Lung sounds during allergen-induced asthmatic responses in patients with asthma.

Abstract

We postulated that the distinct pathophysiologic mechanisms of airway narrowing during the early (EAR) and the late (LAR) asthmatic responses to inhaled allergens are reflected by the generation or transmission of lung sounds in asthma. Therefore, we measured FEV1 and recorded lung sounds in eight mildly asthmatic subjects before a standardized allergen challenge (PRE), during the EAR, during the recovery phase at 2 h (MID), during the LAR at 7 h, and after inhalation of a bronchodilator (POST). The recordings were made during flow- and volume-standardized quiet breathing, and during maximal forced breathing maneuvers. Airflow-dependent power spectra were analyzed for lung sound intensity (LSI), quartile power points (Q25, Q50, Q75), and extent of wheezing (W). These sound characteristics were compared among the various stages of the challenge in the presence (EAR, LAR) and absence (PRE, MID, POST) of acute airway obstruction using ANOVA. LSI, Q25 - Q75, and W were all elevated during airway obstruction. When matched for percent fall in FEV1, during the EAR and the LAR (mean +/- SD: 26.7 +/- 4.0, and 28.9 +/- 5.7, respectively; p = 0.385), the increase in Q25, and Q50 with airflow during quiet expiration, as well as the extent of wheezing, were higher during the LAR than during the EAR (p < or = 0.042 and p < or = 0.012, respectively). At similar levels of FEV1 (p > or = 0.156), LSI on expiration was higher at POST than at PRE or MID (p < or = 0.067), whereas Q25 (p < or = 0.047) and Q50 (p < or = 0.064) were lower at POST than at PRE. During forced expiration W was higher at MID and POST than at PRE (p < or = 0.014). We conclude that LSI, frequency content, and the extent of wheezing vary during the subsequent stages of allergen-induced bronchoconstriction in asthma despite matched values of FEV1. This suggests that airflow-standardized phonopneumography is a sensitive method for detecting differences in the pathophysiology of airway narrowing in asthma.

Detection of methacholine with time series models of lung sounds

“…The small hardware error has mostly been left in the signals which were analyzed, because sometimes the removal by decimating the data turned out to be disadvantageous for detection. Moreover, a comparison with previous studies of the same data [1], [2] remains possible. Synchronization of microphone data with flow measurements enables the detection of the transitions from expiration to inspiration.…”

“…The final goal of this research, however, might be the early detection of asthma by an automatic computer analysis of recorded lung sound signals. Using computerized phonopneumography, lung sounds have been analyzed in intensity and frequency content [1]. As expected, some differences are observed between healthy and asthmatic subjects.…”

“…In this case, the order 100 will be the best model order found with order selection criteria if sufficient data are available. But the best model order for the detection of two classes will probably be AR (1), because all differences between the two AR processes are concentrated in that single reflection coefficient. It is clear that this artificial example will not be representative of practical data.…”

“…The records have been analyzed before [1], [2]. It turned out to be rather difficult to discover reproducible effects of methacholine, that can be used for its detection.…”

“…The limited purpose of this paper is to investigate the feasibility of automatically detecting the presence of methacholine with time series models of lung sounds. Automatically means that no use is made of available a priori knowledge about data properties from present or previous research [1], [2], [4], [5]. Only expiration data from one subject have been used.…”