No abstract

Invasive aspergillosis and other invasive fungal infections are associated with significant morbidity and mortality in immunocompromised patients, in large part due to limitations of existing diagnostic methods for these infections. Detection of species-specific volatile sesquiterpene metabolites of fungal origin in the breath of patients with invasive fungal infections allows the diagnosis and monitoring of these infections in vivo, non-invasively and more rapidly than possible with current diagnostic methods. While detection of exogenous microbial volatile metabolites in the breath has opened up a new and exciting dimension of diagnostic research and development in infectious diseases, we discuss the daunting challenges to volatile diagnostic biomarker discovery and clinical development.

No abstract

BackgroundCRE infections cause significant mortality, in large part because rapid identification of these infections is challenging. We tested the hypothesis that CRE and their isogenic carbapenem-susceptible counterparts have differential metabolic responses to carbapenem therapy.MethodsWe generated isogenic pairs of E. coli, E. cloacae, and K. pneumoniae by inserting a blaNDM-1-containing plasmid into carbapenem-susceptible E. coli, E. cloacae, and K. pneumoniae. We confirmed phenotypic meropenem (MPM) resistance per CLSI breakpoints for Enterobacteriaceae (MIC ≥4) in the NDM-1+ member and susceptibility (MIC≤1) in the NDM-1- member of each pair. We administered 2 × 108 CFU of each isolate intranasally to 23–28 g male C57BL/6J mice, infecting 6 mice with the NDM-1+ member and 6 with the NDM-1− member of each species pair (12 mice per bacterial species). 24 hours after infection, we treated 3 mice in each NDM-1+ and NDM-1− bacterial species cohort with MPM over 4 hours, and the other 3 mice in each cohort with saline over 4 hours as controls, confirming adequate infection (a target of 106 CFU/g of lung tissue) in quantitative lung homogenate cultures. We then collected breath samples from each mouse via tracheostomy using a murine ventilator, identifying all volatile metabolites in each sample using thermal desorption-gas chromatography/tandem mass spectrometry. We used Wilcoxon tests to examine differences in metabolite abundance between MPM and saline-treated control mice in the NDM-1+ and NDM-1− a member of each species pair, with a two-sided P-value threshold of < 0.1.ResultsSeveral breath volatile metabolites changed differentially within each NDM-1+/NDM-1- pair, outlined in Table 1 (E. coli), Table 2 (E. cloacae), and Table 3 (K. pneumoniae). Each listed metabolite that changed with MPM did not change with MPM in mice infected with each isogenic counterpartConclusionThere are differential in vivo metabolic responses with effective vs. ineffective treatment of mice with pneumonia caused by E. coli, E. cloacae, and K. pneumoniae pairs that are genetically identical other than blaNDM-1; this differential treatment response can potentially be used to identify these infections. Disclosures All authors: No reported disclosures.

BackgroundCDI is a frequent cause of morbidity and mortality in hospitalized patients. Despite advances in rapid CDI testing, there are often delays between the onset of symptoms and receipt of test results. We sought to test the hypothesis that the altered CDI intestinal microbiome has a unique volatile metabolite profile, distinct from the profile of patients with other causes of antibiotic-associated diarrhea, which potentially can be used to identify patients with CDI.MethodsWe prospectively collected fresh stool samples from inpatients with suspected CDI at an academic tertiary care hospital from July 2015 to November 2017, adsorbed volatile metabolites from each sample onto sorbent tubes within an hour of sample collection, and used thermal desorption-gas chromatography/tandem mass spectrometry to identify each metabolite. All patients were exposed to at least one antibiotic agent in the prior 90 days, and only patients receiving empiric CDI treatment or with formed stool samples were excluded. We used logistic regression models, adjusting for prior anti-anaerobic antibiotic therapy and CDI severity (serum albumin <3 g/dL and WBC ≥ 15,000/mm3 or abdominal tenderness) and adjusting for multiple testing using Storey’s q-value procedure (with a threshold of q ≤ 0.05), to examine the relationship between CDI, as determined by the reference standard of the cell culture cytotoxicity neutralization assay, and each metabolite.ResultsIn our 565-patient cohort, median age was 61 years (IQR 50, 70) and 277 (49%) were male; 173 (31%) had abdominal pain in the 24 hours before testing, 59 (10%) had fevers in the prior 24 hours, 22 (4%) had an ileus, 74 (13%) had mental status changes in the prior 24 hours, 89 (16%) were hospitalized in the ICU at the time of testing, 45 (7%) were receiving pressors, 82 (15%) had a WBC ≥ 15,000/mm3, and 137 (24%) had a serum lactate > 1.5 mmol/L. Ultimately, 155 patients were diagnosed with CDI. Ten metabolites (Table 1, Figure 1) were differentially distributed in patients with and without CDI.ConclusionWe identified a suite of volatile metabolites that differentiates stool from patients with and without CDI; this profile may ultimately be used to identify patients with CDI. Disclosures All authors: No reported disclosures.