OBJECTIVES: To assess demographic and clinical factors associated with statin selection among elderly patients. METHODS: A retrospective cohort study was conducted to examine predictors of statin selection among patients aged ³65 years, who initiated pravastatin (PS) vs. simvastatin (SS), atorvastatin (AS), or rosuvastatin (RS) between 1/1/2007 and 12/31/2007. Index statin use was defined as the first statin claim following at least 90 days of no statin access. Multiple logistic regression models were employed to assess predictive factors of PS initiation versus other statin initiations. RESULTS: Of 96,450 statin users identified, there were 8,165 PS initiators, 38,099 AS initiators, 11,320 RS initiators, and 38,866 SS initiators. Compared to other statin users, a higher percentage of PS initiators were aged75-85 (PS: 42.3%, SS: 41.0%, AS: 41.1%, RS: 38.0%, P<0.01) and more likely to be female (PS: 56.6%, SS: 51.3%, AS: 51.5%, RS: 55.7%, P<0.01). PS initiators were more likely to have atrial fibrillation (PS: 10.0%, SS: 9.4%, AS: 9.6%, RS: 8.6%, P<0.01) and take warfarin (PS: 11.7%, SS: 9.8%, AS: 10.4%, RS: 9.9%, P<0.01) and triazoles (PS: 3.1%, SS: 2.4%, AS: 2.6%, RS: 2.8%, P<0.01) in the baseline period. A higher percentage of PS initiators took more than 3 unique medications 90-day prior to the index date (PS: 85.4%, SS: 81.8%, AS: 83.0%, RS: 83.9%, P<0.01). After controlling for demographic and clinical characteristics, use of warfarin was associated with initiating PS compared with SS, AS, and RS. Other predictors of PS initiation included prior history of atrial fibrillation, HIV infection and use of calcium channel blockers, ezetimibe, and fenofibrates over the 1-year pre-index period compared with AS and SS. CONCLUSIONS: Patient profiles were different among PS users compared to other statin users. Selected comorbidities and prior use of certain medications were significant predictors of PS initiation among a cohort of Medicare patients.