Seizures and dementia in the elderly: Nationwide Inpatient Sample 1999-2008

This study aimed to examine the association between incidence of admission for a primary diagnosis of "seizure" or "epilepsy" and dementia in a nationally representative database, the Nationwide Inpatient Sample, among the elderly population (55 years of age and above) and to determine whether this relationship is different in individuals with Alzheimer's dementia versus those with non-Alzheimer's dementia.

Data were obtained from the Nationwide Inpatient Sample using appropriate ICD-9 codes. Frequencies and descriptive analysis adjusting for influence of comorbidities and confounders were utilized. A multivariate logistic regression model adjusted for age, gender, and race was used to further explore the relationship.

Those with AD had a higher risk of developing seizures or epilepsy (OR=3.07, 95% CI=2.98-3.16) as compared with cases with NAD (OR=2.20, 95% CI=2.14-2.27). After adjusting for age, the association increased for patients with AD (OR=4.065, 95% CI=3.95-4.17) but not appreciably for patients with NAD (OR=2.68, 95% CI=2.60-2.75). Adding gender and race to the model did not change the relationship for either AD or NAD. Further adjustment for African-American race did not further change the relationship for AD and seizure (OR=3.96, 95% CI=3.854-4.077) as well as for NAD and seizure (OR=2.652, 95% CI=2.575-2.731). Similarly, Hispanic race did not change the relationship significantly for AD (OR=4.1, 95% CI=4.01-4.25) and NAD (OR=2.65, 95% CI=2.56-2.74).

Patients with AD have a higher prevalence of a seizure compared with patients with NAD. Younger patients with AD were more likely to have seizures. Race, when analyzed as a whole and separately as African-American and Hispanic race, did not alter this relationship.