Gut Pathog. 2026 May 30. doi: 10.1186/s13099-026-00838-0. Online ahead of print.
ABSTRACT
BACKGROUND: Gut microbial dysbiosis has been implicated in sepsis-related organ dysfunction. However, the longitudinal dynamics of the gut microbiota and mycobiota-and particularly their cross-kingdom ecological organization-in pneumonia-derived sepsis remain incompletely understood.
METHODS: Patients with pneumonia-derived sepsis were prospectively enrolled. Fecal samples and clinical data (SOFA scores and inflammatory markers) were collected on Day 1 and Day 7. Gut bacterial and fungal communities were profiled using 16 S rRNA and ITS1 sequencing. Longitudinal and outcome-stratified analyses were performed. Trans-kingdom co-occurrence networks and module-based topological analyses were constructed, and associations with clinical parameters were explored.
RESULTS: Global analyses indicated relative compositional stability in the gut microbiota and mycobiota between Day 1 and Day 7, with no significant differences in alpha or beta diversity. The dominant bacteria were Bacillota, Bacteroidota, and Pseudomonadota at the phylum level, and Enterococcus, Bacteroides, and Escherichia-Shigella at the genus level; Escherichia-Shigella showed a decreasing trend and Bacteroides an increasing trend, though neither reached statistical significance (Padj > 0.05). Ascomycota dominated the fungal community, with Candida, Fusarium, and Oligophagozyma as the core genera, with no obvious temporal shifts. However, outcome-based stratification revealed that fungal Chao1 richness increased significantly post-treatment, specifically in the bad-outcome group (P < 0.05). The most notable findings emerged from the trans-kingdom interactome. In the favorable-outcome group, a specific modular configuration (ModM1) was identified post-treatment, containing four microbiota hubs (Parabacteroides, Mediterraneibacter, Serratia, and Enterococcus). While the aggregate abundance of ModM1 lacked clinical correlation, its hub genus, Mediterraneibacter-a prevalent anaerobe-showed a negative association with PCT and TNF-α. Additionally, the fungal-integrated ModM8 showed a potential positive association with IL-8. Conversely, the bad-outcome group showed a lack of such hub-anchored coordination.
CONCLUSIONS: In this small exploratory cohort, early pneumonia-derived sepsis appeared to exhibit relative taxonomic stability but subtle reorganization of cross-kingdom ecological connectivity. Microbial shifts appeared to manifest primarily as changes in network embedding rather than abundance. These observations provide exploratory insights that require further validation regarding topological integration, especially fungal involvement in inflammatory modules, for understanding host-microbiome interactions in critical illness. Larger longitudinal studies are warranted.
TRIAL REGISTRATION: ClinicalTrials.gov, NCT04525677, registered on 14 July 2020.
PMID:42218556 | DOI:10.1186/s13099-026-00838-0
