Continuous electroencephalography in a mixed non-neurological intensive care population, an observational study

Abstract

Purpose: Continuous electroencephalography (cEEG) improves monitoring of the brain in unconscious patients, but implementation at ICU is difficult. The present investigation shows a way to introduce cEEG at an anesthesiological ICU and discusses the first experiences.

Materials and methods: The study analyzed the feasibility of cEEG, assessed the interpretable cEEG time, importance of automatic seizure detection, the incidence of seizures, the predominant background EEG activity, incidence of delirium and mortality.

Results: Fifty-three cEEGs of 50 patients with a median interpretable length of 24 hours [IQR 20 to 42 hours] were recorded. One patient had status epilepticus, while 5 patients had non-convulsive seizures. Automated seizure detection recognized the status epilepticus and 3 of 10 non-convulsive seizures, however, detected 42 false positive seizures. Predominant background EEG activity was alpha (9%), theta (17%), delta (26%), burst-suppression (17%), and suppressed background activity (30%). EEG activity correlated neither with dosage of analgo-sedative drugs nor with incidence of delirium or mortality.

Conclusion: Continuous electroencephalography recording is feasible and manageable. Automatic seizure detection was often false negative/positive; therefore, the interpretation of the cEEG should be supported by EEG-trained neurologists. Background EEG activity was not associated with outcome parameters, which suggests that background activity is a poor outcome predictor.

Keywords: Coma; Critical care; Electroencephalography; Non-convulsive seizures; Seizure detection.