OBJECTIVES: The availability of enhanced Medicare Part D plans with generic-only coverage during the coverage gap (i.e. donut hole) and the genericization of pravastatin and simvastatin in early 2006, respectively, created a unique opportunity to examine how patient cost-sharing differentials for brand-name and generic drugs impact therapeutic substitution and medication adherence with statins.  METHODS:  Using the 2006 5% Medicare files we identified continuous fee-for-service Part D covered patients with hyperlipidemia (ICD-9-CM 272.0-272.4) using the brand-name statins (atorvastatin or rosuvastatin) between 1/1/2006 and 3/31/2006.  The quasi-experimental study examined changes in outcomes across the pre- (prior to coverage gap) and post- (during coverage gap) periods for non-low income subsidy (non-LIS) patients with generic-only coverage in gap who faced substantial increases in brand-name statin cost sharing ($28 to $82 per 30-day supply) but no changes in generic cost-sharing from the pre- to post-period compared to a control group of LIS patients who faced no changes in brand-name ($3) or generic ($1) cost-sharing in the pre- and post-periods.  The two groups were propensity-score matched on sociodemographics and clinical characteristics.  Difference-in-difference regressions were used to examine impact on use of brand-name statins, generic statins, and overall any statin use.  In addition, overall adherence (PDC>0.80) to and discontinuation (30-day gap) of statins was also examined.      RESULTS: Compared to LIS patients, non-LIS patients had a larger decline in brand-name statin use (-0.24 30-day-supply/month, p<0.001), smaller increase in generic statin use (+0.06 30-day-supply/month, p<0.001), and hence, an overall drop in any statin use (-0.18 30-day-supply/month, p<0.001).  Correspondingly, a decrease in adherence (OR: 0.81,p<0.001) and increase in discontinuation (OR: 1.62,p<0.001) on any statins was observed in non-LIS patients relative to LIS patients. CONCLUSIONS: A substantial brand/generic cost-sharing differential was associated with therapeutic substitution of brand-name statins with generic statins; however, it also resulted in declines in adherence and increases in discontinuation of statins.