Despite their familiarity and undoubted success, statins still retain a veneer of mystery that we are only just beginning to scratch through. One issue, of interest to clinicians and marketing teams alike, is why many patients respond less favourably, or not at all, to statin therapy. Now, a study in the Journal of the American Medical Association that examined patients with hypercholesterolaemia that were treated with pravastatin (Pravachol; Bristol-Myers Squibb) is the first systematic study to offer a pharmacogenetic explanation.

In a survey of 148 single nucleotide polymorphisms (SNPs) in 10 genes known to be involved in cholesterol synthesis and statin metabolism — including 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase), squalene synthase, apolipoprotein E and cholesteryl ester transfer protein — Paul Ridker and colleagues found that two polymorphisms in HMG-CoA reductase are associated with reduced efficacy of pravastatin therapy. Individuals heterozygous for SNPs 12 and 29 of the HMG-CoA reductase gene had decreased responses — 22% smaller reduction in total cholesterol and 19% smaller reduction in low-density-lipoprotein cholesterol — compared with individuals who were homozygous for the allele. Looking at the sequence and location of these SNPs, they have no obvious functional effect, so whether they lead to altered HMG-CoA reductase expression or activity, or affect statin binding is uncertain.

As great as the temptation is to translate these results rapidly into a diagnostic test to identify 'non-responders', outstanding questions remain, say the authors. First, do the data extend to other statins? Second, are there dose-response effects? Third, like any other genetic study, can the results be independently confirmed? A further issue is that the proportion of the variance that can be explained by the two HMG-CoA reductase SNPs is small in comparison with the predicted influence of clinical determinants, such as compliance and diet.

But, as Susanne Haga and Wylie Burke state in an accompanying commentary, these findings should force clinicians and policy makers to face two key questions: what evidence justifies the clinical use of a pharmacogenetic test, and what incentives will ensure that the translational research is carried out to provide the evidence? Without attention to these questions, write Haga and Burke, conflicting interests among stakeholders could slow the development of pharmacogenetic approaches to statin therapy or, even worse, lead to changes in practice that are based on speculative or incomplete findings.