Human volatilome analysis using eNose to assess uncontrolled asthma in a clinical setting.

Abstract

Analyses of exhaled volatile organic compounds (VOCs) have shown promising results when distinguishing individuals with asthma. Currently, there are no biomarkers for uncontrolled asthma. Therefore, we aimed to assess, in a real-life clinical setting, the ability of the exhaled VOC analysis, using an electronic nose (eNose), to identify individuals with uncontrolled asthma. A cross-sectional study was conducted, and breath samples from 199 participants (130 females, aged 6-78, 66% with asthma) were analysed using an eNose. A multivariate unsupervised cluster analysis, using the resistance data from 32 sensors, could distinguish three clusters of VOC patterns in the training and testing groups. Comparisons between the clusters were performed using the one-way ANOVA, Kruskal-Wallis and chi-squared tests. In the training set (n = 121), three different clusters covering asthma, lung function, symptoms in the previous 4 weeks and age were identified. The pairwise comparisons showed significant differences with respect to chest tightness during exercise, dyspnoea and gender. These findings were confirmed in the testing set (n = 78) where the training model identified three clusters. The participants who reported fewer respiratory symptoms (dyspnoea and night-time awakenings) were grouped into one cluster, while the others comprised participants who showed similar poor control over symptoms with the distribution of the individuals with asthma being significantly different between them. In a clinical setting, the analysis of the exhaled VOC profiles using an eNose could be used as a fast and noninvasive complementary assessment tool for the detection of uncontrolled asthma symptoms.