Abstracts

Rationale: Most clinical epilepsy research is carried out in referral settings, where the distribution of clinical features is biased towards more severe cases. We studied a large population-based series of incident cases, to define the range of clinical epilepsy features in the general population. Methods: All individuals born in 1920 or later who had incidence of unprovoked seizures while residing in Rochester, Minnesota between 1935 and 1994 were identified in the Rochester Epidemiology Project. Medical records were abstracted on these cases to obtain detailed information on seizure types, syndromes, seizure frequency, EEG and neuroimaging findings, and antecedent etiologic factors. Results: 916 incident cases were identified, of whom 885 (97%) gave permission for review of their medical records. Among these 885 subjects, 699 (79%) had epilepsy (i.e., ≥2 unprovoked seizures) and the remaining 21% had a single unprovoked seizure. Antecedent etiologic factors were identified in 34% of cases overall, and this proportion increased over time, from 28% in incidence cases from 1935-1954, to 32% in those from 1955-1974, and 38% in those from 1975-1994. The increase in the proportion of symptomatic cases was largely attributed to stroke, which comprised 4% of symptomatic cases prior to 1975 and 24% of those from 1975-1994. Overall, 78% of cases with epilepsy were classifiable with respect to broad epilepsy syndrome (primary generalized versus focal). The proportion of epilepsy cases who were classifiable increased from 64% of those diagnosed in 1935-1954 to 82% of those diagnosed in 1975-1994. The most frequent reason for lack of classifiability was absence of epileptiform EEG abnormalities (73% of unclassifiable cases). Among classifiable epilepsy cases, 33% had primary generalized epilepsy, 66% focal, and 1% both. Only 21% of all classifiable epilepsy cases were diagnosed with an idiopathic generalized epilepsy syndrome (IGE). Among the more recent cases (diagnosis 1975-1994, 82% classifiable), the proportion with IGEs was only 13%. The proportion of epilepsy cases who were seizure-free increased from 26% in the first year since diagnosis to 50%, 58%, 65% and 70% in the 2nd, 3rd, 4th, and 5th years respectively. Conclusions: These results provide a very different picture of epilepsy prognosis than is observed in clinical series derived from referral sources. A majority of incident epilepsy cases become seizure-free within 5 years of diagnosis. Although idiopathic generalized epilepsy syndromes are often referred to as 'common epilepsies' they comprise fewer than 25% of all incident cases. Increased survival from stroke has led to a concomitant increase in the number of symptomatic epilepsy cases in the population.