Electrical & Computer Engineering, Department of


Date of this Version







Postoperative delirium is the most common complication among older adults undergoing major surgery. The pathophysiology of delirium is poorly understood, and no blood-based, predictive markers are available. We characterized the plasma metabolome of 52 delirium cases and 52 matched controls from the Successful Aging after Elective Surgery (SAGES) cohort (N= 560) of patients ≥ 70 years old without dementia undergoing scheduled major non-cardiac surgery. We applied targeted mass spectrometry with internal standards and pooled controls using a nested matched case-control study preoperatively (PREOP) and on postoperative day 2 (POD2) to identify potential delirium risk and disease markers. Univariate analyses identifed 37 PREOP and 53 POD2 metabolites associated with delirium and multivariate analyses achieved signifcant separation between the two groups with an 11-metabolite prediction model at PREOP (AUC= 83.80%). Systems biology analysis using the metabolites with diferential concentrations rendered “valine, leucine, and isoleucine biosynthesis” at PREOP and “citrate cycle” at POD2 as the most signifcantly enriched pathways (false discovery rate < 0.05). Perturbations in energy metabolism and amino acid synthesis pathways may be associated with postoperative delirium and suggest potential mechanisms for delirium pathogenesis. Our results could lead to the development of a metabolomic delirium predictor.