Long-term adherence to statins is poor. We assessed the relationship between cardiovascular (CV) risk and atorvastatin adherence in primary- and secondary-prevention patients, adjusting for healthy-adherer bias by incorporating preventive service use into the model. Medical and pharmacy claims from employee-based plans from 2002 to 2008 were analyzed for patients who initiated atorvastatin in 2003-2004. Adherent patients were defined as having ≥60% of days covered in the year after atorvastatin initiation and were required to have pill coverage in months 10-12. CV events were identified as hospitalizations with a primary CV diagnosis and assessed from month 13 after atorvastatin initiation until the end of follow-up (≤36 months). Cox proportional hazards models were used to examine the association between atorvastatin adherence and CV event risk, adjusting for covariates including preventive service use. The study included 94,287 atorvastatin users (79,010 primary- and 15,277 secondary-prevention patients). In both populations, nearly one-half of the patients discontinued atorvastatin after 1 year. During follow-up, ~2% of primary-prevention and ~9% of secondary-prevention patients experienced CV events. After adjusting for covariates, adherent patients in the primary-prevention population had a significantly lower risk of CV events compared with nonadherent patients (hazard ratio, 0.82; 95% confidence interval, 0.74-0.91). In the secondary-prevention population, adherence to atorvastatin was also associated with lower CV risk (hazard ratio, 0.74; 95% confidence interval, 0.66-0.82). Atorvastatin discontinuation rates were high 1 year after treatment initiation. Patients who adhered to atorvastatin treatment were at lower CV risk. Quality-of-care interventions should target improvements to therapy persistence.