Specific Gut Microbes and Bile Acids Linked to Vancomycin-Resistant Enterococcus Risk

Intensive care unit patients who acquire vancomycin-resistant Enterococcus harbor specific baseline microbiome alterations, including lower ratios of Bacteroides to Enterococcus, reduced levels of Clostridium scindens, and depleted concentrations of secondary bile acids, according to a prospective observational investigation conducted at National Taiwan University Hospital. Researchers tracked adults admitted to a medical intensive care unit to determine whether distinct bacterial signatures and metabolic features precede intestinal colonization by this opportunistic pathogen. Vancomycin-resistant Enterococcus, widely designated as VRE, represents a substantial clinical challenge in critical care environments where intestinal expansion frequently precedes invasive infections. Prior animal models and laboratory experiments have suggested that anaerobic commensals and downstream metabolic byproducts contribute to colonization resistance, yet specific human predictive signatures have historically remained difficult to verify. By analyzing serial fecal specimens and surveillance screening swabs, the team documented that patients who began hospitalization free of the pathogen but subsequently acquired it underwent an accelerated decline in microbial diversity compared with patients who remained uncolonized throughout their intensive care stay.

The prospective study enrolled a total of one hundred and eight adult patients admitted to a medical intensive care unit between 2020 and 2021, an environment that predominantly admits individuals facing community-onset acute critical illnesses. At the time of enrollment, initial surveillance screening with rectal anal swabs identified thirty-four patients who were already colonized with vancomycin-resistant Enterococcus, and all thirty-four individuals remained colonized across their hospital admissions. Among the remaining seventy-four patients who initially screened negative, twenty-three individuals acquired the pathogen during their intensive care unit stay, whereas fifty-one patients maintained negative screening results across longitudinal monitoring. For the subset of patients acquiring vancomycin-resistant Enterococcus, the median duration from intensive care admission to the initial positive culture screen was six days, with an interquartile range spanning three to eight days. To delineate microbial dynamics over time, stool samples were systematically paired with regular anal swabs collected upon study admission and twice weekly until hospital discharge, yielding three hundred and two patient specimens alongside reference samples from twenty-seven healthy adult controls.

Taxonomic profiling of patient stool samples using sixteen-S ribosomal RNA gene sequencing revealed stark differences between hospitalized individuals and healthy ambulatory controls, as well as distinct microbial evolutionary trajectories among patient subgroups. When evaluating all three hundred and two collected intensive care specimens, patient samples exhibited significantly higher Enterococcaceae relative abundance, notably reduced Shannon alpha-diversity, and markedly elevated Berger-Parker dominance compared to healthy volunteer specimens. Across the three clinical cohorts, baseline stool collected upon intensive care unit admission demonstrated significant beta-diversity divergence driven primarily by individuals who were already colonized upon arrival. In contrast, patients who entered the unit uncolonized and later acquired vancomycin-resistant Enterococcus initially exhibited alpha-diversity levels that were statistically indistinguishable from those who remained uncolonized. However, longitudinal modeling using generalized estimating equations demonstrated that microbial diversity subsequently eroded far more rapidly in patients acquiring colonization, declining at a significantly steeper rate than observed among persistently uncolonized peers prior to the detection of the pathogen.

Longitudinal evaluations using generalized estimating equations established an average daily reduction in Shannon diversity of zero point zero seven across the overall critical care cohort during the initial two weeks of monitoring. Patients who entered the intensive care unit already harboring vancomycin-resistant Enterococcus consistently maintained lower overall Shannon diversity scores than those who remained negative throughout their stay. Crucially, the rate of diversity loss over time differed markedly between the initially uncolonized patient subsets, with individuals who ultimately acquired colonization exhibiting an additional daily slope decrease of zero point zero eight compared with persistently negative controls. Because of this accelerated microbial degradation, the final fecal specimens collected from patients who acquired the organism exhibited lower Shannon diversity and higher Berger-Parker dominance than persistently negative patients, converging toward the low-diversity profiles of patients colonized at baseline. Parallel evaluations restricted strictly to pre-colonization samples confirmed that this accelerated loss of intestinal richness was already active before diagnostic rectal swabs first detected the hospital-acquired pathogen.

Detailed community evaluations within the predictive cohort, which restricted analyses to patients who provided evaluable baseline fecal specimens prior to documented colonization events, revealed subtle baseline distinctions distinguishing patients who remained protected from those who succumbed to colonization. Although initial global beta-diversity evaluated by weighted UniFrac distances did not achieve a statistically significant difference between these two cohorts, principal coordinate analysis demonstrated distinct taxon loadings along the primary coordinate axis. Bacteroides taxa drove coordinate loadings pointing toward persistently negative patient samples, whereas Enterococcus taxa drove loadings toward patients who subsequently acquired vancomycin-resistant Enterococcus. While standalone absolute relative abundances of Bacteroides and Enterococcus genera did not differ significantly when evaluated individually, the baseline ratio of Bacteroides to Enterococcus was significantly elevated among patients who remained uncolonized. Conversely, broader evaluations of total obligate anaerobic bacteria, categorized computationally through the BugBase phenotype prediction database, showed no significant compositional difference between the cohorts, indicating that broad taxonomic categories were insufficient to capture protective signatures.

Applying analysis of compositions of microbiomes with bias correction identified specific bacterial taxa in baseline fecal samples that were significantly associated with subsequent resistance to vancomycin-resistant Enterococcus colonization. Among all detected organisms, the commensal species Clostridium scindens exhibited a statistically significant inverse association with pathogen acquisition, characterized by a false-discovery-rate adjusted P-value of zero point zero twenty-eight. Clostridium scindens is known for its specialized capacity to express enzymatic machinery required for secondary bile acid biosynthesis, transforming primary host bile acids into inhibitory metabolites. Notably, within this monitored intensive care predictive cohort, not a single patient with an initial Clostridium scindens relative abundance exceeding zero point zero one percent acquired vancomycin-resistant Enterococcus during their intensive care unit course. When researchers examined other known secondary bile acid-producing commensals within the same genus, including Clostridium hylemonae and Clostridium hiranosis, no protective correlation emerged, with Clostridium hylemonae demonstrating comparable abundance across cohorts and Clostridium hiranosis remaining entirely undetected across all fecal specimens.

Targeted biochemical quantification of secondary bile acids within initial fecal specimens corroborated the taxonomic associations identified through sequencing, underscoring a metabolic dimension to colonization resistance. Concentrations of deoxycholic acid and lithocholic acid, quantified via ultra-high-performance liquid chromatography coupled with tandem mass spectrometry, demonstrated a strong positive internal correlation with a Spearman rank correlation coefficient of zero point eighty-four. While independent numerical concentrations of deoxycholic acid and lithocholic acid each exhibited trends toward higher values in uncolonized individuals, their combined absolute concentration reached statistical significance, showing significantly higher baseline levels among patients who remained free of the pathogen. Furthermore, patient stratification into concentration quartiles demonstrated that individuals situated within the highest quartile for deoxycholic acid alone or for combined deoxycholic and lithocholic acid experienced significantly lower colonization rates, with none of the patients in the highest combined quartile acquiring vancomycin-resistant Enterococcus during hospital observation.

To explore functional metabolic pathways that might explain these observed bile acid variations, investigators conducted imputed metagenomic profiling using the PICRUSt2 bioinformatics pipeline against the Kyoto Encyclopedia of Genes and Genomes database. The researchers focused computational queries specifically on the bile acid-inducible operon, commonly designated as the bai operon, which encodes the multistep enzymatic pathway responsible for bacterial seven-alpha-dehydroxylation of host bile salts. In unadjusted comparisons between predictive cohort groups, predicted abundance of the baiA gene was significantly higher in patients who successfully avoided colonization compared to those who acquired the pathogen. However, the study authors emphasize that these imputed functional associations did not retain statistical significance after rigorous correction for multiple hypothesis testing. Because functional metagenomic predictions derived from sixteen-S amplicons are inherently indirect, these computational pathway findings serve primarily as exploratory mechanistic hypotheses supporting the empirical mass spectrometry measurements rather than definitive molecular proof.

Clinical evaluations revealed that patients who acquired vancomycin-resistant Enterococcus had significantly longer overall intensive care unit lengths of stay, recording a median hospitalization of fourteen days compared with seven days among persistently negative patients. Although the median interval of six days from admission to first colonization detection was comparable to the total seven-day hospital duration of uncolonized individuals, extended post-acquisition hospitalization likely reflects a combination of greater acute illness severity and sustained hospital exposure. Importantly, recorded exposure to systemic antimicrobial agents during critical care prior to colonization or censoring did not differ significantly between the two cohorts across anti-anaerobic agents, glycopeptides, or carbapenems. Nonetheless, the researchers emphasized that the modest total number of acquisition events precluded the construction of fully adjusted multivariable models capable of simultaneously accounting for time-varying antibiotic exposures and differential hospital time at risk.

The study authors acknowledged multiple methodological constraints that require these observational conclusions to be interpreted cautiously as hypothesis-generating evidence rather than established causal relationships. Because the research was conducted exclusively within a single medical intensive care unit at one academic medical center, regional and institutional variations in patient acuity and infection control practices limit immediate generalizability. Additionally, sixteen-S ribosomal RNA gene sequencing offers narrower taxonomic resolution compared to shotgun metagenomic sequencing, although two complementary classification algorithms confirmed consistent taxonomic assignment for Clostridium scindens. The healthy comparison cohort was also significantly younger than the elderly intensive care unit population, introducing potential age-related confounding in broad case-control contrasts. Ultimately, while these clinical findings align closely with previously published animal models and laboratory demonstrations of bile acid-mediated pathogen inhibition, the researchers concluded that prospective multi-center studies with larger patient populations are essential before these microbial and metabolic markers can be applied as clinical risk-stratification tools.

Subject of Research: Gut microbiome alterations and secondary bile acids associated with vancomycin-resistant Enterococcus acquisition in intensive care units

Article Title: Gut microbiome features associated with vancomycin-resistant Enterococcus acquisition in intensive care unit patients

Article References: Chuang, Y.-C., Collman, R. G., Ruan, S.-Y., Chien, Y.-C., Huang, Y.-S., Hsu, C.-C., Wang, J.-T., Zou, H.-B., & Chang, S.-C. (2026). Gut microbiome features associated with vancomycin-resistant Enterococcus acquisition in intensive care unit patients. Annals of Clinical Microbiology and Antimicrobials, 25(1), Article 29. https://doi.org/10.1186/s12941-026-00871-6

Image Credits: AI Generated

DOI: 10.1186/s12941-026-00871-6

Keywords: Vancomycin-Resistant Enterococcus, Gut Microbiome, Intensive Care Unit, Secondary Bile Acids, Clostridium scindens, Bacterial Colonization Resistance, Deoxycholic Acid, Lithocholic Acid, microbiome, features, associated, vancomycin-resistant