Statin use in postmenopausal women is associated with an increased risk of incident diabetes mellitus
- 1Centre for Population Health Sciences, University of Edinburgh, Edinburgh, UK
- 2Nutrition and Metabolism Academic Unit, Faculty of Medicine, University of Southampton, UK
- 3University Hospital of Southampton, National Institute for Health Research, Biomedical Research Centre, Southampton, UK
- Correspondence to: Christopher D Byrne
University Hospital Southampton, University of Southampton, Southampton SO166YD, UK;
The evidence that statins reduce both all-cause and cardiovascular disease mortality when used for both primary and secondary prevention is incontrovertible. However, recent meta-analyses of trials have also shown that statins increase risk of incident diabetes.1 ,2 Culver and colleagues who are from the Women's Health Initiative (WHI) have investigated the association between usage of different statins and incident diabetes in a large cohort of postmenopausal women to assess the balance of risks and benefits of statin treatment in this population.
The WHI included 161 808 postmenopausal women aged 50–79 years recruited in the USA between 1993 and 1998 in a trial of dietary modification, hormone replacement therapy (HRT) and calcium/vitamin D. A further 93 676 women participated in a linked observational study with follow-up data until 2005. Together there were 153 840 eligible consenting women included in this analysis. Data on statin use were collected from medication labels at several points and statins were classified as low (fluvastatin, lovastatin and pravastatin) or high (simvastatin and atorvastatin) potency based on their low-density lipoprotein-cholesterol lowering effects. Incidence of diabetes was identified from validated self-report of a new physician diagnosis of treated diabetes at each follow-up (every 6 months for trial participants, annually for observational study participants).
The mean age of the participants was 63.2 years and >80% were of European ancestry. At baseline, 7% of women were taking statins and statin users who were older, reported less physical activity, had higher body mass index (BMI), different dietary patterns and lower educational attainment, were less likely to be smokers or current HRT users and were more likely to have a family history of diabetes. There were 10 242 incident cases of diabetes giving an unadjusted HR of 1.71 (95% CI 1.61 to 1.83). After adjusting for potential confounders, statin use at baseline was associated with increased risk of diabetes (HR 1.48, 1.38 to 1.59) with similar estimates for the strata of low and high statin potency, history of cardiovascular disease and combinations of statin use at baseline and follow-up.
The strength of the association varied with ethnicity (HR 1.49, 1.38 to 1.62), (HR 1.18, 0.96 to 1.45), (HR 1.57, 1.14 to 2.17) and (HR 1.78, 1.32 to 2.40) for whites, African Americans, Hispanics and Asians, respectively, and BMI (HR of 1.89 (1.57 to 2.29), 1.66 (1.48 to 1.87) and 1.20 (1.09 to 1.33) for BMI of <25.0, 25.0 to 29.9 and >30.0 kg/m2, respectively).
The authors report a significant association between use of any statin and the risk of incident diabetes that is only partially attenuated by adjustment for confounding factors. The HR for lean women was higher than for obese women but there was a markedly higher absolute risk of incident diabetes among obese than lean women regardless of statin use. There were no statistically significant differences in HR by ethnicity, perhaps due to limited power of the study to detect such effects. The authors acknowledge the potential for bias and residual confounding and these limitations may contribute to the apparently stronger association between statin use and incident diabetes in this observational study than observed in trials.
Two meta-analyses of randomised statin trials reported a small increase in odds of incident diabetes associated with statin treatment (OR 1.09 for both meta-analyses, 95% CI 1.02 to 1.16 and 1.02 to 1.17, respectively).1 ,2 However, longer-term follow-up is needed to determine whether statins affect risk of microvascular disease. The accumulating evidence suggests that statins produce a small increase in plasma glucose through a peripheral effect on insulin sensitivity, and that higher doses of the more potent statins may increase glucose levels over the threshold for diagnosing diabetes. It is also not clear whether the myalgia that some people experience as a side effect of statin treatment results in a decrease in physical activity levels. There is little evidence that statin treatment has any deleterious impact on pancreatic endocrine function and insufficient evidence to decide whether screening for diabetes among people treated with statins is warranted. In conclusion, we recommend that current guidance about statin usage to decrease cardiovascular risk should not change given their effectiveness in reducing all-cause mortality. The evidence for increased risk of diabetes reinforces the importance of lifestyle change (eg, increased physical activity and weight loss if appropriate) for patients receiving statin treatment.