Objective: We systematically synthesized the epidemiologic literature on mortality in patients with epilepsy (PWE) by epilepsy-related clinical characteristics with an aggregate data meta-analysis.
Methods: We systematically searched 15 electronic databases, browsed reference lists of pertinent publications, and contacted authors in the field. We were interested in cohort studies that reported the relative risk of death in representative epilepsy populations relative to the general population, with exclusion of highly selected subpopulations of PWE, such as patients with intellectual disabilities or epilepsy surgery series. Search, data abstraction, and study quality assessment with the Newcastle-Ottawa Scale were all performed in duplicate.
Results: Pooled mortality was threefold (relative risk 3.33, 95% confidence interval 2.83-3.92) in 38 epilepsy cohorts including 165,879 patients (79.6% from Nordic countries). Among incident cases, idiopathic epilepsies did not associate with materially increased mortality (1.29, 0.75-2.20; 4 studies), whereas mortality was almost twofold in cryptogenic epilepsy (1.75, 1.20-2.54; 5 studies), and highly elevated in patients with symptomatic epilepsy (4.73, 3.27-6.83; 12 studies) and especially in epilepsies due to congenital or developmental causes (10.3, 4.03-26.2; 2 studies). Newly diagnosed patients who attained seizure freedom did not have elevated mortality (0.97, 0.73-1.30; 2 studies).
Conclusion: Excess mortality was highly related to the etiology of epilepsy in all ages. In adult patients without neuroradiologic abnormalities or other identifiable cause of epilepsy, only patients with cryptogenic epilepsy exhibited excess mortality. Risk of premature death was lowest in idiopathic epilepsy and in PWE who attained seizure freedom.
© 2014 American Academy of Neurology.