Abstracts

Rationale: During critical illness, the threshold for clinical and subclinical seizures is lowered. Several studies have shown that subclinical seizures and status epilepticus are common in the ICU (10-20% of patients), and not detected with a standard 30-minute EEG. It is unclear whether ICU patients should receive continuous EEG monitoring, and whether this has an impact on clinical decision-making and outcome. We addressed these questions by doing 24-hour video-EEG monitoring on ICU patients who would normally have received only a 30-minute study. Methods: During a prospective 2-year study, ICU patients, for whom medical staff requested a standard 30-minute EEG at any point during their ICU stay, instead got 16-24 hours of continuous video-EEG. Patients who initially needed long-term EEG monitoring were excluded from this study. Some patients were excluded for logistical reasons and received only a 30-minute EEG. ICU nurses were asked to press the alarm button for clinical seizures, and all video-EEG data were reviewed by an epileptologist. Abnormalities were noted and categorized. Hospital charts were reviewed for treatment decisions and outcomes. Results: Altogether, 175 ICU patients got EEGs. Of these, 67 from the outset needed and received continuous video-EEG monitoring. 31, for logistical reasons, had only a 30- minute routine EEG. The remaining 77 patients, who were in the study, had continuous video-EEG for 16-24 hours. In a few, this was continued for longer. All EEGs, except one, were abnormal. Non-epileptic abnormalities included generalized slowing (76), burst-suppression (4), triphasic waves (1). 17 patients showed epileptiform EEG activity. 10 had structural pathology such as hemorrhage or tumor, 5 had suffered cardiac arrest, 2 had metabolic derangements. The most common finding was bilateral or lateralized periodic epileptiform discharges (BiPEDs or PLEDs) in 11 patients. Other abnormalities were myoclonus with generalized EEG discharges (2), subclinical generalized seizures (1), focal motor seizures (3). 9 patients showed epileptiform activity in the first 30 minutes, of whom 1 had a clinical seizure. In 8 patients the epileptiform activity evolved overnight, with 5 patients showing definite seizure activity (1 electrographic, 4 clinical). No patient had non-convulsive status epilepticus. 39 patients died. 38 went to a residential facility. 6 patients with seizures were treated, and 4 were discharged. Conclusions: This study suggests that in an unselected ICU population, overnight EEG monitoring, compared to a standard 30-minute EEG, adds only a little information, and has little effect on treatment or outcome. Additional epileptiform abnormalities were detected in only 8 out of 77 patients, and only 1 of those had clinically undetectable seizures. 5 out of 6 patients with seizures were treated on clinical grounds. The benefit of treating purely electrographic abnormalities is unproven. Although prolonged monitoring for >24 hours might lead to a different conclusion, we believe that routine long-term EEG monitoring of all ICU patients is not warranted.