Skip secondary menu

Better Prediction Could Mean Better Control over Epileptic Seizures


For release: Tuesday, March 13, 2007

Despite conventional wisdom that epileptic seizures are random and unforeseeable, a new study shows that people can sometimes anticipate them, hinting at the possibility of treatments that could quell an oncoming seizure.

The study, which was published in Neurology,* involved about 70 epilepsy patients who kept diaries with daily assessments about their likelihood of having a seizure.  Overall, prediction of a seizure within the next 24 hours was associated with a two-fold increased risk of having one.  And among a subgroup of 12 patients, prediction of a seizure was associated with a three-fold increased risk of having one

Since epileptic seizures are viewed as unpredictable, “most patients live in fear of the next attack,” said the study’s lead author, Sheryl Haut, M.D, a neurologist at Montefiore Medical Center and the Albert Einstein College of Medicine in New York City.  She said the ability to sense an imminent seizure could allow patients to avoid risky activities, such as swimming, and could open the door to preemptive treatments, including drugs that wouldn’t be practical for daily use.

Epilepsy affects about 3 million people in the U.S.  Epileptic seizures – marked by convulsions, a lapse in consciousness, or both – occur when cells in the brain send out abnormal electrical discharges.  Medication and other interventions are available to control seizures, but 25 to 30 percent of patients either don’t respond to treatment or simply have fewer seizures.

Daily use of powerful drugs like benzodiazepines, a class of sedatives that includes Valium, isn’t an option for most patients.  However, benzodiazepines or similar drugs could be used to supplement a regular treatment regimen when a seizure feels imminent, Dr. Haut said.

Dr. Haut noticed that some of her patients at Montefiore seemed able to predict their seizures.  Her study, which was funded by the National Institute of Neurological Disorders and Stroke (NINDS), is the first systematic analysis of the phenomenon.  She asked more than 100 patients to keep diaries, with a daily answer to the question “Do you think you will have a seizure in the next 24 hours?”  Only data from the 71 patients who made seizure predictions for at least 30 days were included in the analysis.

Overall, those patients predicted 32 percent of their seizures and 83 percent of their seizure-free days.  A group of 12 patients who, on average, were younger and had more frequent seizures than the larger group showed consistently higher accuracy.  They predicted 57 percent of their seizures and 88 percent of their seizure-free days.

“We believe these numbers can be improved with training,” said Dr. Haut. 

The ability to predict seizures could be associated with prodrome, a set of symptoms – including mood changes, headache, and stomach upset – that can precede seizure by hours or even days, she said.  She hopes to teach patients to become better at recognizing these warning signs when they’re present.  Also, certain changes in brain activity, measurable by electroencephalography (EEG), might signal an oncoming seizure, she said.  By studying these EEG changes and monitoring how patients feel when they occur, she hopes to identify sensations that usually go unrecognized but might serve as harbingers of an oncoming seizure.

Predicting seizures and developing better ways to control them are major goals of the Benchmarks for Epilepsy Research program, developed by NINDS in collaboration with the research community and patient advocacy groups.

-By Daniel Stimson, Ph.D.

*Haut SR, Hall CB, LeValley AJ, Lipton RB.  “Can Patients with Epilepsy Predict Their Seizures?”  Neurology, January 23, 2007, Vol. 68, pp. 262-266.

Last Modified December 30, 2008