Alterations in the gut microbiome and metabolome in elderly patients with postoperative delirium: A prospective nested case-control study

To elucidate the role of gut microbiota and their metabolites, including short-chain fatty acids (SCFAs) and targeted metabolomics, in the development of postoperative delirium (POD) in elderly patients. Prospective nested case-control study. A Chinese tertiary hospital. Elderly patients underwent elective orthopedic surgery. Participants were assessed for POD using the 3-min Diagnostic Confusion Assessment Method (3D-CAM). Biological samples, including feces and plasma, were collected. A 1:1 propensity score matching (PSM) was conducted to match POD cases with non-POD cases. 16S ribosomal RNA (rRNA) sequencing and metabolomics analyses were performed on the matched case series. Predictive models were developed using logistic regression analysis, incorporating bacterial genera and metabolites that exhibited significant differences between the two groups as predictors. Among 234 patients who were followed up, 41 were diagnosed with POD. A total of 39 cases were matched for both the POD and control groups using PSM. No significant differences were found in the α-diversity and β-diversity of preoperative gut microbiota between the two groups. However, specific bacterial genera, including Romboutsia, Bacteroides faecalis, Blautia mucilaginosa, and Eggerthella lenta, exhibited significant differences. The risk of POD was associated with higher postoperative plasma levels of propionic acid, histidine, aspartate, and ornithine. Logistic regression and receiver operating characteristic curve analyses revealed that indicators derived from the gut microbiota and metabolites could predict POD, with an area under the curve of 0.8413 (95 % confidence interval (CI): 0.7393-0.9434). This study identified four preoperative bacterial genera and four postoperative plasma metabolites associated with an increased risk of POD in elderly orthopedic patients, suggesting the potential of gut microbiota and metabolite profiles as biomarkers for improving risk prediction and guiding interventions.