Abstracts

There is no doubt that antiepileptic drugs reduce the risk of seizure recurrence for people with a single seizure or early epilepsy. However, existing data do not allow an accurate prediction of outcome for the two policies of immediate or no immediate treatment with antiepileptic drugs. In order to address this question we undertook a multicentre randomized controlled trial (RCT), which is currently the largest completed epilepsy RCT.
This multicentre RCT recruited adults or children following a single seizure or with early epilepsy for whom both clinician and patient were uncertain as to whether antiepileptic drug treatment was required. Participants were randomized to either a policy of delayed treatment or a policy of immediate treatment. For those allocated immediate treatment, the choice of treatment was made by the treating clinician. For those allocated to delayed treatment, where necessary, treatment could be initiated at any time post randomization. Recruitment started in 1993 and ended in December 2000 and participants were followed up to the end of 2001. Outcomes included time to first seizure and time to two year remission. Multiple regression models were used to assess the effect of clinical factors upon outcome and to create prognostic models.
1443 participants were randomized, 722 to immediate(45% carbamazepine, 45% valproate, 10% other) and 721 to delayed treatment. The median follow up was 4.3 years. 717 were from the UK, 624 from other European countries, 34 from India and 68 from South America. Age ranged from 5 months to 92 years. Patients allocated to delayed treatment were more likely to have a recurrence, hazard ratio 1.41 (1.21, 1.64). For patients with a single seizure, at 2 years, 10 patients would need treatment to prevent one recurrence, whilst 5 patients with 2 or more seizures would need treatment to prevent a single recurrence. Patients allocated to immediate treatment achieved an earlier 2 year remission (p=0.001), but there was no difference in the proportion of patients achieving a 2 year remission by 7 years, 95.3% immediate versus 95.8 delayed treatment. Factors that were of prognostic importance for time to recurrence include number of seizures, abnormal EEG, Age at randomization, age at first seizure, neurological deficit.
Immediate treatment for single seizures or early epilepsy reduces the risk of seizure recurrence, but does not alter the long term prognosis of epilepsy. Predictive models will enable patients and their physicians to make better informed treatment decisions.
[Supported by: Medical Research Council, UK]