(684) Rapid Response EEG: Enhancing Utilization and Protocol Refinement Through Clarity Software

Introduction: The emergence of rapid response EEG (RR-EEG) has shown promise in expediting the diagnosis and treatment of seizures in the ICU. Ceribell offers real-time monitoring and automated analysis through its proprietary Clarity software that allows early detection of non-convulsive seizures (NCS) and epileptiform patterns (EP), like lateralized periodic discharges (LPD), generalized periodic discharges (GPD), and triphasic waves (TP). This study aimed to enhance utilization and cut down on calls to epileptologists (EPI).

Methods: A total of 428 ICU patients underwent evaluation of possible seizures with RR-EEG. Automated detection of seizure burden throughout a recording (calculated as the percentage of ten-second epochs with seizure activity in any 5-min EEG segment) was performed with Claritγ (> 90% triggering a status epilepticus alert which prompts treatment and EPI review, 10 - 90 % burden prompts call to EPI and then treatment based on the review, and < 10 % indicates absence of NCS or EP and no treatment).

Results: A total of 428 RR-EEGs were performed for an average duration of 9.5 +/- 0.3 hours: age (60 +/- 17 years), gender (54 % male), APACHE III scores (57 +/- 28), acute physiologic scores (46 +/- 26), and reason for EEG (unexplained encephalopathy in 64 %, seizure-like activity in 55 %, automatisms in 28 %, convulsive seizures in 8%, and post-cardiac arrest in 10%). In < 10 % (n = 282), EPI read no seizures/NCS, 2 LPD (0.7 %), 2 GPD (0.7 %), and 1 TP (0.35 %), for a negative predictive value (NPV) of 100 %. In the 10-90% (n = 118), EPI read 6 seizures (5 %), 4 LPD (3.4 %), no GPD, and 2 TP (1.7 %). In the > 90% (n = 28), EPI read 6 seizures/NCS (21.5 %), 2 LPD (7 %), 2 GPD (7 %), and 3 TP (11 %). When calculated for < 50% (n = 367), EPI read 4 seizures/NCS (1.1 %), 5 LPD (1.3 %), 2 GPD (0.5 %), and 1 TP (0.3 %), for a negative predictive value (NPV) of 87 % (86% - 88%).

Conclusions: Ceribell Clarity system demonstrated comparable performance to expert epileptologist interpretations, exhibiting a 100% NPV for < 10% and an 87% NPV for < 50% as assessed by Clarity. Institutional protocol adjustment from < 10% to < 50%, denoting absence of NCS or EP and no call to epileptologist unless high clinical suspicion, results in a notable 58% reduction in epileptologist consultation.