Abstracts

Rationale: Determining the epidemiology of epilepsy in the general population is essential to providing appropriate care. The objective of this study was to determine the ten-year incidence, prevalence and mortality of epilepsy among residents of the province of Saskatchewan, Canada, stratifying by age, sex group and Register Indian status. Methods: A population-based retrospective cohort design was used and data was extracted from the ten provincial health administrative databases linked to each individual by a unique health services number. This cohort included individuals with the first identification of epilepsy from January 01, 2001 to December 31, 2010, based on meeting criteria for case definition of epilepsy from one of two administrative health databases (hospital separation data and physician's service data). The case definition was fulfilled if the subject had at least one hospital separation with a diagnosis of epilepsy (ICD-9 345 or ICD-10-Ca G40) during the ten year period or had two physician visits within 730 days before or within the ten-year period of the study (sensitivity= 88.9 95%CI 85.4-92.5 and specificity =92.4 95%CI 85.7-92.8) (1). Results: A total of 6,396 incident cases of epilepsy were identified during 2001/2010. The incidence rate decreased from 109.14 in 2001 to 43 person per 100,000 in 2010. The incidence of epilepsy was higher in elderly, but not in early life. Furthermore, the incidence of epilepsy was invariably higher in Registered Indians (RR= 1.77, CI 1.46-2.16), and slightly more frequent in males than in females (RR= 1.06, CI 0.94, 1.19). The point prevalence of epilepsy in Saskatchewan increased from 3.01 in 2001 to 8.2 per 1000 person in 2010. The prevalence was higher among individuals aged 35-45 years. Over the ten-year follow-up, the standardized mortality ratio among patients with epilepsy was between 3.49 and 5.2 higher than the mortality in the general population Conclusions: This is the first Canadian study of incidence and mortality of epilepsy involving the entire age spectrum. We did not find a classical U shape incidence curve as it has been found in other developed countries, finding only a high incidence in the elderly population. The prevalence and mortality are well within the range of those reported in the Rochester studies(2,3). Our new findings are foundational for public awareness and policy recommendations, health promotion and prevention strategies, appropriate health resource planning, and research priorities. Funding: Department of Medicine, University of Saskatchewan