Increased risk of glucose intolerance and type 2 diabetes with statins

BMJ 2011; 343 doi: http://dx.doi.org/10.1136/bmj.d5004 (Published 08 August 2011) Cite this as: BMJ 2011;343:d5004
  1. Christopher D Byrne, professor of endocrinology and metabolism1,
  2. Sarah H Wild, reader in epidemiology and public health2
  1. 1Institute of Developmental Sciences, University of Southampton, Southampton General Hospital, Southampton SO16 6YD, UK
  2. 2Centre for Population Health Sciences, University of Edinburgh, Edinburgh, UK
  1. cdtb{at}soton.ac.uk

In people with a moderate or high risk of cardiovascular disease, benefits of statins still outweigh the risks

Statins are considered a 21st century panacea to the extent that some people propose they should be taken by everyone over 55 years of age.1 Convincing evidence shows that statins reduce all cause mortality and prevent or postpone cardiovascular disease,2 and that in general they are well tolerated, with 11 cases of myopathy (the most serious side effect) occurring per 100 000 person years of treatment.3

However, accumulating evidence suggests that statins worsen glucose tolerance and possibly cause type 2 diabetes. This suggests a parallel with thiazides and β blockers, which are also known to increase plasma glucose, thus creating concerns over the risks of treatment versus benefit.4 An analysis of West of Scotland Coronary Prevention Study (WOSCOPS) data concluded in 2001 that randomisation to pravastatin significantly reduced the risk of developing diabetes (hazard ratio after adjusting for age, body mass index, white blood cell count, blood pressure, lipids, alcohol intake, smoking, and glucose at baseline 0.70, 95% confidence interval 0.50 to 0.99).5 This conclusion was based on 1.9% of 2999 men in the pravastatin group and 2.8% of 2975 men in the comparison group developing diabetes during a five year follow-up. However, this finding has not been replicated in subsequent statin trials. In 2009, a systematic review of five other randomised statin trials reported data on new onset (incident) diabetes and found an increased risk of diabetes with statins (relative risk 1.13, 1.03 to 1.23). This calculation was based on 4% of 25 843 people in the statin group and 3.5% of 25 776 people in the placebo group developing diabetes over a mean follow-up of 3.9 years.6

Last year a meta-analysis provided further information on the association between statins and the risk of diabetes.7 The results have generated uncertainty among clinicians and anxiety among their patients about the balance between the adverse effects and benefits of statins. The analysis of data from 13 statin trials with 91 140 participants showed that 4278 (2226 assigned to statins and 2052 assigned to control treatment) developed diabetes during a mean of four years of follow-up. Statins were associated with a small 9% increased odds of incident diabetes (odds ratio 1.09, 1.02 to 1.17), with little heterogeneity (I2=11%) between trials. Meta-regression showed that the odds of developing diabetes with statin treatment was highest in trials with older participants, and that neither baseline body mass index nor change in low density lipoprotein-cholesterol concentrations accounted for variations in risk. The investigators concluded that treatment of 255 (150 to 852) patients with statins for four years would result in one extra case of diabetes. Consequently, a paradox arises in that a treatment used to reduce cardiovascular disease risk increases the risk of developing diabetes, which is known to increase cardiovascular risk.8

There has been a flurry of recent activity to analyse data from the newer statin trials to confirm or refute these findings. One analysis of three large trials assessed the risk of diabetes with atorvastatin.9 The authors found that 80 mg per day of atorvastatin was associated with a significantly increased risk of new onset diabetes compared with placebo in the SPARCL study, with no significant difference in the incidence of diabetes compared with lower doses of atorvastatin or simvastatin in the other two trials, suggesting a dose-response effect. Although confounding by risk factors for diabetes (such as age, body mass index, membership of a high risk ethnic group) could explain the association between use of statins and risk of diabetes in observational studies, randomisation in trials means that these risk factors should be roughly equally distributed between the statin and placebo groups and should not have influenced the results.

Assuming that the statin effect on glucose tolerance is real, what is the explanation? One study found evidence that different statins have differential metabolic effects on several factors that affect glucose tolerance, providing biological plausibility for the trial findings.10 The effect of statins on inhibition of hydroxymethylglutaryl CoA reductase not only affects cholesterol biosynthesis but also influences other key downstream metabolites that have profound effects on intermediary metabolism. Statins inhibit production of isoprenoids, which modify key proteins, allowing them to interact with membranes and act as “molecular switches;” it can therefore potentially affect a range of biological functions. Statins may also directly affect insulin secretion and peripheral insulin sensitivity.

Despite uncertainties about the precise mechanisms responsible for reduced glucose tolerance with statins, clinicians should be reassured that, for most people at moderate or high risk of cardiovascular disease, the benefits of statins outweigh the risks. For every 1000 people treated with statins for secondary prevention for an average of 4.2 years, about 37 can expect to have a cardiovascular event prevented, and five can expect a serious adverse event (such as rhabdomyolysis or myopathy).11 The meta-analysis described above suggests that about four extra people for every 1000 treated with statins can be expected to develop diabetes compared with an untreated group.7

People at risk of type 2 diabetes can be fairly easily identified using simple risk scores, such as the Cambridge risk score, which includes age, sex, prescription of steroids and antihypertensive drugs, family history of diabetes, body mass index, and smoking status).12 Currently, it is not clear whether screening for diabetes in people treated with statins is appropriate, and whether statins influence the performance of the different glucose or glycated haemoglobin based tests for diabetes. However, in people at risk of diabetes, lifestyle changes to improve diet, increase physical activity, and, where appropriate, lose weight should reduce the risk of diabetes and cardiovascular disease.


Cite this as: BMJ 2011;343:d5004


  • Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Not commissioned; externally peer reviewed.