Feed

Volatile organic compounds (VOCs) in exhaled breath reflect biological processes in the lungs. Exploring VOC profiles in severe asthma patients starting biologics may reveal treatment-related biological changes over time. This study aimed to assess the performance and feasibility of applying a sparse partial least squares-discriminant analysis (sPLS-DA) analytical approach in detecting longitudinal changes in exhaled VOCs of patients with severe asthma initiating biologic treatment. Exhaled breath was collected in the 3TR-ABC study, a multicenter observational prospective cohort study, at baseline and 4 months after starting biologics (benralizumab or dupilumab). VOCs were trapped on thermal desorption tubes and analyzed by gas chromatography-mass spectrometry. Paired sPLS-DA was used to distinguish baseline from post-treatment VOC profiles, and Kendall rank correlation assessed associations with clinical parameters at baseline. Data from two sites (Amsterdam and Copenhagen) were available of 25 severe asthma patients. The sPLS-DA resulted in a model containing five VOCs (methyl isobutyl ketone, 2-ethyl-1-hexanol, 5-methyl octadecane, 2-methylundecane and 1,1'-oxybis(decane)) with an area under the receiver operating characteristic curve of 0.89 (95% confidence interval: 0.79-0.99) for distinguishing baseline and 4 months post treatment. At baseline, a positive correlation was observed between 2-ethyl-1-hexanol and asthma control (Asthma Control Questionnaire); τ = 0.40, p < 0.01. A negative correlation was found between 1,1'- oxybis(decane) and blood eosinophil counts (τ = -0.30, p=0.04). Based on our results, the proposed analytical approach appears feasible for analyzing longitudinal data in patients with severe asthma initiating biologics. Further validation in larger cohorts is warranted to confirm robustness and generalizability.