Part of what makes epilepsy frightening is that for many people seizures happen at random. Yet other people with epilepsy report that they know when they are going to have a seizure. What proportion of patients can do this? Are there any types of seizures that are more predictable than others? These are the questions that this study, by researchers from the Comprehensive Epilepsy Management Center in New York, aimed to find answers to.

They followed a group of 71 adult patients with partial epilepsy for a year. All the participants in the study had had at least one seizure in the previous twelve months, but didn't have seizures every day. They were asked to keep a diary of their seizures, recording when they occurred and what they were like. In addition, every day they had to answer the question "Do you think you will have a seizure in the next 24 hours?", choosing an answer from the following list: "extremely likely", "somewhat likely", "somewhat unlikely", and "extremely unlikely".

Of the 1,488 seizures that occurred within the study, 32% occurred on days when seizures had been predicted. Of the 13,691 seizure-free days in the study, 83.2% had been predicted as unlikely days for seizures. Generally, seizures were twice as likely to follow a positive prediction as a negative one. There was a subgroup of twelve patients whose predictions were even more accurate, with seizures three times more likely if predicted than if not. These people were younger than the average and had a higher seizure rate.

This is the first time seizure occurrence has been studied using a "prospective" study design: that is, asking patients to predict a seizure and then checking whether they were right, rather than looking back over their seizures (a "retrospective" study) to estimate predictability.

Dr Sheryl Haut, who led the research published this month in the journal Neurology, said that this shows that some people with partial epilepsy have "a significant predictive ability" for their seizures. She suggested that with training, people with epilepsy could become more aware of seizure triggers and warning signs, and learn to adjust their daily activities to take this into account. Being able to predict when a seizure will not occur is just as important as being able to predict when it will, in this regard.

There's also the possibility of being able to develop pre-emptive treatments for seizures: an anti-epileptic drug to be taken (or a relaxation technique to be followed) only if a seizure is on its way. This would reduce the amount of medicines patients have to take. The researchers said that they now want to focus on narrowing the predictive window for seizures from 24 hours to shorter periods. They will also analyse the warning signs patients reported in this study, to see whether some are better predictors than others.

Read more here (free signup required)