The application of the forced oscillation technique to assess reversibility of airflow obstruction was compared with that of indices of forced expiration and plethysmographic airway resistance (Raw). In 125 patients with airflow obstruction, we measured total respiratory resistance (Rrs) and reactance (Xrs), Raw and specific airway conductance (sGaw), maximal flow-volume curves and forced expiratory volume in 1 s (FEV1), before and 30 min after 2 x 20 micrograms salbutamol by MDI. Salbutamol induced significant change in mean value of all measured indices. The changes in impedance data consisted of decrease in mean value and of negative frequency dependence of Rrs, an increase in Xrs with slight decrease of its positive frequency dependence. Multivariate analysis of differences between pre- and postbronchodilator values showed that the single indices with the greatest sensitivity to detect the effect of salbutamol were, in decreasing order, (1) in relative change (% baseline value): Raw, Rrs at 6 Hz (Rrs6), forced vital capacity (FVC), FEV1, and (2) in absolute change: FVC, sGaw or Raw, Rrs6, FEV1, maximal expiratory flow (MEF50). The effect of salbutamol was described best in (1) by a combination of Raw and FVC and in (2) by sGaw and FVC. For individual detection of bronchodilator effect, threshold values were calculated from mean reproducibility of the three baseline values of the various indices, attempting to estimate whether response to a bronchodilator is statistically significant. The greatest number of significant responses were observed for Raw, sGaw, FEV1, and FVC in that succession, Rrs6 being markedly less sensitive. This discrepancy is due to the lack of Rrs6 response to bronchodilators in patients with severe airway obstruction.(ABSTRACT TRUNCATED AT 250 WORDS)