OBJECTIVES: To determine how switching to a new statin after treatment initiation affects new statin users' compliance and persistence, and to evaluate switching rates among five statins: atorvastatin, fluvastatin, lovastatin, pravastatin, and simvastatin. METHODS: The sample includes 24,035 patients over the age of 18 who were continuously eligible for two-years. These patients are all new statin users as defined by a one-year wash-out period before the first statin prescription. Patients with only one claim or with claims covering more than 30 days of supply were excluded. Their compliance is measured with the medication possession ration (MPR) and their persistence with the duration of continuous therapy. Logistic regression (simple or weighed with propensity score for drug treatment assignment) was used for compliance and censored regression for persistence. RESULTS: Switching negatively impacts persistence (-14.5% duration of continuous therapy) vs. no switch (p<0.0001). The time of switching also matters for persistence: Switching during the first quarter after the initial statin reduces persistence less (-8.4%, p=0.0073) than switching during the following quarters (-20.7% Q2, -17.2% Q3, -14.6% Q4, p<0.0001 for all coefficients). Switching significantly reduces the probability of high compliance defined as MPR°Ý0.8 (odds ration 0.856) vs. no switch (p<0.0001). Finally, atorvastatin users are the least likely to switch to another statin (odds ratio for probability to switch vs. atorvastatin: fluvastatin 2.66, pravastatin 2.67, lovastatin 1.54, simvastatin 1.88; p<0.0001 for all coefficients). CONCLUSION: Switching lowers both persistence and compliance, specifically by reducing the number of long-term users. Patients started on atorvastatin are less likely to be switched to another statin during the course of their treatment.