Résumé de recherche
Rationale
Recent advances in oscillometry have revealed the diagnostic potential of intra-breath changes in the mechanical impedance of the respiratory system (Zrs). In particular, single-frequency tracking oscillometry has shown that the changes in resistance (R) and reactance (X) between end-expiration and end-inspiration (AR and AX, respectively) are sensitive to lung inhomogeneities (Eur RespirJ 71: 907, 2017), and the areas of the X vs volume (AXV) and X vs flow loops (AXV') can quantify expiratory flow limitation. Additionally, AR and AX recorded during tidal breathing were highly sensitive and specific in the detection of airway obstruction in children where conventional oscillometry failed (Thorax 71: 907, 2016).
We aimed to examine the feasibility to extract intra-breath information from conventional multi-frequency oscillometry data.
Methods
Zrs was measured in healthy subjects (n=10), patients with asthma (n=13) and COPD (n=30) with the tremoFlo oscillometry device (Thorasys Inc., Montreal, QC, Canada) using a pseudorandom signal in the 5-37-Hz range. Once at least 3 reproducible Zrs spectra were collected from 16-s recordings, a 30-s measurement with a single 10-Hz signal was made with a research modality of the tremoFlo. From the former, the frequency component nearest to 10 Hz (11 Hz) was extracted and analysed in the time domain. In addition to the intra-breath measures AR, AX, AXV, and AXV' the standard mean R and X for the whole respiratory cycle (Rmean and Xmean, respectively) were calculated and compared in terms of variability using the interquartile range (IQR).
Results
The median values estimated at 11 Hz did not differ statistically significantly from the 10-Hz data despite an expected slight frequency dependence of Zrs, and their correlations were strong for all measures except AR and AX (Table 1). However, all IQR values were significantly larger in the extracted 11-Hz data. Conclusions This preliminary study indicates that even at lower signal/noise levels resulting in larger variabilities it is possible to obtain acceptable intra-breath data from multi-frequency oscillometry recordings. Therefore, re-processing of archives of conventional oscillometry may succeed in the extraction of Zrs measures that are able to identify subtle changes in lung function.
