Abstracts

Rationale: Nonconvulsive seizures (NCS) are common in patients with acute brain injury. Most NCS are unrecognized at bedside and usually require EEG for diagnosis. Although prolonged EEG monitoring is essential in the diagnosis of NCS, it is time consuming, labor intensive, expensive and technically challenging. In addition, it is still unclear whether treating subclinical seizures will improve the outcome. The aim of this study was to determine the clinical utility and prognostic importance of prolonged EEG monitoring in identifying NCS in the intensive care unit. Methods: We reviewed 831 bedside routine and prolonged EEG recordings performed between January 1, 2000 and September 30, 2007. The indications for these EEGs were generalized convulsive status epilepticus and persisting encephalopathy with or without seizures. For this study, only patients who had prolonged EEG monitoring for at least 72 hours were evaluated. Patients who required prolonged EEG monitoring for titration of antiepileptic medications or pentobarbital were excluded. We used multivariate logistic regression to identify factors that are associated with NCS documented on prolonged EEG monitoring. Multinomial logistic regression was used to determine the association between the variables and the outcome. Results: One hundred and thirty-eight patients were identified. The mean age was 56 years. Eighty-nine patients (64%) did not have NCS on routine EEGs or prolonged EEG monitoring. NCS were detected by routine EEG recordings in 42 patients (30%). Only seven patients (5%) had unremarkable routine EEGs but developed NCS during prolonged EEG monitoring. Six of these patients had NCS within the first 12 hours of recording. The other patient developed NCS within the first 24 hours. Among patients with NCS, the history of epilepsy was associated with electrographic seizures (OR 7.4, 95% CI 2.2 to 25.0). Level of consciousness and age were poor predictors of NCS on prolonged EEG monitoring (P>0.05). The outcome was identified in 115 patients. Among these, there were death in 62 patients, vegetative state in 4 patients, severe disability in 21 patients, moderate disability in 17 patients and mild disability in 11 patients. We found that only coma was associated with death (OR 11.2, 95% CI 3.7-34.2). The presence of NCS, number of antiepileptic medications administered, age, underlying medical conditions and duration of illness were not of prognostic importance (P>0.05). Conclusions: At least 12 hours of prolonged EEG monitoring is essential in the diagnosis of NCS in the intensive care unit. Patients with a history of epilepsy have a higher risk of developing NCS. Although it seems reasonable to detect subclinical seizures, treatment of NCS in comatose patients does not appear to improve clinical outcome.