Parallel Offline Breath Sampling for Cross-Validated Analysis of Volatile Organic Compound Metabolites

Volatile organic compounds (VOCs) in breath are potential biomarkers for medical conditions that may be used for non-invasive health monitoring. One challenge that still exists is determining the fidelity of reported VOC biomarkers. The lack of universally accepted sampling methods makes it difficult to identify reliable candidates, thus allowing for the potential of false discovery.

The purpose of this study was to robustly profile VOCs in breath samples collected from relatively healthy participants using two offline methods for collection/analysis via solid phase microextraction (SPME) coupled to gas chromatography - mass spectrometry (GC-MS).

158 cross-sectional volunteers provided one-time samples using two methods, one which directly sampled breath via SPME and another which collected breath in Tedlar bags. Using both methods, 10 volunteers provided an additional nine longitudinal samples. Ambient air samples were collected routinely, and a robust data processing schematic was used to ensure high quality reporting of on-breath VOCs.

Data screening and processing led to the identification of > 30 unique VOCs in both methods. Hierarchical clustering and correlation analyses demonstrated volatile terpene/-oids showed homologous trends in both data sets. Of the 12 VOCs identified using both methods, 11 analytes displayed statistically significant correlations (p < 0.05) in healthy breath samples. Finally, both methods were benchmarked regarding VOC reproducibility, and analyses showed that longitudinally collected samples were more reproducible compared to cross-sectional.

The quantitative results from both sampling methods mirrored each other, thus increasing the reliability and fidelity of VOCs reported along with the results from biostatistical analysis.