Abstracts

Rationale: Critically ill neurologic patients often require prolonged EEG monitoring to monitor depth of sedation. This is important in patients who are receiving anesthetic drugs for the goal of achieving burst suppression to manage intracranial pressure and prevent awareness during pharmacological paralysis. Although bispectral EEG analysis (BIS) has been studied for years as a monitoring device in the operating room, its utility in patients with neurologic injury in the ICU remains unclear. Compared to traditional EEG, BIS is a small device with a simplified numeric output that can be read at the bedside and could represent a cost effective, labor saving alternative to continuous EEG. Therefore, the aim of this study was to investigate the correlation between BIS and scalp EEG recording in patients with neurologic injury in order to assess the potential for BIS as a substitute to full montage EEG monitoring in select NICU patients.Methods: 8 patients in the NeuroICU underwent simultaneous EEG and BIS monitoring. Information on the clinical diagnosis, EEG reports and radiographic findings was retrospectively collected. Persyst software was used to calculate the suppression ratio (SR), from the left and right hemisphere at 10 second intervals. BIS indices were recorded in approximately 60 second intervals. BIS and SR data points from each case were first inspected visually with scatter plots. Pearson correlation was then performed on all ranges. For selected cases a second correlation with values of SR over 40% based on previous BIS studies showing more validity in these ranges.Results: 6 of the 8 patients were female. The primary diagnosis was subarachnoid hemorrhage (SAH) for 6/8 patients and cerebral edema with intraparenchymal hemorrhage for the other 2 patients. All were sedated with midazolam or propofol and were paralyzed with cisatracurium. The median BIS indices between patients ranged from 13-49 with a mean of 34. The median SR between patients ranged from 7.8 to 66% with a mean of 31%. Of the 8 patients, 5 had strong inverse correlation (r = -0.582 to -0.938) of BIS to SR and 1 had a moderate inverse correlation (-0.265 on the left and -0.313 on right) that was statistically significant (p > 0.001). For two of these cases the correlation was only present when the suppression ratio is greater than 40%. Of the two patients in whom the BIS did not correlate with the SR, one had significant EEG asymmetry with alpha rhythm on the same side (left) as the unilateral BIS recording.Conclusions: In this case series there is evidence of inverse linear correlation between BIS indices and suppression ratio from scalp EEG. For some cases the correlation is best when the SR is greater than 40%. However, this is a small sample and further study is needed to verify these findings over multiple levels of sedation and to investigate possible causes of lack of correlation in BIS monitoring before BIS monitoring can be considered a substitute to full montage EEG for monitoring depth of sedation in NICU patients.