Abstract Background Statins are a class of medications that reduce cholesterol by inhibiting 3-hydroxy-3-methylglutaryl-coenzyme A reductase. Whether statins can benefit patients with dementia remains unclear because of conflicting results. We hypothesized that some of the confusion in the literature might arise from differences in efficacy of different statins. We used a large database to compare the action of several different statins to investigate whether some statins might be differentially associated with a reduction in the incidence of dementia and Parkinson's disease. Methods We analyzed data from the decision support system of the US Veterans Affairs database, which contains diagnostic, medication and demographic information on 4.5 million subjects. The association of lovastatin, simvastatin and atorvastatin with dementia was examined with Cox proportional hazard models for subjects taking statins compared with subjects taking cardiovascular medications other than statins, after adjusting for covariates associated with dementia or Parkinson's disease. Results We observed that simvastatin is associated with a significant reduction in the incidence of dementia in subjects ≥65 years, using any of three models. The first model incorporated adjustment for age, the second model included adjusted for three known risk factors for dementia, hypertension, cardiovascular disease or diabetes, and the third model incorporated adjustment for the Charlson index, which is an index that provides a broad assessment of chronic disease. Data were obtained for over 700000 subjects taking simvastatin and over 50000 subjects taking atorvastatin who were aged >64 years. Using model 3, the hazard ratio for incident dementia for simvastatin and atorvastatin are 0.46 (CI 0.44–0.48, p