BackgroundPostoperative delirium (PD) and subsyndromal delirium (PSSD) are frequent complications in older patients associated with poor long-term outcome. It has been suggested that certain electroencephalogram features may be capable of identifying patients at risk during surgery. Thus, the goal of this study was to characterize intraoperative electroencephalographic markers to identify patients prone to develop PD or PSSD.MethodsWe conducted an exploratory observational study in older patients scheduled for elective major abdominal surgery. Intraoperative 16 channels electroencephalogram was recorded, and PD/PSSD were diagnosed after surgery with the confusion assessment method (CAM). The total power spectra and relative power of alpha band were calculated.ResultsPD was diagnosed in 2 patients (6.7%), and 11 patients (36.7%) developed PSSD. All of them (13 patients, PD/PSSD group) were compared with patients without any alterations in CAM (17 patients, control group). There were no detectable power spectrum differences before anesthesia between both groups of patients. However, PD/PSSD group in comparison with control group had a lower intraoperative absolute alpha power during anesthesia (4.4 ± 3.8 dB vs. 9.6 ± 3.2 dB, p = 0.0004) and a lower relative alpha power (0.09 ± 0.06 vs. 0.21 ± 0.08, p < 0.0001). These differences were independent of the anesthetic dose. Finally, relative alpha power had a good ability to identify patients with CAM alterations in the ROC analysis (area under the curve 0.90 (CI 0.78-1), p < 0.001).DiscussionIn conclusion, a low intraoperative alpha power is a novel electroencephalogram marker to identify patients who will develop alterations in CAM – i.e., with PD or PSSD – after surgery.

BACKGROUND: Patients with low cognitive performance are thought to have a higher risk of postoperative neurocognitive disorders. Here we analyzed the relationship between preoperative cognition and anesthesia-induced brain dynamics. We hypothesized that patients with low cognitive performance would be more sensitive to anesthetics and would show differences in electroencephalogram (EEG) activity consistent with a brain anesthesia overdose. METHODS: This is a retrospective analysis from a previously reported observational study. We evaluated cognitive performance using the Montreal cognitive assessment (MoCA) test. All patients received general anesthesia maintained with sevoflurane or desflurane during elective major abdominal surgery. We analyzed the EEG using spectral, coherence, and phase-amplitude modulation analyses. RESULTS: Patients were separated into a low MoCA group (<26 points, n = 12) and a high MoCA group (n = 23). There were no differences in baseline EEG, nor end-tidal age-corrected minimum alveolar concentration (MACage). However, under anesthesia, the low MoCA group had lower α-β power (high MoCA: 2.9 [interquartile range {IQR}: 0.6–5.8 dB] versus low MoCA: −1.2 [IQR: −2.1 to 0.6 dB], difference 4.1 [1.0–5.7]) and a lower α peak frequency (high MoCA: 9.0 [IQR: 8.3–9.8 Hz] versus low MoCA: 7.5 [IQR: 6.3–9.0 Hz], difference 1.5 [0–2.3]) compared to the high MoCA group. The low MoCA group also had a lower α band coherence and a stronger peak-max phase–amplitude coupling (PAC). Finally, patients in the low MoCA group had longer emergence times (high MoCA 663 ± 345 seconds versus low MoCA: 960 ± 352 seconds, difference 297 [15–578]). Multiple linear regression shows up that both age and MoCA scores are independently associated with intraoperative α-β power. CONCLUSIONS: All these EEG features, together with a prolonged emergence time, are consistent with the possibility that older patients with low cognitive performance are receiving a brain anesthesia overdose compare to cognitive normal patients.