Automated Seizure Detection and Management using CERIBELL: A single center cohort study
Ajay Rajshekar1, James Siegler1, Jared Wolfe1, Miranda Flamholz1, Kenyon Sprankle1, Manisha Koneru1, Stefan Gillen1
1Cooper University Hospital
Objective:
To determine whether point-of-care electroencephalography (POC-EEG) monitoring, using the Ceribell device can reduce time from hospital arrival to EEG and successful treatment of seizure activity when compared to conventional 10-20 EEG.
Background:
Ceribell is a point-of-care electroencephalography (POC-EEG) device that includes machine-learning algorithms to detect status epilepticus without the requirement of a trained technologist. We compared time to monitoring and outcomes with POC-EEG against conventional EEG.
Design/Methods:
A process improvement paradigm was implemented in which POC-EEG was used for suspected seizure activity when 10-20 EEG monitoring was not immediately available (off-hours 6pm-8am, June 2021 - June 2022). Data regarding diagnostic testing and seizure treatment were compared to patients monitored in the preceding period using the conventional paradigm (Mar 2020 - May 2021).
Results:
Of the 142 patients with suspected seizure activity for whom an EEG was requested during off-hours, 85 (60%) were monitored using POC-EEG. Compared to those monitored using a conventional 10-20 system, those monitored with POC-EEG had no significant clinical differences between groups (age, race, ethnicity, history of seizure, indication for EEG). Among patients monitored within 24h of hospital arrival for presumed seizure (n=97), POC-EEG was associated with shorter time to monitoring (median 225min [IQR 76-422] vs. 567min [IQR 290-914], p<0.001). Only 8 monitored patients experienced electrographic seizures (n=3 for POC-EEG), precluding comparisons in time to seizure detection, treatment, or cessation. In 72% of patients monitored, POC-EEG was thought to have expedited diagnostic testing or treatment.
Conclusions:
Compared to the 10-20 conventional EEG, POC-EEG may allow for more rapid diagnostic evaluation of patients with suspected seizure. While few included patients were diagnosed with seizures in this cohort, earlier exclusion of seizure may reduce unnecessary treatment and expedite second tier diagnostic testing for altered mentation or abnormal movements.