Authors :
Presenting Author: Gregory Krauss, MD – Johns Hopkins University
Samyak Shah, PhD – Johns Hopkins University
James Wheless, BScPharm, MD, FAAP, FACP, FAAN, FAES – LeBonheur Children’s Hospital
Michael Sperling, MD – Thomas Jefferson University, Philadelphia, Pennsylvania, USA.
Dewi Depositario-Cabacar, MD – Children's National Hospital
William Gaillard, MD – Children's National Hospital
Richard Chiaccherierini, PhD – R. P. Chiacchierini Consulting
Nathan Crone, MD – Johns Hopkins University
Rationale:
Tonic clonic seizures (TCSs) are associated with risks for injuries and sudden unexpected death in epilepsy (SUDEP). Mobile devices can monitor for TCSs and alert caregivers to provide assistance during seizures. Current devices for detecting TCSs have high false alarm rates (FAR), which can cause “alarm fatigue”, discouraging device usage and responses to seizure alerts. EpiWatch is an app for Apple Watch that continuously monitors for TCSs and can alert caregivers of TCSs in persons with epilepsy. We performed a large-scale phase 3 multicenter study to evaluate the performance of EpiWatch for detecting TCSs (sensitivity and FAR) in children and adults during inpatient video-EEG monitoring.
Methods:
Subjects were admitted for clinical evaluations at 4 pediatric and 2 adult epilepsy monitoring units (EMU). Subjects ages 5 and older were consented for study participation with IRB approval. During the participation period, subjects were continuously monitored, blinded to any detections from the EpiWatch app. Video-EEG of motor events were evaluated by an independent and blinded central panel. All events were classified as TCS, other seizures, and detection events, with discordant classifications resolved by consensus review. False alarms were tabulated with respect to detection of TCSs. Adverse events were monitored during the study and reviewed by a DSMB.
Results:
The study enrolled 242 subjects across 6 epilepsy monitoring units. Subjects included a broad demographic range of children and adults (See Table I) There were no serious adverse events (SAEs) or adverse device effects.
Sensitivity (positive percent agreement):
The 242 study subjects had 16,189 hours of study monitoring. They had 47 TCSs confirmed by the central panel; 46 (98%) of these were detected by EpiWatch. The EpiWatch PPA (sensitivity) was 0.94 adjusted for all age groups; 1.0 for ages 5-12 years, 0.95 for ages 13-21 years and 1.0 for adults ( >21 years). The single undetected TCS was in an adolescent whose watch arm was held down by a caregiver. (See Table II.)
False alarm rate (FAR):
There were 56 false alarms of which 20 were seizures with motor signs. This resulted in a FAR for detecting TCSs at 0.083 per 24-hour period. This point estimate equals an estimated rate of one false alarm in 12.4 days. The age groups all had similarly low FARs. (See Table II). Additionally, more than 1/3 of the false alarms were seizures with motor signs.
Conclusions:
EpiWatch detected TCSs during continuous inpatient monitoring with a high sensitivity and a low false alarm rate (FAR), which was one tenth or less the rate published for other TCS detection devices. EpiWatch’s low FAR mitigates the risk of “alarm fatigue”, which can discourage seizure monitoring and responses to seizure alerts. Continuous monitoring for TCSs using a commercial, non-stigmatizing device such as Apple Watch may be beneficial in alerting caregivers and reducing SUDEP and other risks from TCSs.
Funding: EpiWatch Inc.