Recent work has shown that dysfunctional brain EEG responses to anesthetic drugs can be an indicator of both preoperative cognitive impairment and postoperative delirium risk. However, since excessive anesthetic dosage can also cause abnormal EEG brain responses, it is unclear how to tell to what extent such abnormal brain EEG responses reflect latent neurocognitive impairment versus excessive anesthetic dosage. Further, it is unclear what underlying mechanisms might underlie the link between phenotypes (such as delirium and cognitive impairment) and these abnormal neurophysiologic responses to anesthetic drugs.

Dual center prospective cohort design. 139 total older surgical patients from two academic centers underwent intraoperative EEG monitoring with the bispectral index (BIS) EEG monitor during anesthesia and surgery, and postoperative delirium screening by geriatrician interview (Duke cohort) or by trained research staff (Mt Sinai cohort). We developed the Duke Anesthesia Resistance Scale (DARS), defined as the average BIS EEG values divided by the quantity 2.5 minus the age adjusted end tidal anesthetic gas concentration). We then examined the relationship between the DARS and postoperative delirium risk using the Youden index to identify an optimal low DARS threshold for delirium risk, and we used multivariable logistic regression to control for potential confounders.

Neither BIS scores nor inhaled anesthetic dosage differed significantly between patients with vs without postoperative delirium. Yet, patients with delirium had lower DARS scores than those who did not develop delirium (27.92 vs 32.88, p=0.015). A DARS threshold of 28.7 maximized the Youden index for the association between the DARS and delirium. In multivariable models adjusting for site (Duke vs Mt Sinai) and individual patient risk factors, DARS values <28.7 were associated with a 3.79 fold increased odds ratio (95% CI 1.63-9.10; p=0.03) for postoperative delirium. These results remained unchanged after adjusting for intraoperative medications including opioids, benzodiazepines, propofol, phenylephrine and ketamine. Patients with structural/functional MRI or CSF biomarker evidence of preclinical/prodromal Alzheimer's disease and/or neurovascular pathology were more likely to show altered anesthetic-induced EEG activity patterns.

Lower scores on a processed EEG-based scale of neurophysiologic resistance to anesthetic induced brain activity changes were independently associated with a nearly 4 fold increased delirium risk. The altered anesthetic-induced brain EEG patterns in patients who go on to develop postoperative delirium may reflect latent pre-clinical/pro-dromal Alzheimer's disease and/or neurovascular pathology.