Gut microbiota dysbiosis in patients with sepsis: a meta-analysis of microbial diversity and taxonomic abundance.

BACKGROUND

Gut microbiota alterations are frequently reported in sepsis, but their magnitude, consistency, and clinical specificity remain uncertain. We evaluated microbial diversity and taxonomic abundance in patients with sepsis, distinguishing healthy/reference controls from nonseptic clinical comparators.

METHODS

PubMed, Embase, Cochrane Library, and Web of Science Core Collection were searched from inception to 20 September 2026, without language or geographic restrictions, supplemented by targeted retrieval of primary studies and supplementary data. Comparative human observational studies were included. Standardized mean differences (Hedges g) and 95% confidence intervals (CIs) were pooled using random-effects models. Healthy/reference and nonseptic clinical controls were analyzed separately. Subgroup and sensitivity analyses were performed to explore heterogeneity and assess the stability of pooled estimates. Small-study effect tests and meta-regression were restricted to outcomes with sufficient studies.

RESULTS

Twenty-three studies comprising 1201 participants were included. Compared with healthy/reference controls, patients with sepsis had lower Shannon diversity (20 studies; g = - 1.12, 95% CI - 1.38 to - 0.85; I² = 59.9%) and lower Firmicutes abundance (g = - 0.78, 95% CI - 1.32 to - 0.24), but higher Proteobacteria (g = 0.96, 95% CI 0.67 to 1.25) and Enterococcus abundance (g = 0.68, 95% CI 0.38 to 0.98). These associations remained statistically significant after correction for multiple comparisons. Estimates for the remaining outcomes were less consistent or imprecise. The reduction in Shannon diversity was smaller against nonseptic clinical controls (eight studies; g = - 0.29, 95% CI - 0.54 to - 0.05) and was not statistically significant after multiplicity correction.

CONCLUSIONS

Sepsis was associated with reduced gut microbial diversity and alterations in selected bacterial taxa, most consistently in comparisons with healthy/reference controls. Clinical heterogeneity and incomplete control of antibiotic exposure and illness severity limit attribution of these changes to sepsis itself. Prospective studies are needed to establish their prognostic and therapeutic relevance.