Intended for healthcare professionals


New analysis fuels debate on merits of prescribing statins to low risk people

BMJ 2014; 348 doi: (Published 26 March 2014) Cite this as: BMJ 2014;348:g2370
  1. Zosia Kmietowicz
  1. 1BMJ

The controversy concerning the prescribing of statins to a wider section of the population showed no signs of abating this week after a new analysis—this time of high quality observational studies involving data on millions of patients—found that the risk of harmful effects from statins was very small and similar to those seen in clinical trials.1

The finding is fuelling the debate on whether statins should be extended to healthy people who are at low risk of cardiovascular disease, as recommended in draft guidance from the National Institute for Health and Care Excellence (NICE).2

The latest research comes just two weeks after the publication of a meta-analysis of 29 randomised controlled trials that found that serious adverse events occurred just as often in patients taking a placebo (14.9% of patients in primary prevention trials and 11.2% of those in secondary prevention trials) as they did in patients taking statins (14.6% for primary and 9.9% for secondary prevention).3

That the two recent analyses concur gives leverage to people who say that statins should be more widely used to prevent cardiovascular disease because the risks associated with taking them are very small.

Arguments about how far statins should be used in preventing cardiovascular disease intensified at the weekend when the Oxford University academic Rory Collins accused the BMJ of potentially putting lives at risk by publishing articles that questioned the benefits of statins in this section of the population.4

In the articles John Abramson, from Harvard University, and Aseem Malhotra, a cardiologist at Croydon University Hospital in London, argued that the side effects of statins were too great for the drugs to be used in people at low risk of cardiovascular disease.5 6

But Collins said that emphasising the harms of statins in low risk people was confusing GPs and the public and was “a serious disservice to British and international medicine” that he claimed was probably killing more people than had been harmed as a result of Andrew Wakefield’s paper on the measles, mumps, and rubella vaccine.

Collins is co-director of the Cholesterol Treatment Trialists’ Collaboration, which was set up in 1994 by Oxford University’s clinical trials service unit to answer uncertainties about the effects of lowering cholesterol concentrations. A meta-analysis from the collaboration, published in 2012, concluded that people who had a 10% risk of developing cardiovascular disease in the next five years could benefit significantly from statins and that this benefit “greatly [exceeded] any known hazards of statin therapy.”7 It was this study that led to NICE recommending in February that the threshold for starting preventive treatment with statins should be halved from a 20% risk over 10 years as assessed by the QRISK2 tool to a 10% risk.2

However, when Abramson examined the collaboration’s data he calculated that statins did not reduce the risk of death or serious illness in people at low risk and that side effects occurred in a fifth of people treated.6 Doctors would do better to talk to patients about lifestyle factors that could reduce cardiovascular risk, he said.

In the Guardian on 21 March Abramson argued that only the Oxford group has seen the patient level data that drug companies hold on statin trials and that these data should be made publicly available to end the “ongoing secrecy” that “only raises the public’s level of suspicion.”4

But the latest analysis of observational studies, published in BMC Medicine, indicates that the prevalence of side effects in clinical trials did not differ from the prevalence seen in large observational studies.1 The researchers, from the London School of Hygiene and Tropical Medicine, did a meta-analysis of 90 studies involving 4.5 million people.

They found that use of statins, when compared with non-use, was associated with a lower risk of dementia and cognitive impairment (odds ratio 0.74 (95% confidence interval 0.62 to 0.87)) and an increased risk of myopathy (2.63 (1.50 to 4.61)), liver enzyme changes (1.54 (1.47 to 1.62)), and diabetes (1.31 (0.99 to 1.73)).

When they compared these findings with those from randomised controlled trials the researchers found the risk of side effects to be “broadly similar.” They said that myopathy and elevated risk of liver enzyme changes were both rare side effects that were more likely with higher doses of statins and that there were recommendations for monitoring patients for signs of type 2 diabetes.

They concluded that the absolute risk of harms from statins was very small when compared with the benefits in preventing major cardiovascular events.

Shah Ebrahim, professor of public health and policy at the London School of Hygiene and Tropical Medicine and coordinating editor of the Cochrane Heart Group, was an author of the new study. He told the BMJ, “Patients and doctors should be reassured that no major or minor potential unintended effects of statins were found in our review of observational studies, which included patients more typical of those treated in routine medical practice than those studied in trials.

“In the future it is hoped that the pharmaceutical industry will refine their approach to collection of data and reporting of unintended effects of drugs.”

Fiona Godlee, editor in chief of the BMJ, dismissed Collins’s comparison between the BMJ articles and the Wakefield paper that wrongly suggested a link between the measles, mumps, and rubella vaccine and autism. She said, “We are talking about vastly extending the use of statins to healthy people at low risk of cardiovascular disease. The balance between benefits and harms is much more nuanced in this wider population, and this decision must be open to all possible scrutiny and debate.

“It is no longer acceptable for a decision of this magnitude to be based on a single meta-analysis of industry funded data that only a few people have seen. The BMJ articles have attracted many responses from doctors, which shows that this issue needs clarifying for those who are expected to implement national guidelines.”

Godlee said that the BMJ had already invited Collins to respond to the Abramson and Malhotra articles and would gladly publish his critique.

The consultation on the draft NICE guidance on lipid modification ends on 26 March (, and the full guideline is due to be published in July.8


Cite this as: BMJ 2014;348:g2370


View Abstract