OBJECTIVES: To assess the long-term cost-effectiveness of titration from initial doses of rosuvastatin (RSV) and generic simvastatin (SIM). METHODS: Efficacy data from the STELLAR clinical trial (TC, HDL-C, TG) were used as input to the model. Markov models ran in 4-year cycles for 20 years, from age 55 to 76 years to predict primary and secondary CHD based on Framingham risk equations in four gender/risk cohorts. In year one quarterly titration up to a maximum dose of 40mg (RSV) or 80 mg (SIM) was based on a total cholesterol (TC) target of 5 mmol/l. Risk was calculated using the average TC:HDL-C ratio of 1000 simulated patients, with adjustment for Framingham's hypothesised over-prediction of UK risk. RSV and generic SIM prices for September 2004 and recent UK CHD event cost data were applied. Relative mortality risks and health-state utilities were used to derive quality-adjusted life years (QALYs). Discounting was performed at 3.5% (costs and outcomes). RESULTS: The STELLAR trial found RSV 10mg lowered total cholesterol significantly more than SIM 10–40mg (-32.9% vs. –20.3%, -25.7%, 27.9%, respectively), and RSV 20mg lowered total cholesterol significantly more than SIM 80mg (-37.6% vs. –32.9%). Based on this model less CHD events and deaths are expected among patients on RSV compared with SIM. Hence RSV delivers more QALYS at an acceptable cost per patient, e.g. £ 3458 per QALY for high-risk males compared with SIM. Insensitivity analysis when generic SIM is priced at zero, the cost per QALY gained for RSV ranged from £ 11,169 (male high risk) to £ 21,752 (female base case). CONCLUSIONS: More CHD events are likely to be avoided using RSV than SIM. RSV will be a cost-effective strategy, as defined by UK NICE thresholds for cost per QALY gained, even as the generic SIM price approaches zero.