Within-breath variations of forced oscillation resistance in healthy subjects.

Respiratory resistance (Rrs) was measured by the forced oscillation technique at 10, 20 and 30 Hz in 54 healthy subjects. The sinusoidal pressure oscillations were applied around the head, rather than at the mouth, so as to minimize transmural pressure across extrathoracic airway walls and the corresponding artefact (Peslin et al., J Appl Physiol, 1985, 59, 1790-1795). The flow (V') and volume (V) dependences of Rrs during the respiratory cycle were analysed by least square regression according to: Rrs = K1 + 2.K2.[V']#- K3.V, where K1 and K2 are Rohrer's constants, and where K3 expresses the (negative) volume dependence of Rrs. The analysis was made separately on the inspiratory and expiratory phases. A good fit was usually found between the data and the model, with a root-mean-square error averaging 15% of the mean Rrs at 10 Hz. At all frequencies K2 and K3 were substantially and significantly larger, and K1 slightly lower during expiration than during inspiration. Rrs, K1 and K3 were minimum at 20 Hz, while K2 exhibited a strong positive frequency dependence. The decrease of Rrs from 10 to 20 Hz was entirely explained by the variations of its linear component, and its increase from 20 to 30 Hz was largely due to its flow dependent component. Both the phasic variations and the frequency dependence of the coefficients suggest that the model is purely descriptive and that coefficients K2 and K3 reflect a number of phenomena, including the variations in glottic aperture during the respiratory cycle.