Background: Non-convulsive seizures have been reported to be common in neurocritical care patients. Many jurisdictions do not have sufficient resources to enable routine continuous electroencephalography (cEEG) and instead use primarily intermittent EEG, for which the diagnostic yield remains uncertain. Determining risk factors for epileptiform activity and seizures could help identify patients who might particularly benefit from EEG monitoring.
Methods: We performed a cohort study involving neurocritical care patients with admission Glascow Coma Scale (GCS) scores ≤ 12, who underwent ≥ 1 EEG. EEGs were reviewed for presence of interictal discharges, periodic epileptiform discharges (PEDs), and seizures. Multivariate analysis was used to identify predictors of these findings and to describe their prognostic implications.
Results: 393 patients met inclusion criteria. 34 underwent cEEG, usually because epileptiform activity was first detected on a routine EEG. The prevalence of PEDs or electrographic seizures was 13%, and was highest with anoxic encephalopathy and central nervous system infections. Other independent predictors for epileptiform activity included a history of convulsive seizure(s), increasing age, deeper coma, and female gender. Although patients with epileptiform activity had higher mortality, this association disappeared after adjustment for confounders.
Conclusion: Approximately 7-8 neurocritical care patients must undergo intermittent EEG monitoring in order to diagnose one with PEDs or seizures. The predictors we identified could potentially help guide use of resources. Repeated intermittent studies, or cEEG, should be considered in patients with multiple risk factors, or when interictal discharges are identified on an initial EEG. It remains unclear whether aggressive prevention and treatment of electrographic seizures improves neurologic outcomes.