Continuous EEG predictors of outcome in patients with altered sensorium

PURPOSE To evaluate EEG predictors of outcome in patients with altered sensorium suspected to have seizure and to assess whether short term EEG is as effective as long term continuous EEG (cEEG) in predicting the outcome of patients with altered sensorium due to neurological causes. METHODS We identified 99 consecutive critically ill patients with altered sensorium in whom nonconvulsive seizures were suspected. They underwent cEEG. Functional outcome was assessed with the modified Rankin Scale (mRS) at 4 weeks, discharge or death. We compared efficacy of short term EEG and long term continuous EEG in recording the abnormal patterns on EEG. Logistic regression analysis was performed to identify EEG findings associated with poor outcome, defined as mRS 4-6 (dead or moderately to severely disabled). RESULTS Poor outcome was associated with nonconvulsive seizures (NCS), nonconvulsive status epilepticus (NCSE), periodic lateralized epileptiform discharges (PLEDS), PLEDS plus, generalized periodic epileptiform discharges (GPEDS) and abnormal EEG background. Short term EEG can detect seizure activity and other electrographic markers of poor outcome but the values are statistically insignificant. CONCLUSION cEEG monitoring provides independent prognostic information in patients with altered sensorium and suspected seizures. Unfavorable findings include nonconvulsive seizures, periodic epileptiform discharges and abnormal background. Short term EEG is ineffective in detecting seizures on EEG in patients with altered sensorium and should not be used as substitute for Long term EEG.