Background: Mechanisms of postoperative delirium remain poorly understood, limiting development of effective treatments. We tested the hypothesis that intraoperative oxidative damage is associated with delirium and neuronal injury and that disruption of the blood–brain barrier modifies these associations. Methods: In a prespecified cohort study of 400 cardiac surgery patients enrolled in a clinical trial of atorvastatin to reduce kidney injury and delirium, we measured plasma concentrations of F₂-isoprostanes and isofurans using gas chromatography-mass spectrometry to quantify oxidative damage, ubiquitin carboxyl-terminal hydrolase isozyme L1 to quantify neuronal injury, and S100 calcium-binding protein B using enzyme-linked immunosorbent assays to quantify blood–brain barrier disruption before, during, and after surgery. We performed the Confusion Assessment Method for the Intensive Care Unit twice daily to diagnose delirium. We measured the independent associations between intraoperative F₂-isoprostanes and isofurans and delirium (primary outcome) and postoperative ubiquitin carboxyl-terminal hydrolase isozyme L1 (secondary outcome), and we assessed if S100 calcium-binding protein B modified these associations. Results: Delirium occurred in 109 of 400 (27.3%) patients for a median (10th, 90th percentile) of 1.0 (0.5, 3.0) days. In the total cohort, plasma ubiquitin carboxyl-terminal hydrolase isozyme L1 concentration was 6.3 ng/ml (2.7, 14.9) at baseline and 12.4 ng/ml (7.9, 31.2) on postoperative day 1. F₂-isoprostanes and isofurans increased throughout surgery, and the log-transformed sum of intraoperative F₂-isoprostanes and isofurans was independently associated with increased odds of postoperative delirium (odds ratio, 3.70 [95% CI, 1.41 to 9.70]; P = 0.008) and with increased postoperative ubiquitin carboxyl-terminal hydrolase isozyme L1 (ratio of geometric means, 1.42 [1.11 to 1.81]; P = 0.005). The association between increased intraoperative F₂-isoprostanes and isofurans and increased postoperative ubiquitin carboxyl-terminal hydrolase isozyme L1 was amplified in patients with elevated S100 calcium-binding protein B (P = 0.049). Conclusions: Intraoperative oxidative damage was associated with increased postoperative delirium and neuronal injury, and the association between oxidative damage and neuronal injury was stronger among patients with increased blood–brain barrier disruption. Editor’s Perspective: What We Already Know about This Topic: What This Article Tells Us That Is New:

Background: Reduced preoperative cognition is a risk factor for postoperative delirium. The significance for type of preoperative cognitive deficit, however, has yet to be explored and could provide important insights into mechanisms and prediction of delirium. Objective: Our goal was to determine if certain cognitive domains from the general cognitive screener, the Mini-Mental State Exam (MMSE), predict delirium after cardiac surgery. Methods: Patients completed a preoperative MMSE prior to undergoing elective cardiac surgery. Following surgery, delirium was assessed throughout ICU stay using the Confusion Assessment Method for ICU delirium and the Richmond Agitation and Sedation Scale. Results: Cardiac surgery patients who developed delirium ( n = 137) had lower total MMSE scores than patients who did not develop delirium ( n = 457). In particular, orientation to place, working memory, delayed recall, and language domain scores were lower. Of these, only the working memory and delayed recall domains predicted delirium in a regression model adjusting for history of chronic obstructive pulmonary disease, age, sex, and duration of cardiopulmonary bypass. For each word not recalled on the three-word delayed recall assessment, the odds of delirium increased by 50%. For each item missed on the working memory index, the odds of delirium increased by 36%. Of the patients who developed delirium, 47% had a primary impairment in memory, 21% in working memory, and 33% in both domains. The area under the receiver operating characteristics curve using only the working memory and delayed recall domains was 0.75, compared to 0.76 for total MMSE score. Conclusion: Delirium risk is greater for individuals with reduced MMSE scores on the delayed recall and working memory domains. Research should address why patients with memory and executive vulnerabilities are more prone to postoperative delirium than those with other cognitive limitations.

Hypoxemia, as measured by pulse oximetry (SpO2), is common in postanesthesia care unit (PACU) patients. The temporal distribution of desaturation has managerial implications because treatment may necessitate the presence of an anesthesiologist. We retrieved SpO2 values recorded electronically every 30 to 60 seconds from 137,757 PACU patients over n = 80 four-week periods at an academic medical center. Batch mean methods of analysis were used. Onset times of hypoxemic episodes (defined, on the basis of previous studies, as SpO2 <90% lasting at least 2 minutes) were determined and resolution at 3, 5, and 10 minutes was assessed. Episodes beginning <30 minutes and ≥30 minutes after PACU admission were compared. Patients undergoing intubation in the PACU were identified by doing a free text search of electronically recorded nursing notes for phrases suggesting intubation, followed by a confirmatory manual chart review. Intervals from PACU admission to intubation were determined. Fewer than half (31.2% ± 0.05%) of episodes of PACU hypoxemia lasting ≥2 minutes occurred 50%) occurred ≥30 minutes after admission (P < 0.0001). Few (<1%) anesthesia providers transporting patients to the PACU were still present in the PACU 30 minutes after arrival in the PACU. Fewer than half (37%; 95% confidence interval, 27.4% to 48.8%) of PACU intubations occurred 50%) occurred ≥30 minutes after admission (P = 0.029). Hypoxemic episodes in the PACU resolved more slowly than episodes in operating rooms (P < 0.0001). After 3 minutes, 40.9% ± 0.6% were unresolved in the PACU versus 23% (99% upper confidence limit) in operating rooms, and 32.6% ± 0.5% vs 9% (99% upper confidence limit) after 5 minutes. Because most (68.8%) hypoxemic episodes in the PACU occur ≥30 minutes after admission, a time by which the anesthesia provider who transported the patient usually would no longer be present (>99% of cases), the PACU needs to be considered when anesthesiologist operating room staffing and assignment decisions are made.