Gut microbiota and SCFA signatures in predicting postoperative infection sites in neurocritical care patients

Background Postoperative pulmonary infection (PI) and intracranial infection (IC) are major complications in neurocritical care patients; however, reliable approaches for predicting infection site susceptibility remain unavailable. Increasing evidence suggests that the gut microbiota and its metabolites influence systemic immune responses through the gut–lung and gut–brain axes. However, whether early postoperative gut microbial signatures could discriminate between subsequent PI and IC remains unknown. Methods In this cross-sectional study, 90 neurocritical care patients undergoing neurosurgical procedures were enrolled and subsequently categorized based on their postoperative infection status into three groups: IC ( n  = 30), PI ( n  = 30), and non-infected controls (CON, n  = 30). Fecal samples were collected within 24 h after surgery for 16S rRNA gene sequencing and gas chromatography–mass spectrometry-based quantification of short-chain fatty acids (SCFAs). Microbial diversity, taxonomic composition, differential microbial biomarkers, functional profiles, and microbiota–metabolite correlations were systematically analyzed. Results Compared with CON patients, both infection groups exhibited significantly reduced microbial diversity and altered community structures. Distinct gut microbial and metabolic patterns were observed between IC and PI patients. The IC group showed enrichment in Enterococcus, Staphylococcus , and Clostridium , increased butyrate levels, reduced propionate levels, and elevated branched-chain SCFA proportions. In contrast, the PI group was characterized by expansion of Proteobacteria , enrichment of Klebsiella and Escherichia-Shigella , increased propionate concentrations, and decreased butyrate levels. Correlation analyses demonstrated strong associations between characteristic microbial genera and SCFA profiles. A six-genus microbial panel, comprising Enterococcus, Faecalibacterium, Dialister, Escherichia-Shigella, Agathobacter , and Blautia , effectively discriminated infection categories, achieving an area under the curve of 0.901. Functional prediction further revealed distinct metabolic and pathogenicity-related pathways among the three groups. Conclusion Early postoperative gut microbiota and SCFA signatures are associated with subsequent infection sites in neurocritical care patients. These findings highlight the potential of gut microbial-metabolic profiling as a non-invasive strategy for early risk stratification and personalized prevention of postoperative pulmonary and intracranial infections.