Statins are one of the most widely used drug classes, with approximately 50 million prescriptions dispensed in England alone in 2008. Their efficacy and safety in significantly reducing cardiovascular events in moderate-to-high-risk patients has been well documented, both in primary and secondary prevention.1 Commonly reported side-effects include muscle aches and increases in liver enzymes, but, in general, statins are well tolerated with a low incidence of side-effects. However, the recent collaborative meta-analysis of 13 major placebo-controlled statin trials by Sattar and colleagues reports a 9% increased risk for incident diabetes over 4 years (OR 1.09; 95% CI 1.02-1-17) in patients randomised to statins compared to those assigned to placebo. Heterogeneity between trials was low (I²=11%), suggesting that this risk appears to be a true class effect, despite known differences in lipophilicity and metabolic clearance pathways between individual statin drugs.

Although these findings are of concern, it is worth pointing out that all meta-analyses have inherent limitations. Importantly, as …