Detection of electrographic seizures with continuous EEG monitoring in critically ill patients

Objective: To identify patients most likely to have seizures documented on continuous EEG (cEEG) monitoring and patients who require more prolonged cEEG to record the first seizure. Methods: Five hundred seventy consecutive patients who underwent cEEG monitoring over a 6.5-year period were reviewed for the detection of subclinical seizures or evaluation of unexplained decrease in level of consciousness. Baseline demographic, clinical, and EEG findings were recorded and a multivariate logistic regression analysis performed to identify factors associated with 1) any EEG seizure activity and 2) first seizure detected after >24 hours of monitoring. Results: Seizures were detected in 19% (n = 110) of patients who underwent cEEG monitoring; the seizures were exclusively nonconvulsive in 92% (n = 101) of these patients. Among patients with seizures, 89% (n = 98) were in intensive care units at the time of monitoring. Electrographic seizures were associated with coma (odds ratio [OR] 7.7, 95% CI 4.2 to 14.2), age <18 years (OR 6.7, 95% CI 2.8 to 16.2), a history of epilepsy (OR 2.7, 95% CI 1.3 to 5.5), and convulsive seizures during the current illness prior to monitoring (OR 2.4, 95% CI 1.4 to 4.3). Seizures were detected within the first 24 hours of cEEG monitoring in 88% of all patients who would eventually have seizures detected by cEEG. In another 5% (n = 6), the first seizure was recorded on monitoring day 2, and in 7% (n = 8), the first seizure was detected after 48 hours of monitoring. Comatose patients were more likely to have their first seizure recorded after >24 hours of monitoring (20% vs 5% of noncomatose patients; OR 4.5, p = 0.018). Conclusions: CEEG monitoring detected seizure activity in 19% of patients, and the seizures were almost always nonconvulsive. Coma, age <18 years, a history of epilepsy, and convulsive seizures prior to monitoring were risk factors for electrographic seizures. Comatose patients frequently required >24 hours of monitoring to detect the first electrographic seizure.