OBJECTIVES: Statin therapy reduces the rate of coronary heart disease (CHD), but high costs in combination with a large population eligible for treatment ask for prioritizing. One of the tools of priority setting is cost-effectiveness analysis (CEA). Even though all trials agree on the size of the benefit, CEA of statins report contradictory results. We reviewed CEA comparing statins with no pharmacological therapy and sought to synthesize cost effectiveness ratios (CER) for categories of risk of CHD and age. The level of heterogeneity among CER and its potential sources were evaluated. METHODS: Systematic review of the published statin CEA in Medline, the British National Health Service Economic Evaluation database and authors' reference lists. Outcomes were standardized by inflation and currency and stratified by risk and age. CER under US$ 20,000 per life year saved were "cheap", over US$ 40,000 were "expensive" and in between were "moderate". Sources of heterogeneity in the outcomes after stratification were analyzed using a regression model. RESULTS: Twenty-four studies were included, yielding 216 CER. The range of CER varied 10-fold (savings to $489,394). Disagreement existed at levels of risk lower than 4%. After adjusting for risk in the regression model, the only significant variable was funding source. CONCLUSIONS: Statin therapy is cheap for high levels of risk, but discrepancies exist at lower levels. Scarce evidence exists for younger and older ages (<45, > 65). Competing interests between pharmaceutical companies wanting to sell more and governments wanting to spend less may have strong impact on the results. Clear guidelines for CEA could reduce their vulnerability to bias and assure reproducibility.