Background: Asthma control does not yet meet the goals of asthma management guidelines. Non-invasive monitoring of airway inflammation may help to improve the level of asthma control in children.
Objectives: (1) To identify a set of exhaled volatile organic compounds (VOCs) that is most predictive for an asthma exacerbation in children. (2) To elucidate the chemical identity of predictive biomarkers.
Methods: In a one-year prospective observational study, 96 asthmatic children participated . During clinical visits at 2 month intervals, asthma control, fractional exhaled nitric oxide, lung function (FEV1, FEV1/VC) and VOCs in exhaled breath were determined by means of gas chromatography time-of-flight mass spectrometry. Random Forrest classification modeling was used to select predictive VOCs, followed by plotting of receiver operating characteristic-curves (ROC-curves).
Results: An inverse relationship was found between the predictive power of a set of VOCs and the time between sampling of exhaled breath and the onset of exacerbation. The sensitivity and specificity of the model predicting exacerbations 14 days after sampling were 88% and 75%, respectively. The area under the ROC-curve was 90%. The sensitivity for prediction of asthma exacerbations within 21 days after sampling was 63%. In total, 7 VOCs were selected for the classification model: 3 aldehydes, 1 hydrocarbon, 1 ketone, 1 aromatic compound, and 1 unidentified VOC.
Conclusion: VOCs in exhaled breath showed potential for predicting asthma exacerbations in children within 14 days after sampling. Before using this in clinical practice, the validity of predicting asthma exacerbations should be studied in a larger cohort.