The role of the standard 20 minute EEG recording in the comatose patient

Abstract

Non-convulsive seizures and non-convulsive status epilepticus (NCSE) are believed common in comatose patients and are suggested to worsen outcome. The purpose of this study was to prospectively evaluate outcome in patients in critical care units in whom NCSE was suspected to determine how often evidence of seizure activity existed based on an isolated standard 20 minute electroencephalogram (EEG) and to determine what clinical factors predicted outcome. We prospectively reviewed EEGs and clinical charts of patients admitted to a critical care unit at a tertiary care center who were suspected to have non-convulsive seizures. Outcomes were correlated with EEG findings, clinical factors, and acute therapies using univariate and multivariate logistic analyses. Of 189 patients, complete information was available in 169. Eighty-one (47.9%) patients died, 67 (39.6%) were discharged home, and 21 (12.4%) were discharged to long-term care. Four patients had electroencephalographic seizures, two of whom had no clinical manifestations (i.e. non-convulsive). On univariate analysis, increased age, an admitting diagnosis of cardiac arrest, a Glasgow Coma Scale (GCS) score ⩽ 8, and burst suppression were correlated significantly with poor outcome. A past history of seizures and unequivocal tonic–clonic convulsions were correlated significantly with a better outcome. On multivariate analysis, increased age, cardiac arrest, and a GCS score ⩽ 8 were associated with increased mortality ( p < 0.05). Clinical factors, including age, underlying etiology and GCS score are the most important predicators of outcome in coma. A standard 20 minute EEG did not correlate with a high detection rate of seizure activity. Furthermore, EEG patterns and treatment with anticonvulsant medications did not correlate with outcome.