Introduction

The nutrients we ingest also feed the microbiome that lives in our digestive system. The resulting metabolic products may have beneficial or deleterious effects on our health. For example, ethanol and 2-3-butanediol, produced by carbohydrate fermentation, were found to be associated with non-alcoholic steatohepatitis (NASH). Interestingly, several microbiome-derived metabolites are volatile organic compounds (VOCs) that can be measured in breath. In this ongoing study, we explored the feasibility of microbiome metabolic characterization before and after carbohydrate or amino acid ingestion.

Aims & Methods

A total of 13 healthy subjects were recruited with age ≥ 18 years and body weight ≥ 50 Kg. Subjects were instructed to fast overnight and to not drink alcoholic beverages the day before the experiments. Breath samples were collected before, and up to 90 minutes after ingestion of 75 g of glucose, cysteine, or tryptophan, with a time resolution of 5 minutes. Compounds of interest were measured using Soft Ionization Flow Tube Mass Spectrometry (SIFT-MS) with direct sampling. Results were expressed as part per billion (PPB v) as a function of time post-glucose ingestion.

Results

Intraoral swishing of 250 mg/mL glucose in water for 30 seconds, followed by spitting, elicits a peak of ethanol within 10-15 minutes that disappears after 25-30 minutes. However, ingestion of 300 mL of the same solution showed a variable response among different subjects. Some individuals showed a peak after 15 minutes, others, after 30-50 minutes (Fig. 1). Median (M) and interquartile range (IQR) baselines levels (before glucose administration) of ethanol, were 99.4 [71.6-182.4] PPB. Post glucose ingestion, breath peak levels ranged between 130, and 1600 PPB. Similarly, glucose administration induced spikes of acetoin, and propanoic acid, with ranges of 7-20 PPB and 18-160 PPB, respectively. Administration of tryptophan and cysteine elicits spikes of p-cresol, phenol, and hydrogen sulfide. Additional compounds were also detected, such as trimethylamine (M: 42.3, IQR: 33.6-54.3 PPB), methanol (M: 443, IQR: 199-531.4 PPB), and methane (M: 26988, IQR: 10100-76213 PPB).

Conclusion

Breath analysis is suitable for the clinical characterization of microbiome metabolism in the digestive system in healthy and disease populations to establish correlations between metabolites and disease severity and progression, as well as investigate the interaction of gut microbiota in response to therapeutic interventions. We aim to expand this approach to carbohydrates of different complexity, other amino acids, and nutrients (Table 1). This non-invasive method can replace the current need for blood collection enabling easier scaling to large cohort populations.

Disclosure

^​Tawfike A. ^​Murgia A. ^​Banda I. ^​Ahmed Y. ^​Manhota M. ^​Fonseca D. ^​Ricciardi F.Sweeney K. ^​Nicholson-Scott L. McConville L. ^​Gandelman O. ^​Allsworth, M. ^​Boyle B. ^​Ferrandino G. are employees of Owlstone Medical Limited