliance and resonant frequency. For these comparisons, subjects were classified according to four respiratory status factors: smoking status, cough, expectoration and airway obstruction assessed from the maximum expiratory flow volume curve. Logistic regression models using the set of four forced oscillation parameters (excluding resonant frequency, which is derived from compliance and inertance) were then used to analyse the ability of each method to classify the subjects in each group. Significant differences between the two methods were observed for the mean values obtained for all five parameters. However, when each parameter was considered separately, the correlations between the values for each method were significant. Each method possessed the necessary ability to separate subjects into our group classification, but the significant relationships were not always found for the same parameters. Finally, logistic regression models showed that the two methods led to almost the same classification of the subjects. According to our results, the standard method of applying forced oscillation at the mouth seems an acceptable device for measuring respiratory impedance for epidemiological purposes.
