BACKGROUND: Outcome studies in people with epilepsy have largely focused on the prognosis in the early stages and factors predictive of early remission. Few studies have examined prognosis in chronic refractory epilepsy. METHODS: We determined the pattern of remission and relapse of epilepsy in a cohort of people with refractory epilepsy (seizures in the past two years, at least five years after onset and who have been treated with at least 2 appropriate antiepileptic drugs during that time) to investigate whether any clinical or demographic features are predictive of seizure patterns. Seizure patterns were defined as intermittent (at least one previous period of remission of two or more years with a subsequent relapse) or continuous (no periods of remission of two years or more since seizure onset). We correlated clinical variables with these patterns. We devised a prognostic model summarising patterns of remission and relapse over time in epilepsy. RESULTS: 290 people were recruited, of whom 70% had a continuous pattern of seizures with the remaining 30% having an intermittent pattern. The only clinical variables which significantly differed between the two groups were a higher total number of antiepileptic drugs taken by those in the continuous group (P=0.01) and fewer seizures in the previous year in the intermittent group (P<0.001). A prognostic model of epilepsy is proposed. CONCLUSION: There is considerable heterogeneity in long-term seizure patterns in people who do not enter long-term remission in the early years after diagnosis.
BACKGROUND: Outcome studies in people with epilepsy have largely focused on the prognosis in the early stages and factors predictive of early remission. Few studies have examined prognosis in chronic refractory epilepsy. METHODS: We determined the pattern of remission and relapse of epilepsy in a cohort of people with refractory epilepsy (seizures in the past two years, at least five years after onset and who have been treated with at least 2 appropriate antiepileptic drugs during that time) to investigate whether any clinical or demographic features are predictive of seizure patterns. Seizure patterns were defined as intermittent (at least one previous period of remission of two or more years with a subsequent relapse) or continuous (no periods of remission of two years or more since seizure onset). We correlated clinical variables with these patterns. We devised a prognostic model summarising patterns of remission and relapse over time in epilepsy. RESULTS: 290 people were recruited, of whom 70% had a continuous pattern of seizures with the remaining 30% having an intermittent pattern. The only clinical variables which significantly differed between the two groups were a higher total number of antiepileptic drugs taken by those in the continuous group (P=0.01) and fewer seizures in the previous year in the intermittent group (P<0.001). A prognostic model of epilepsy is proposed. CONCLUSION: There is considerable heterogeneity in long-term seizure patterns in people who do not enter long-term remission in the early years after diagnosis.