Longitudinal Monitoring of Volatile Organic Compounds in Exhaled Breath Samples Collected from Children Diagnosed With Cystic Fibrosis

Abstract Introduction: Despite the introduction of disease-modifying treatments, people with cystic fibrosis (CF) remain at risk for decreased quality of life as well as life expectancy. People with CF experience pulmonary exacerbations (PEx), which are episodes of decreased lung function accompanied by cough and other respiratory/systemic symptoms. There are no existing biomarker-based tests for PEx diagnosis or monitoring response to treatment, and there is interest in developing a point-of-care test that can ultimately enable early treatment and recovery of lung function. Volatile organic compounds (VOCs) in exhaled breath may be a promising route for exploration, as they are noninvasive and a rich source of biomarkers for many conditions. Methods: Breath was collected longitudinally from children with CF ages 8-18 to benchmark VOC reproducibility while stable at baseline pulmonary health, to identify potential biomarkers of PEx, and define changes during PEx recovery. Whole breath samples were collected in 3L Tedlar bags, and VOCs were cryothermally transferred to headspace vials. To identify VOCs, samples were analyzed by gas chromatography-mass spectrometry. Biostatistical methods were employed to estimate VOC reproducibility as well as identify discriminatory features of PEx. Results: 55 breath samples were collected from a total of 22 children during clinical visits. Data screening revealed an abundant number of VOCs, with over 600 features present in at least 25% of all samples. Analysis was limited to 336 analytes (present in 50% of samples), and reproducibility was assessed for participants who were stable and not experiencing PEx. Relative standard deviation (RSD) values for intersubject measurements spanned 18-300% and were statistically significantly higher than intrasubject RSD values (p-value = 1.8x10-13). A total of 84 VOCs were differentially expressed between stable clinic visits and during PEx, with significant p-value or log2 fold change. These VOCs are visualized in a hierarchical heatmap, displaying relatively high interclass and low intraclass variation. Principal component analysis was implemented on this panel of VOCs, and the first dimension alone could distinguish PEx with over 80% accuracy. Conclusions: Breath collected from children with CF is a rich source of VOCs. Reproducibility for children with CF at baseline indicated that VOC measurements within individuals are significantly more reproducible than between individuals. Additionally, a pool of 84 VOCs was identified to be potentially useful for noninvasive monitoring of PEx. Future work will involve identifying these compounds as well as exploring their trends in samples collected from those receiving antibiotic treatments.