Diagnostic yield of electroencephalography in the medical and surgical intensive care unit

To determine the incidence of electrographic seizures during continuous electroencephalography (cEEG) in the medical and surgical ICU.

We retrospectively reviewed the records of all adults who underwent cEEG in our medical and surgical ICU over a 4.5 year period. Patients with acute brain injury were excluded. Our primary outcome was cEEG documentation of an electrographic seizure, defined as a rhythmic discharge or spike and wave pattern demonstrating definite evolution and lasting at least 10 s. To assess inter-rater variability in cEEG interpretation, two electrophysiologists independently reviewed all available cEEGs of subjects with electrographic seizures documented on their clinical cEEG report and those of an equal number of randomly selected subjects from the remaining cohort.

Kappa analysis showed a value of 0.88, indicating excellent inter-rater agreement. Electrographic seizures were identified in 12 of 105 patients (11 %, 95 % CI 5-18 %). This rate did not change after excluding patients with a history of seizure, remote brain injury, or seizure-like events before cEEG. In an ordinal logistic regression model controlling for age, sex, and SOFA score, electrographic seizures were associated with lower odds of good outcomes on the Glasgow Outcome Scale at discharge (OR 0.3, 95 % CI 0.1-0.8).

In a tertiary care medical and surgical ICU, electrographic seizures were seen on 11 % of cEEGs ordered for the evaluation of encephalopathy, and were associated with worse functional outcomes at discharge. Our findings confirm the results of a prior study suggesting a substantial burden of electrographic seizures in critically ill encephalopathic patients.