The potential of processed EEG monitoring systems in predicting postoperative cognitive deficits

Postoperative neurocognitive disorders (PNDs), particularly postoperative delirium (POD), are significant yet frequently overlooked complications following surgery. PND is marked by cognitive impairment, confusion, and attention deficits. It is especially prevalent among elderly patients, leading to worsened postoperative outcomes and increased mortality. Research into risk factors has emerged. This study aims to evaluate the association between preoperatively recorded processed electroencephalogram (pEEG) indices and POD.

One hundred and ten patients who underwent surgical intervention under general anesthesia were analyzed. Inclusion required artifact-free preoperative EEG recordings and cognitive assessments using the brief Confusion Assessment Method (bCAM) post-surgery. We included pEEG indices from three different systems, i.e., the Bispectral Index (BIS) and the spectral edge frequency (SEF) from the BIS monitor, the SEDLine Patient State Index (PSI), and the Entropy Module's State Entropy (SE). We used the Mann-Whitney U test for pairwise comparisons, and the discriminatory ability was evaluated via the effect size area under the curve.

Thirty patients had to be excluded due to corrupted baseline data. Out of the remaining 80 patients with a median age of 71 (1st quartile 67, 3rd quartile 79) years, 24 were screened positive for POD. Analyses revealed significant differences in pEEG values between patients with and without POD, particularly in the eyes-closed EC condition. All monitor indices except SEF showed significantly lower EC baseline indices in patients with PND, but the effect sizes were rather small with AUC ranging from 0.62 to 0.65. Effects were similar for the indices at eyes-open EO and their change from EC to EO. The strongest effect size was found for the PSI at EO with AUC=0.71 [0.58;0.83] (p<0.01).

This study highlights the potential of preoperative pEEG monitoring to identify patients at risk for POD in an elderly surgical population. Its integration into preoperative assessments may improve patient care and reduce PND risks. Future work is essential to validate these findings across larger, diverse populations and to elucidate the mechanisms connecting pEEG patterns with postoperative cognitive outcomes.