Abstracts

Rationale: Sudden Unexplained Death in Epilepsy (SUDEP) is a devastating epilepsy-related event, usually seen in the context of a generalized convulsion in an unattended patient. This study is aimed to investigate the Emfit mattress monitor as an alarm system that can alert caregivers to movements associated with seizure activity and therefore potentially avoid nocturnal SUDEP. This device is a piezoelectric under-mattress sensor that detects rhythmic movements through changes in mattress pressure which activates an audible alarm system. Until now, there have been no studies performed comparing the Emfit monitor with video EEG (vEEG) detected seizures.Methods: Patients admitted for continuous monitoring were recruited from our Epilepsy Monitoring Unit over a 3 month period. Background information recorded included age, age at onset, handedness, gender, weight, number of previous and current anti-epileptic drugs, MRI findings, and epilepsy type. Data collected from the Emfit monitors were reviewed and compared to video EEG findings. Patient events were categorized into Generalized Tonic-Clonic (GTC) seizures, other clonic seizures, other motor seizures, non-motor clinical seizures, non-epileptic behaviors, and no visible changes on video or to direct testing. Results: Twenty-two patients were enrolled with 12 being female (55%), a mean age of 42 years (range 20-81), and a mean weight of 82 kg (range 61-130). Twenty-two Emfit alarms transpired over the course of 1600 hours of vEEG monitoring. Ten GTCs occurred, all in patients with focal epilepsy, resulting in 8 true alarms. All 14 false alarms occurred during wakefulness with 11 of those being due to rhythmic activity (e.g. tooth flossing, limb scratching). Overall, there were 12 non-convulsive seizures, 13 subclinical seizures, 5 non-epileptic seizures, and 3 auras, none of which triggered the device.Conclusions: The Emfit monitor was able to accurately alarm for 80% of GTCs when compared to the gold standard of vEEG monitoring and is therefore sensitive to GTC detection. Most false alarms (79%) could be explained by rhythmic movement related to activities of daily living, and none of them were noted during sleep. The device was not triggered by non-convulsive seizures, subclinical seizures, or auras, and is thus specific to rhythmic movement. Also of note, non-epileptic seizures did not activate the alarm indicating that the device is specific to epileptic convulsions. Based on our findings to date, the Emfit monitor is an effective method of GTC detection and may be a novel way of preventing SUDEP.