Autonomic Vulnerability Phenotype, Regional Cerebral Oxygen Saturation, and Postoperative Delirium in Elderly Patients Undergoing Non-Cardiac, Non-Neurological Surgery: A Propensity Score–Matched Cohort Study

Background and Objectives: We investigate whether a preoperative autonomic vulnerability phenotype and intraoperative regional cerebral oxygen saturation (rcSO2) variables are associated with postoperative delirium (POD) in elderly patients undergoing non-cardiac, non-neurological surgery, and whether autonomic vulnerability modifies the association between cerebral desaturation and POD. Materials and Methods: This retrospective propensity score–matched cohort study included patients aged 65 years or older who underwent general anesthesia with intraoperative rcSO2 monitoring. The preoperative autonomic vulnerability phenotype was defined using clinical features documented before surgery, including autonomic neuropathy, diabetic autonomic neuropathy, orthostatic hypotension, syncope or presyncope suggestive of autonomic dysfunction, and unexplained resting bradycardia or chronotropic incompetence not attributable to rate-limiting medication. The primary outcome was POD within 5 postoperative days. Patients were matched 1:1 using nearest-neighbor propensity score matching with a caliper of 0.2 standard deviations of the logit of the propensity score, and conditional logistic regression was used in the matched cohort. Results: A total of 412 patients were included; 112 had the phenotype and 300 did not. After matching, 98 pairs were analyzed. POD occurred in 27.6% of patients with the phenotype and 14.3% of patients without it. In the matched cohort, the phenotype (odds ratio [OR] 2.12, 95% confidence interval [CI] 1.18–3.82, p = 0.012), rcSO2 decrease ≥20% (OR 2.45, 95% CI 1.31–4.58, p = 0.005), and longer duration of rcSO2 < 80% of baseline (OR 1.02 per min, 95% CI 1.01–1.04, p = 0.008) were independently associated with POD. The phenotype-by-desaturation interaction was exploratory (OR 2.10, p = 0.032) and was not uniformly robust across sensitivity analyses. Conclusions: A preoperative autonomic vulnerability phenotype and intraoperative cerebral desaturation were independently associated with POD. The association between rcSO2 decrease and POD appeared stronger in patients with autonomic vulnerability, but this interaction should be interpreted as hypothesis-generating rather than confirmatory.