“Statins raise your chance of diabetes,” said the Daily Mail , but apparently the benefits in terms of heart disease protection still outweigh the risks. Taking the cholesterol-lowering drugs reportedly increases the chance of developing type 2 diabetes by 9%.
This report is based on a well-conducted review investigating the association between statin treatment and risk of diabetes. The research combined the results on 91,140 people from 13 trials. It estimated that over four years the risk of diabetes was 9% greater in those using the drugs than in those using no treatment. However, the actual number of people who developed diabetes was small, with the researchers estimating that treating 255 people with statins for four years would result in one extra case of diabetes. As a result, the benefits of statin treatment in people at risk of cardiovascular disease still appear to outweigh any small increase in risk of diabetes.
The authors of this review also conclude that the benefits of statins outweigh the small risk of diabetes, saying “clinical practice in patients with moderate or high cardiovascular risk or existing cardiovascular disease should not change”.
Non-diabetics taking statins for potential cardiovascular problems are already likely to have their blood sugar monitored periodically by their doctors, and the results of this study support this practice.
Where did the story come from?
This research was conducted by Professor Naveed Sattar of the University of Glasgow and colleagues from other centres in the UK, Ireland, Europe and the US. The review itself received no funding but trials included in the review and individual researchers conducting the review had received funding from the pharmaceutical industry. The study was published in the peer-reviewed medical journal The Lancet.
News coverage has generally reflected the findings of this review accurately.
What kind of research was this?
This was a systematic review and meta-analysis, which combined the results of previous trials in order to investigate a possible association between statin use and the development of type 2 diabetes. The best method of assessing the effects of a particular treatment is to analyse existing evidence through a well-conducted systematic review of all relevant randomised controlled trials. However, combined results are inevitably limited by the differences in the methods and results of the trials included.
What did the research involve?
The review included both published and unpublished prior research. The reviewers searched a number of medical databases for trials conducted between 1994 and 2009. Suitable trials had to have been designed to investigate the effects of statins upon cardiovascular outcomes, have included more than 1,000 people (all of whom had to be free of diabetes at the start of the study) and to have followed people for at least a year.
The reviewers only looked at trials comparing a statin with a placebo (dummy) pill or usual care, but not those trials comparing different statin drugs with each other. The reviewers used standard diagnostic criteria for diagnosing diabetes. When combining results, they applied statistical methods that took into account differences in results between the trials.
What were the basic results?
The researchers found 13 relevant trials, which included 91,140 people without diabetes at the start of the studies. Of the participants, who were followed for an average of four years, 45,521 were assigned statins and 45,619 were assigned a control treatment. In total, 4,278 participants (4.7%) went on to develop diabetes: 2,226 that had received statins and 2,052 that had been given control treatment or a placebo. However, within the individual trials there was high variability in the rate of participants developing diabetes, ranging from about 2 to 14%.
When the individual trials were analysed in isolation the association between statin use and development of diabetes was non-significant in 11 trials and significant in two. However, when the reviewers combined the results of all 13 trials in their meta-analysis, statin use increased the risk of developing diabetes by 9% overall. This association was just significant (odds ratio 1.09, 95% confidence interval 1.02 to 1.17).
Further sub-analysis of the results on each brand of statin drug found mostly non-significant results for each statin individually. There were also no differences between the risks from each statin brand.
The reviewers performed another sub-analysis to try to investigate the reason for the slight differences in risk between the 13 trials. They found that the increase in diabetes risk associated with statins was highest in trials of older participants. Neither body mass index (BMI) nor cholesterol levels at the start of the studies seemed to have any effect upon the statin-diabetes association.
The reviewers calculated that, overall, treating 255 people with statins for four years would result in one extra case of diabetes on average.
How did the researchers interpret the results?
The review concluded that statin treatment is associated with a slightly increased risk of development of diabetes, but the amount of the risk is low and outweighed by the reduction in coronary events that statins provide. The reviewers say that: “clinical practice in patients with moderate or high cardiovascular risk or existing cardiovascular disease should not change”.
This was a large and well-conducted review, which has found that the overall risk of diabetes is increased 9% by statin treatment. It should be noted that the risk of participants developing diabetes was relatively low to begin with. This means that even after the 9% increase associated with statins, the actual risk remained low.
There are further points to consider when interpreting these findings:
- Normally when combining results from different trials differences in the individual trials’ methods and results can lead to inaccuracies in the final results. However, in this case, heterogeneity (differences) between the results of the trials was non-significant, so we can have more confidence in the combined result.
- When taking the results of the trials individually, only two studies out of 13 were found to have significant associations between statins and diabetes risk. When all these results were combined into a meta-analysis, the 9% risk increase was only marginally significant.
- Only 4.7% of the overall sample (those taking statins or placebos) developed diabetes, meaning a 9% difference between the two treatment groups is still going to be small. In terms of actual numbers, there were only 174 more cases of diabetes across all the trials’ statin groups, which the researchers calculated to be one additional case of diabetes out of the 255 people treated with statins for four years. Therefore, it may be considered that this is a relatively small risk increase.
- The specific aim of the review was to investigate the association between diabetes and statin use, and so the review did not present the benefits of statins. For example, it did not compare the number of new cases of cardiovascular disease, heart attacks, CVD-related mortality in the statin and placebo groups. The benefits of statins have been shown in numerous pieces of prior research.
- As the researchers say, there may be various unidentified factors that are confounding (influencing) the observed relationship between statins and diabetes. For example, there may be more new diabetes cases in the statin group because there have been greater numbers in the control group who have actually died as a result of cardiovascular disease.
The overall conclusion of the review seems appropriate when weighing the small increase in risk of diabetes against the benefits of cholesterol treatment in people at risk of cardiovascular disease.
“Clinical practice in patients with moderate or high cardiovascular risk or existing cardiovascular disease should not change”. Clinicians could continue periodically to monitor blood sugar control in non-diabetics who are at risk of cardiovascular disease and treated with statins.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
Daily Mail, 17 February 2010
BBC News, 17 February 2010
The Daily Telegraph, 17 February 2010
Links to the science
The Lancet, [Early Online Publication], February 17 2010