Abstracts

Rationale: To investigate the utility of prolonged EEG monitoring in the ICU and the value of specific EEG patterns in predicting prognosis in patients in coma/stupor (CS+) or with altered mental status (AMS). Methods: We retrospectively analyzed (2005-2011) continuous EEG monitoring reports of 40 adults patients with CS+ (22) or AMS (18) in the ICU of the University of Minnesota. The EEG reports(first and last)of patients in a coma/stupor (CS+) condition (n=22) were compared to patients with AMS. Table 1 summarizes clinical data of the two groups. The differences in modified Rankin Scale score (MRS) at the time of discharge, the EEG patterns seen, i.e periodic (GPED, PLEDS,Bi-PlEDS), slow (theta,FIRDA,deltas) or mixed between the two groups were analyzed using Fisher's exact probability and t test. Results: EEG monitoring was done mean 3.0 +/-1.61 on CS+ and 7.0 +/-9.73 on AMS group (p.09)No difference in ICU stay days between groups (p=.26) EEG seizures were found, in the first EEG, in 25 out of the 40 (62.5%) patients. Group 1 (CS+)had 15 and group 2 (AMS) had 10. Fourteen (14) patients(35%) were actually in non-convulsive status (NCS). Etiology of both groups was mixed (Table 1). There was a significant difference in the MRS between the two groups (P < 0001) with those in the CS + (group 1) having a higher disability score at discharge 34 (85%) or death 6(15%) than Group 2 (AMS). There was no correlation between the MRS and different EEG patterns described(P = 0.73)(Table 2). Conclusions: EEG in the ICU in CS+ and AMS patients was very useful for detection of seizures activity and NCS. High MRS correlated directly with presence of CS + at time of monitoring. No correlation between MRS and specific EEG patterns were found (P = 0.73). A larger study with more subjects and more specific EEG patterns is recommended.