Page 381 - ILAE_Lectures_2015

The implications of the observed ictal asystole in a small cohort of patients to a larger, more
representative, group of epilepsy patients is unknown. If this finding is confirmed, the
potential role of pacemaker insertion in preventing a proportion of SUDEP cases needs to be
assessed.

Supervision of patients with epilepsy has emerged as the only clinically important protective
factor, independent of seizure control. The basis for this remains unclear but may relate to
body positioning and alleviation of obstructive apnoea or possibly brainstem arousal
mechanism11,18,22,93.

Existing techniques for monitoring apnoea in other clinical contexts suffer from limitations
which make them unsuitable for SUDEP prevention. These include the size and weight of
monitors, duration of real-time monitoring, difficulty of use in unsupervised conditions and,
most importantly, very poor sensitivity and specificity, mostly due to signal artefacts.
Oximetry, for example, suffers from artefacts and false alarms, and the delay between
beginning of apnoea and detection of oxygen saturation drop causes warnings to come late.

The first clinical study of a novel wearable apnoea detection device (WADD) has been
undertaken, which proved that the device works even in the presence of artefacts in healthy
subjects and individuals with sleep apnoea. WADD was evaluated in 20 healthy subjects and
10 individuals who had been referred for evaluation of possible sleep apnoea. WADD had
99.2% sensitivity, and 99.6% specificity for detection of voluntary 15–30 second apnoea
whereas the automatic software of the state-of-the-art FDA approved ambulatory SOMNO
system (SOMNOmedics Gmbh) had 37.8% sensitivity and 90.5% specificity. For
spontaneous apnoea during natural sleep and considering the expert clinician scorer as the
gold-standard WADD had 91% sensitivity and 99.5% specificity. In contrast the SOMNO
system had 12.1% sensitivity and 99.4% specificity.

The device that has been created for epilepsy patients is smaller than the prototype used in
the above study, weighing 7.5 g instead of 17 g. The device is attached to the neck with a
hydrocolloid plaster of less than 5 cm diameter and can remain in place at least overnight; it
can provide over 90% sensitivity and specificity for detection of potentially dangerous
apnoeas123.

The ideal system for monitoring a patient’s movements should have a high sensitivity and
specificity, be easy to operate and be unobtrusive. Several attempts have been made to
develop devices in order to alert patients and carers to an ongoing seizure, but unfortunately
these attempts have universally had a very low sensitivity and specificity124,125. Current
approaches for SUDEP prevention are primarily based on detecting rhythmic movement
caused by tonic-clonic seizures, with devices that are either worn by the person or installed
in the bed. The ideal device should be validated with simultaneous ictal EEG recordings. The
stationary bed seizure monitors function by either detecting noises originating from the
rhythmic banging on the bed during the clonic phase of a generalised tonic-clonic seizure
(GTCS) or from the bed springs124, or changes in mattress pressure during abnormal
movements125. Several products are commercially available, but none of them have been
tested in a clinical setting, and sensitivity and specificity are disappointingly low. The
wearable devices are based on accelerometry signals, either alone126 or in combination with
other modalities127. A commercially available wireless wrist accelerometer sensor tested in a
clinical setting detected GTCS with a high sensitivity (90%) and a low rate of false alarms
(0.2/day)128. The next step would be to test it in a more naturalistic environment. Surface
electromyography (EMG) during convulsive seizures is another way to detect ongoing
epileptic seizures. The tonic phase of GTCS is characterised by a marked increase in