Robin D. Couch, Karl Navarro, Masoumeh Sikaroodi, Pat Gillevet, Christopher B. Forsyth, Ece Mutlu, Phillip A. Engen, Ali Keshavarzian

Recent studies have illustrated the importance of the microbiota in maintaining a healthy state, as well as promoting disease states. The intestinal microbiota exerts its effects primarily through its metabolites, and metabolomics investigations have begun to evaluate the diagnostic and health implications of volatile organic compounds (VOCs) isolated from human feces, enabled by specialized sampling methods such as headspace solid-phase microextraction (hSPME). The approach to stool sample collection is an important consideration that could potentially introduce bias and affect the outcome of a fecal metagenomic and metabolomic investigation. To address this concern, a comparison of endoscopically collected (in vivo) and home collected (ex vivo) fecal samples was performed, revealing slight variability in the derived microbiomes. In contrast, the VOC metabolomes differ widely between the home collected and endoscopy collected samples. Additionally, as the VOC extraction profile is hyperbolic, with short extraction durations more vulnerable to variation than extractions continued to equilibrium, a second goal of our investigation was to ascertain if hSPME-based fecal metabolomics studies might be biased by the extraction duration employed. As anticipated, prolonged extraction (18 hours) results in the identification of considerably more metabolites than short (20 minute) extractions. A comparison of the metabolomes reveals several analytes deemed unique to a cohort with the 20 minute extraction, but found common to both cohorts when the VOC extraction was performed for 18 hours. Moreover, numerous analytes perceived to have significant fold change with a 20 minute extraction were found insignificant in fold change with the prolonged extraction, underscoring the potential for bias associated with a 20 minute hSPME.