Statins and the risk of developing diabetes

BMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f3156 (Published 23 May 2013) Cite this as: BMJ 2013;346:f3156
  1. Risto Huupponen, professor1,
  2. Jorma Viikari, professor2
  1. 1Department of Pharmacology, Drug Development and Therapeutics, University of Turku, 20520 Turku, Finland
  2. 2Department of Medicine, University of Turku, Turku, Finland
  1. risto.huupponen{at}utu.fi

Overall benefits of statins still clearly outweigh potential risks

Statins are the cornerstone of treatment for dyslipidaemia, but a recent meta-analysis of randomised trials found an association between their use and incident diabetes.1 In a linked observational study (doi:10.1136/bmj.f2610), Carter and colleagues provide an important contribution to this topic.2 Using an epidemiological approach, they analysed administrative claims data from Ontario, Canada, to retrospectively examine the comparative risk of incident diabetes for different statins in users aged over 66 years. They found that, compared with pravastatin, treatment with atorvastatin, rosuvastatin, or simvastatin—but not fluvastatin or lovastatin—was associated with an increased risk of incident diabetes in statin-naive older patients without diabetes. The risk of incident diabetes did not differ between patients with or without cardiovascular disease. Importantly, in line with a meta-analysis showing increased risk of diabetes for an intensive statin dose versus a moderate dose,3 Carter and colleagues found an association between development of overt diabetes and potency and dose of statins. The increase in risk became non-significant for patients treated with rosuvastatin when dose was taken into account.

These findings may be interpreted to support a true pharmacological class effect of statins on the development of diabetes. However, an alternative explanation—prescribing bias—cannot be ruled out. Patients with an unfavourable risk profile and higher risk of developing diabetes may have been more likely to be prescribed a more potent statin and at a higher dose. As the authors acknowledge, information on body weight, family history, and laboratory parameters was not available and could not be included in the analysis, which is an important limitation.

A recent meta-analysis of 13 randomised controlled trials with more than 90 000 participants found a 9% increased risk for incident diabetes after four years of statin treatment, particularly in older people.1 Two hundred and fifty five patients had to be treated with a statin for one case of diabetes to be diagnosed (95% confidence interval 150 to 852). Diabetes is only one risk factor, although an important one, for cardiovascular disease. For each 1 mmol/L reduction of low density lipoprotein-cholesterol achieved through statin treatment, there is a 10% proportional reduction in all cause mortality, a 20% reduction in deaths from coronary heart disease, and a 22% reduction in major vascular events.4 On that background, the overall benefit of statins still clearly outweighs the potential risk of incident diabetes.

The present results lead us once again to the basic principles of pharmacotherapy. All drugs have beneficial and harmful effects, and the doctor must find a justified balance between them. For statins, prevention of cardiovascular events and the association with incident diabetes seem to go hand in hand; the greater the potency and higher the dose, the larger the effect for both outcomes. The balance favours use of statins, provided the treatment is targeted properly. Patients with dyslipidaemia and low total risk for cardiovascular events may benefit most from lifestyle modifications, which might be enough to avoid pharmacotherapy.5 However, even if statin treatment is started, dietary modification should not be forgotten. A Mediterranean-type diet—high in fruit and vegetables and low in saturated fats—has a clear additive cholesterol lowering effect when used together with statins.6 When total cardiovascular risk favours statin treatment, a low dose, low potency agent should be used to begin with. The dose should be tailored according to the response, bearing in mind that, in addition to the laboratory values, the total cardiovascular risk score of the patient should guide treatment. Antihypertensives should be added to treat the total cardiovascular risk if needed. The most potent statins, at least in higher doses, should preferably be reserved for patients who do not respond to low potency treatment but who have a high total risk of cardiovascular events. In patients with low total cardiovascular risk the risk-benefit ratio of an intensive statin treatment may be unfavourable.

Paradoxically, statins have been shown to reduce cardiovascular events in patients with diabetes, so these drugs play an important role in the treatment of these patients. Each 1 mmol/L reduction in low density lipoprotein-cholesterol achieved with statin treatment results in a 21% reduction in major vascular events and 9% reduction in total mortality over four years in patients with diabetes.7 Statins have been suggested to have multiple pleiotropic actions,8 and their effects on glucose homeostasis and the cardiovascular system may be mediated through different mechanisms. The mechanism of the diabetogenic action of statins is still unclear, but—as has been suggested recently—statins may inhibit insulin secretion through their effects on pancreatic β cells.9


Cite this as: BMJ 2013;346:f3156


  • Research, doi:10.1136/bmj.f2610
  • Competing interests: We have read and understood the BMJ Group policy on declaration of interests and declare the following interests: JV has received personal fees from MSD Finland; RH has received grants from the Social Insurance Institution of Finland and the Academy of Finland.

  • Provenance and peer review: Commissioned; not externally peer reviewed.