Seizure Prediction Device Shows Proof of Concept

May 14, 2013

Researchers in Australia have shown for the first time in humans that epileptic seizures can be predicted by an implantable device measuring electroencephalography (EEG) patterns. They suggest that arming patients with such information can help them make practical decisions about their daily activities, and additional interventions could be given in an attempt to prevent the seizures during high-risk periods.

"Much of the disability related to seizures is not knowing when they are going to occur," lead author, Mark J. Cook, MD, from St. Vincent's Hospital, Melbourne, Australia, commented to Medscape Medical News. "This puts huge limits on what patients can do for fear of a seizure occurring. If we could predict when a seizure has a high likelihood of occurring then this would be extremely useful from a practical point of view in allowing patients to go about their day-to-day lives."

The device, which is implanted into the brain, analyzes electrical activity and communicates with an external hand-held unit that shows 3 different-colored lights: red for high risk for a seizure, white for moderate risk, and blue for low risk.

Dr. Mark J. Cook

"If you are on a red day, then you can make the decision not to go work, or put yourself in any position which could be dangerous," Dr. Cook said. "It may also be possible to take some preventative measures aimed at stopping the seizure from happening, such as taking a short-acting antiseizure drug."

Dr. Cook likens the approach to that used for predicting earthquakes. "We collected an enormous amount of EEGs to identify which patterns were associated with imminent seizures. This is the first time long-term intracranial EEG data has ever been recorded."

Implantable lead assemblies, telemetry unit, and device that advises the patient of pending seizure activity

Their results were published online in Lancet Neurology on May 2.

Like Predicting Earthquakes

For the study, the device was implanted into 15 volunteers with drug-resistant epilepsy. Data were collected so that individual brain patterns that preceded seizures could be identified. In patients in whom such patterns were identified, the device was then programmed to look for these patterns. The process was modified over time as more seizures occurred.

There were 11 device-related adverse events within 4 months of implantation, 2 of which were considered serious (device migration, seroma); an additional 2 serious adverse events occurred during the first year after implantation (device-related infection, device site reaction) but were resolved without further complication.

The devices generated suitable algorithms in 11 patients. Three patients' algorithms did not meet performance criteria, and 1 patient required device removal because of an adverse event before sufficient training data were acquired.

In terms of efficacy, the researchers found that the device showed sensitivities for a high likelihood of a seizure (over moderate) ranging from 65% to 100%. Clinical effectiveness measures did not significantly differ between baseline and 4 months after implantation.

The mean warning time of the red advisory light was 114 minutes before the seizure occurred, although this time varied widely, from 5 to 960 minutes. Patients with the lowest proportions of time in the red advisory typically reported the highest satisfaction with the device and were able to make lifestyle adjustments, such as going swimming when at low risk, or going to bed early or warning bystanders if at high risk, the authors note.

Dr. Cook added that the device had a very high negative prediction rate. "There were only a couple of occasions where a seizure occurred on a blue light, and these were explained by other external factors."

The researchers conclude, "If these findings are replicated in larger, longer studies, accurate definition of preictal electrical activity might improve understanding of seizure generation and eventually lead to new management strategies."

The main disadvantage of the device is the invasive procedure needed for its implantation. On this, Dr. Cook said, "It does require leads to be implanted intracranially, but many of the patients had had previous surgery for epilepsy and already had a hole in their scull, and we could use that. For those who hadn't had previous surgery, then, yes, the requirement for craniotomy is a big deal. This is the major limitation of this approach at present, but we are working on a less invasive device."

The researchers had wondered whether the requirement for invasive brain surgery would discourage people from volunteering for the study, but this does not appear to be the case.

"Since initial news reports of our study, we have been inundated with patients wanting to try the device. This reflects how poorly the medical world regards the uncertainty component of epilepsy. Many patients are prepared to undergo a major procedure to remove this uncertainty from their lives," Dr. Cook told Medscape Medical News.

New Sponsor Needed

Unfortunately, the company that manufactured the device, NeuroVista, no longer exists, and so patients are being told they have to wait until another sponsor is found to fund further development. "We are hoping that another company will come along and pick it up and that the current publication will help in this regard. We need to get the show back on the road," Dr. Cook said.

Dr. Cook and colleagues also looked at the approach of giving patients a short-acting antiseizure drug such as clonazepam on red light days to see whether that could prevent seizures. "While some patients said this was effective, we didn't have sufficient numbers to show anything meaningful, but this is something we can look at in future studies," Dr. Cook commented.

In the paper, the researchers write, "Patients' abilities to interpret and apply predictive data to gain clinical benefit were anecdotally varied and unproven by quantitative metrics." Dr. Cook also suggested that this device could be combined with already available devices that provide electrical stimulation to the brain to stop seizures, so that seizure prediction would be paired with a therapeutic strategy in 1 unit.

In an accompanying comment, Christian E. Elger and Florian Mormann, University of Bonn Medical Centre, Germany, say that the clinical implications of the study are unclear, but "the presented results suggest that at least some patients would view the warning device as beneficial."

They point out another potential beneficial finding from the study: that the implanted units picked up more seizures than patients reported.

"Until now, uncertainty in seizure counting has been ignored by much of the epilepsy community," they conclude. "We hope the findings presented by Cook and colleagues will increase awareness of this crucial problem and open up a new research avenue for EEG-based, automated seizure-documentation devices that could eventually replace seizure diaries."

This study was funded by NeuroVista. Some of the coauthors were employees of the company.

Lancet Neurol. Published online May 2, 2013. Article Comment


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: