Among 491 surgical patients aged ≥65 undergoing elective laminectomy or joint replacement, 10% developed postoperative delirium within 24–48 hours. In the subset with complete biomarker and microbiome data (n=86–139), co-occurrence network analysis revealed distinctly different microbial interaction patterns between delirium and non-delirium groups. Critically, significant associations among specific gut microbiota, the tryptophan-derived metabolite indole-3-propionic acid (IPA), and plasma Tau phosphorylation were detectable exclusively in delirium patients — not controls. Machine learning classifiers using microbiome-derived features achieved AUC values above 50%, a modest but directionally meaningful signal.
Postoperative delirium afflicts up to 50% of older surgical patients in some populations and is independently linked to accelerated cognitive decline and dementia. This work is notable for triangulating three previously siloed research threads — neuroinflammation (Tau), gut dysbiosis, and tryptophan metabolism — into a coherent gut-blood-brain axis hypothesis. IPA is a known neuroprotective metabolite produced by Clostridium sporogenes; its association with Tau pathology here opens a mechanistically plausible therapeutic avenue. However, the analytical sample is small (86 patients with microbiome data), the design is observational, and AUC >50% is a low bar for a diagnostic claim. Causality remains unestablished. This is best characterized as a hypothesis-generating, signal-finding study — incremental but scientifically coherent — that warrants validation in larger, interventional cohorts before any clinical translation.
