Implantable cardiac monitor shows promise for automated seizure detection

08 May 2023 byJairia Dela Cruz
Implantable cardiac monitor shows promise for automated seizure detection

A subcutaneously implanted cardiac monitor device, which incorporates an algorithm designed based on heart rate variability (HRV), may accurately detect focal seizures in individuals with ictal autonomic changes.

“The HRV seizure detection algorithm was based on the modified cardiac sympathetic index (CSI) determined from 100 successive R-R peaks. This approach was feasible only for the patients who had marked autonomic ictal changes, ie, an ictal change in heart rate >50 beats per minute (bpm),” according to a team of investigators from Aarhus University Hospital, Aarhus, Denmark, led by Prof Jesper Jeppesen. [Epilepsia 2019;60:2105-2013; Epilepsia 2020;61Suppl1:S41-S46; Seizure 2015;24:1-7]

“Yet, most patients with focal seizures (53–57 percent) had such [ictal] changes, and among them, the algorithm had a detection sensitivity of 83.3–90.5 percent for the nonconvulsive focal seizures, with a false alarm rate of 1/24-hours,” they added.

In a proof-of-concept study, the implantable cardiac monitor (ICM) device was able to detect 50 of 54 focal seizures, with a corresponding sensitivity of 92.6 percent (95 percent confidence interval [CI], 85.6–99.6). Furthermore, only four of the 141 nonseizure (control) events were false positives, with a false alarm rate of 2.7 per 24 hours. [Epilepsia 2023;doi:10.1111/epi.17612]

“Using heart rate increase instead of our HRV-based algorithm would have resulted in a much higher number of false detections,” Jeppesen said.

Accurate and safe

The proof-of-concept study included six patients (three women) for admission to an epilepsy monitoring unit (EMU) for long-term video-EEG monitoring. The age of these patients was 23, 35, 46, 50, 51, and 64 years, respectively. All of them were likely to have ictal autonomic changes, with one having seizures with tachycardia previously documented in the EMU and five having focal-to-bilateral tonic-clonic seizures in their history.

An ICM loop recorder was subcutaneously placed between the sixth and seventh ribs on the left side of the chest, approximately 5 cm away from the centre, in all patients. The device was connected to a mobile application installed on the patients’ smartphones. The patients, their close relatives, and the EMU staff were taught how to use the app and mark an event after experiencing a seizure.

Of the 54 focal seizures recorded overall, 13 occurred in the EMU and 41 in an out-of-hospital setting. The ICM identified 12 and 38 of them, respectively, with a sensitivity of 92.3 percent (95 percent CI, 77.2–100) and 92.7 percent (95 percent CI, 84.7–100).

“The device was tolerated well by all patients and no adverse effects occurred during the monitoring period (up to 8 months),” according to Jeppesen.

The patients did not report any significant discomfort with the ICM. On the 7-point Likert scale, the median was 2 (slightly uncomfortable) for the device implantation and 1 (not uncomfortable) for the monitoring.

More convenient

Compared with other wearables that use surface electrode patches, subcutaneous ICM devices have several advantages, as Jeppesen pointed out.

“The implanted devices do not need daily maintenance, the lifetime of the devices is long (up to 3 years with current settings), their size is much smaller, and skin-irritation caused by prolonged use of electrode patches can be avoided. The ICM provides a stable contact and good quality signal throughout the monitoring, avoiding device deficiency periods due to patients forgetting to change the electrodes or misplacing them,” he said.

As for the need for an implantation, Jeppesen noted that the subcutaneous insertion of the device is minimally invasive and is already a routine procedure in cardiology.

“The results of our proof-of-concept study suggest that indications of automated HRV-analysis on subcutaneous ICM may be extended to automated seizure detection in patients with ictal autonomic changes. Nevertheless, these findings need to be confirmed in large, phase III trials in the future,” he added.