Abstracts

The majority of patients diagnosed with epilepsy gain complete control of seizures with AED therapy. Identifying those who are unlikely to enter remission early in the course of the disorder can help target specialist intervention including epilepsy surgery.
890 patients diagnosed with epilepsy and prescribed their first antiepileptic drug between August 1982 and May 2001 were evaluated longitudinally for treatment outcomes. Factors of prognostic importance were collected for each patient by review of research case notes.
Treatment outcomes were known for 780 (88%) patients newly diagnosed with epilepsy (405 male, 375 female; median age 29 years). 222 patients had idiopathic generalised epilepsy and 558 had localisation related epilepsy. Longitudinal evaluation revealed 3 main outcome groups [ndash] patients who had been seizure free for longer than 12 months and were in that state at the end of follow up (remission), those who never had a 12 month period of seizure freedom (uncontrolled) and those who were seizure free for 12 months or more before becoming uncontrolled (relapse). At the end of a median period of 79 (range 24 to 240) months of follow up, 462 (59%) patients were in remission, 276 (35%) were uncontrolled and 42 (6%) had relapsed. Patients who entered remission did so within the first year of treatment in 83% of cases. Patients with idiopathic generalised epilepsy were more likely to enter remission than those with localisation related epilepsy [odds ratio (OR) 1.47, p= 0.019]. Analysis by age groups showed that elderly patients (age [gt] 65 years at diagnosis, n = 90, 85% remission) were significantly more likely to enter remission than other adults (age 20 to 64, n = 520, 53% remission) or adolescents (age [lt]20, n = 170, 65% remission). History of head injury (n = 86, OR 3.11, p [lt] 0.001), febrile convulsions (n = 35, OR 2.69, p = 0.02), family history of epilepsy (n = 91, OR 2.02, p = 0.02), partial seizures with or without secondary generalisation (n = 357, OR 1.54, p = 0.002), presence of psychiatric co-morbidity (OR 2.13, p [lt] 0.001), younger mean age at onset (p = 0.001) and large number of pre-treatment seizures (p [lt] 0.001) were associated with the probability of not achieving remission. Duration of epilepsy prior to starting treatment, seizure clustering, previous status epilepticus, results of brain imaging, surface electroencephalographic findings and presence of learning disability or neurological deficit were found not to have a significant correlation with outcome. Multivariate analysis of all prognostic factors will be presented.
Prediction of outcomes in individual patients with epilepsy is possible using information available early in the course of the disorder. A clinical prediction model for each syndrome, appropriately validated, will help identify patients requiring early specialist intervention.