Children with epilepsy and their parents, and adults with epilepsy obviously want to know when it is sensible to stop anti-epileptic medication when they have been free of seizures for two or three years. What risks of recurrence do they have, and should they stop the drugs very slowly, or does the speed with which they do this matter very much? 20 years after diagnosis, 50 per cent of a community sample will have been free from all seizures without anti-epileptic drugs for at least five years, and many will have abandoned their drugs far earlier. What advice can be given?
First of all, it must be recognized that many people are anxious about the possibility of recurrence of seizures, not least because if one occurs, a driving licence regained will be lost again for one year. However, it seems sensible to try and avoid potential adverse effects from very long-term use. In children, there may be anxieties about continued medication and potential effects on cognitive function and learning. Women in their child-bearing years may be anxious about the possible effects of anti-epileptic drugs upon the prenatal development of their babies.
Factors which indicate a significant risk of relapse of seizures on stopping anti-epileptic drugs include the epilepsy syndrome (juvenile myoclonic epilepsy being particularly likely to relapse), and the duration of epilepsy, the number of tonic-clonic seizures so far, and the need to take more than one anti-epileptic drug before control was established. All these factors, if present, suggest ‘difficult’ epilepsy, so it is not surprising if seizures recur if anti-epileptic drugs are stopped.
The EEG may occasionally be helpful about deciding when to withdraw drugs but only in children, in whom it has been shown that the presence of persisting generalized spike-wave activity makes relapse more likely. The evidence is much less impressive in adults,
Top curve: probability of completing a period of five consecutive years without seizures. For example, six years after diagnosis 42 per cent of patients have been seizure free for five years.
Middle curve: the probability of being in remission, at any time, for at least the past five years. The difference between the top and middle curve is due to relapse after achievement of a five-year remission. For example at 20 years after diagnosis 70 per cent of patients are currently free from seizures and have been for five years and a further 6 per cent have had at least one seizure-free period of at least five years’ duration, but have subsequently relapsed.
Lowest curve: the probability of being free of seizures for at least five years whilst not taking anticonvulsant drugs.
In summary, 20 years after diagnosis 50 per cent of patients were free from seizures without anticonvulsants for at least five years. A further 20 per cent continue to take anticonvulsant medication and have also been free of seizures for at least five years. Seizures continue, in spite of medication, in 30 per cent. Data from Dr J.F. Annegers and colleagues.
But it may be assumed that a markedly abnormal EEG, by indicating widespread nerve cell abnormalities, makes it rather more likely that further seizures will occur. However, the finding of any abnormality does not imply that seizures necessarily will recur, and the absence of any abnormality does not guarantee that seizures will not recur. For this reason, many specialists do not carry out an EEG before stopping treatment, but decide on the basis of the type of seizure or epilepsy syndrome that the patient has and the interval that the patient has been seizure-free.
It is generally felt, if a decision is made to withdraw anti-epileptic drugs, this should be done gradually over about two-three months or so. This is particularly important for phenobarbitone and for the benzodiazepine group of drugs (for example, diazepam or Valium and clobazam frizium) of drugs; in patients taking these drugs, abrupt withdrawal may precipitate a burst of seizures, rather like what is known to happen in someone who suddenly stops drinking after many years of abusing alcohol.
*73\188\2*
