Long-term prognosis of epilepsy, prognostic patterns and drug resistance: a population-based study.

BACKGROUND AND PURPOSE: Seizures in most people with epilepsy remit but prognostic markers are poorly understood. There is also little information on the long-term outcome of people who fail to achieve seizure control despite the use of two antiepileptic drugs (drug resistance). METHODS: People with a validated diagnosis of epilepsy in whom two antiepileptic drugs had failed were identified from primary care records. All were registered with one of 123 family physicians in an area of northern Italy. Remission (uninterrupted seizure freedom lasting 2 years or longer) and prognostic patterns (early remission, late remission, remission followed by relapse, no remission) were determined. RESULTS: In all, 747 individuals (381 men), aged 11 months to 94 years, were followed for 11 045.5 person-years. 428 (59%) were seizure-free. The probability of achieving 2-year remission was 18% at treatment start, 34% at 2 years, 45% at 5, 52% at 10 and 67% at 20 years (terminal remission, 60%). Epilepsy syndrome and drug resistance were the only independent predictors of 2- and 5-year remission. Early remission was seen in 101 people (19%), late remission in 175 (33%), remission followed by relapse in 85 (16%) and no remission in 166 (32%). Treatment response was the only variable associated with differing prognostic patterns. CONCLUSION: The long-term prognosis of epilepsy is favourable in most cases. Early seizure remission is not invariably followed by terminal remission and seizure outcome varies according to well-defined patterns. Prolonged seizure remission and prognostic patterns can be predicted by broad syndromic categories and the failure of two antiepileptic drugs.