To characterize the association between electrographic status epilepticus (ESE), as diagnosed by an automated algorithm, and functional outcomes at discharge.

Point-of-care electroencephalogram (POC EEG) helps triage seizure activity, and some systems are capable of running automated algorithms that monitor seizure burden (e.g. Clarity by Ceribell Inc.). Recently, FDA cleared an algorithm for the diagnosis of ESE in POC EEG recordings. Previous work has shown associations between functional outcomes at discharge and automated 5-min seizure burden.

We performed a cohort analysis on the retrospective SAFER-EEG trial.  The original study included hospitalized adult patients who received EEG monitoring across four academic centers.  Here, we included the POC EEG cases from three sites with valid outcome scores.  The ESE diagnosing algorithm was run post-hoc.  The discharge outcomes were categorized as unfavorable if the patient had modified Rankin Scale (mRS) score ≥ 4.  The odds of unfavorable outcome were estimated using a multivariate logistic regression, adjusting for demographics and clinical presentation pre-EEG.

From 359 patients included for analysis, 36 cases (10.0%) had ESE per the algorithm.  Patients with ESE were more likely to have admission diagnoses of acute brain injury (e.g. stroke, traumatic brain injury, inflammatory, etc.), post-cardiac arrest or history of epilepsy than those without a positive finding (75.0% vs. 65.3%, p = 0.004).  After adjusting for covariates, we found that patients with ESE diagnosis had increased odds of unfavorable outcomes at discharge compared to those without ESE (adj. OR = 3.0, 95% C.I. = 1.28-7.69, p = 0.014).  Some other variables that significantly affected the outcomes were age ≥ 65, mRS ≥ 4 pre-EEG, severe head injury, and having over 72h between admission and EEG start.

In a cohort of POC EEG patients, we showed significant associations between presence of ESE as determined by an automated algorithm and poor short-term functional outcomes.