A nationwide, retrospective, data-linkage, cohort study of epilepsy and incident dementia

Christian Schnier,Susan Duncan,Tim Wilkinson,Gashirai K Mbizvo,Richard F M Chin,Gashirai K. Mbizvo,Richard F.M. Chin
DOI: https://doi.org/10.1212/WNL.0000000000010358
IF: 9.9
2020-09-23
Neurology
Abstract:Objective To determine the association of epilepsy with incident dementia by conducting a nationwide, retrospective data-linkage, cohort study to examine whether the association varies according to dementia subtypes and to investigate whether risk factors modify the association. Methods We used linked health data from hospitalization, mortality records, and primary care consultations to follow up 563,151 Welsh residents from their 60th birthday to estimate dementia rate and associated risk factors. Dementia, epilepsy, and covariates (medication, smoking, comorbid conditions) were classified with the use of previously validated code lists. We studied rate of dementia and dementia subtypes in people with epilepsy (PWE) and without epilepsy using (stratified) Kaplan-Meier plots and flexible parametric survival models. Results PWE had a 2.5 (95% confidence interval [CI] 2.3–2.6) times higher hazard of incident dementia, a 1.6 (95% CI 1.4–1.8) times higher hazard of incident Alzheimer disease (AD), and a 3.1 (95% CI 2.8–3.4) times higher hazard of incident Vascular dementia (VaD). A history of stroke modified the increased incidence in PWE. PWE who were first diagnosed at ≤25 years of age had a dementia rate similar to that of those diagnosed later in life. PWE who had ever been prescribed sodium valproate compared to those who had not were at higher risk of dementia (hazard ratio [HR] 1.6, 99% CI 1.4–1.9) and VaD (HR 1.7, 99% CI 1.4–2.1) but not AD (HR 1.2, 99% CI 0.9–1.5). Conclusion PWE compared to those without epilepsy have an increased dementia risk.
clinical neurology
What problem does this paper attempt to address?