RESUMO
OBJECTIVE: To assess the association between statin therapy and risk of Alzheimer disease (AD) in a prospective cohort study with documented statin exposure and incident dementia. METHODS: This is a prospective, cohort study of statin use and incident dementia and probable AD. A cohort of 2,356 cognitively intact persons, aged 65 and older, were randomly selected from a health maintenance organization (HMO), and were assessed biennially for dementia. Statin use was identified using the HMO pharmacy database. A proportional hazards model with statin use as a time-dependent covariate was used to assess the statin-dementia/AD association. RESULTS: Among 312 participants with incident dementia, 168 had probable AD. The unadjusted hazard ratios (HRs) with statin use were 1.33 (95% CI 0.95 to 1.85) for all-cause dementia and 0.90 (CI 0.54 to 1.51) for probable AD. Adjusted corresponding HRs were 1.19 (CI 0.82 to 1.75) and 0.82 (CI 0.46 to 1.46). A subgroup analysis of participants with at least one APOE-epsilon4 allele who entered the study before age 80 produced an adjusted HR of 0.33 (CI 0.10 to 1.04). CONCLUSION: Employing time-dependent proportional hazards modeling, the authors found no significant association between statin use and incident dementia or probable AD. In contrast, when the data were analyzed, inappropriately, as a case-control study, the authors found an OR of 0.55 for probable AD, falsely indicating a protective effect of statins. Study design and analytic methods may explain the discrepancy between the current null findings and earlier findings.