Intended for healthcare professionals

Feature Publication Ethics

Risks in the balance: the statins row

BMJ 2014; 349 doi: https://doi.org/10.1136/bmj.g5007 (Published 07 August 2014) Cite this as: BMJ 2014;349:g5007
  1. Nigel Hawkes, freelance journalist, London
  1. nigel.hawkes1{at}btinternet.com

Nigel Hawkes asks where the fallout over The BMJ’s correction of two articles leaves patients and doctors

Imagine, if you will, a patient newly prescribed a statin by his general practitioner: let’s call him Mr Lowrisk. It is Thursday 15 May, and as Mr Lowrisk sits down to breakfast the Today programme on BBC Radio 4 is broadcasting a discussion about the risks and benefits of statins. He listens carefully—a golf partner has warned him that statins can make your muscles ache. Defending statins, Rory Collins, an Oxford professor, asserts confidently, “There is a very, very low risk of muscle problems,” a reassuring message for Mr Lowrisk. On his way to work he picks up his prescription, and at coffee time he reads the patient information leaflet.

To his surprise, listed under common side effects that may affect “up to one in 10 people” are joint pain, muscle pain, and back pain. “But I thought that professor said there was a very, very low risk of muscle problems,” he mutters. Mr Lowrisk is confused. The manufacturers seem to be owning up to a side effect that the Oxford professor said barely even existed.

This vignette, oversimplified as it may be, encapsulates an argument over the risks and benefits of statins that has been raging ever since The BMJ published two articles on the subject in October last year.1 2 The reputation of the journal has been called into question; so has that of the authors responsible, charged by Collins with deliberately misconstruing the evidence and unfitted, in his view, from ever contributing to the journal again. Comparisons have been drawn, again by Collins, with the scare over the measles, mumps, and rubella vaccine, in which parents were discouraged from getting their infants vaccinated by false evidence linking MMR to autism. “It is a serious disservice to British and international medicine,” he told the Guardian, adding in reference to the MMR scare: “I would think the papers on statins are far worse in terms of the harm they have done.”3

Calls for retraction

Collins’s complaint was that the two articles exaggerated the harms of taking statins and as a result concluded that the drugs should not be prescribed to people at low risk of cardiovascular disease, like our imaginary golfer. Collins called for the articles to be retracted. A panel set up by The BMJ to consider the matter and led by Iona Heath, former president of the Royal College of General Practitioners, has now concluded that the errors he and others identified, already the subject of corrections, were insufficient, when judged by the criteria of the Committee on Publication Ethics, to justify retraction.4 Collins, not unexpectedly, disagrees.

The row has set a powerful cadre of experts, led by Collins and including the Cochrane Collaboration, the British Heart Foundation, and the National Institute for Health and Care Excellence (NICE), against a more diverse group of doctors, statisticians, and epidemiologists who question whether the evidence is really so compelling as to justify further extending the availability of statins to lower risk groups.

Among these sceptics is Klim McPherson, a colleague of Collins at Oxford and chair of the UK Health Forum, an alliance of professional and public interest organisations working to reduce the risk of non-communicable diseases. He asked, “Where are the data? It’s quite obvious to me that muscular myopathy is much more common than Rory believes and that it’s quite disabling,” he said. “It’s very worrying. NICE has dodged the issue completely by not seeking those data, when it should have done. What is clear is that we don’t have adequate data about the side effects, and those data that we do have are misleading.”

Others questioning the need to medicalise so many people wrote to the Times accusing those responsible for the NICE guideline on statins of having ties to the companies that made them—“a completely unjustified attack on their integrity,” retorted Mark Baker of NICE.5 The signatories to that letter included some heavyweights, such as Richard Thompson, president of the Royal College of Physicians, and Clare Gerada, former chair of the Royal College of General Practitioners.6

Cochrane Collaboration U turn

How did the row ignite, and why did it burn so bright? New readers start here.

At the end of January 2013 the Cochrane Collaboration published an analysis of the use of statins in the primary prevention of cardiovascular disease.7 Reversing a conclusion it had reached just two years earlier, it said that the evidence now justified the use of statins in people at low risk of cardiovascular disease. The evidence came from the Cholesterol Treatment Trialists (CTT) collaboration, which is led from the Clinical Trials Support Unit at Oxford, co-directed by Collins. The 2013 Cochrane review included as an appendix an exchange with Collins in which he had criticised the 2011 version for not including evidence from a paper the CTT had published in 2010. He called for urgent revision and insisted, successfully, on the correction of the 2011 press release, which he called “dangerously misleading” for saying that statin use in people at low risk might do as much harm as good.

It was in this exchange that Collins first drew a parallel with the MMR scare. Calling for a public retraction by the Cochrane Collaboration of the press release, he wrote, “In public health terms, it is potentially a far more serious misrepresentation than that of the risks of MMR by Wakefield and The Lancet.” Since readers of the review had not seen the press release, sent only to journalists, the collaboration responded to this demand by sending the same journalists a correction and noting it on its website, in what Collins described as a “misleadingly half-hearted” way. There the matter rested, but a fuse had been lit.

Cochrane did indeed change its mind in the direction Collins wanted, but some experts doubted that the new evidence, including a meta-analysis published by the CTT in 2012, was enough to justify the U turn. In June 2013 John Abramson of Harvard Medical School and colleagues submitted a paper to The BMJ challenging the claim that statins reduced all cause mortality and cardiovascular events in people at low risk without any increase in adverse events. The CTT’s own figures, they said, showed no reduction in all cause mortality. Given this absence of benefit, exposing low risk people to the side effects of statins was unjustified. The paper was published in October 2013.1

The claims were challenged in a rapid response to The BMJ by the lead author of the Cochrane group, Mark Huffman.8 While conceding that no strong evidence of a reduction of all cause mortality had been shown, he said that this was because the numbers were small: only 1% of the control group died over four years. He also disagreed with Abramson’s interpretation of the evidence.

Personal visits

No blood was shed in this exchange, nor any call for a retraction made. But a few days later Collins visited The BMJ’s editor in chief, Fiona Godlee, and raised the temperature. His complaint was that Abramson’s statements about the side effects of statins were misleading, claiming that this was worse than the MMR scare and demanding retraction of the Abramson paper and of an Observations article by the cardiologist Aseem Malhotra that had been published in the same issue and included the same claim about the prevalence of side effects.2 Godlee invited him to write an article putting the points he had made to her on record—an invitation never taken up.

Corrections to the two articles were in due course made—too tardily in the view of the Heath panel—but meanwhile Collins had gone public with his criticisms in the interview in the Guardian on 21 March 2014.3 This generated plenty of follow-up coverage, not all of it entirely accurate. The Daily Record said that the papers had been withdrawn, when they hadn’t; the Daily Telegraph included the same claim in a standfirst over an otherwise accurate article; and in his column in the Western Mail Rhodri Morgan, former first minister in the Welsh Assembly, asserted that the papers had included “schoolboy howlers in the statistics, huge miscalculations.”

The reality was more modest. Both papers had included the claim that a fifth of people who took statins experienced side effects. This figure came from an observational study by H Zhang and colleagues that actually showed that 18% of statin users had “statin-related clinical events that may be interpreted as adverse reactions by patients or clinicians.”9 The original paper made no claim of causation; the two papers in The BMJ did. They were entitled to cite the original work but not to misrepresent it. Most people would not deem this a hanging offence, but Collins’s vivid language made it seem worse, and Godlee’s initial slip in claiming that the error had been missed by peer reviewers further riled him. In fact, the statistic had been added to Abramson’s paper in proof; the reviewers were blameless.

“Not a reasonable matter for debate”

In real life, many people who take statins have pains that they attribute to the drugs. Are they deluding themselves by inferring causation where none exists? That is the view of Collins and of the NICE panel—their argument is that in the clinical trials just as many people taking a placebo had joint and muscle pain as in the active drug group. The reason Collins gave for not submitting the response Godlee wanted was, in a letter marked “not for publication”: “This is not a reasonable matter for debate, but is instead one of fact: a ‘statin-related adverse event’ (as studied by Zhang et al) is not necessarily caused by, or a side-effect of, a statin so it is just plain wrong to claim that it is.”

It’s fair to say, however, that this is a distinction that might be lost on Mr Lowrisk.

Collins’s response to the report of the Heath panel was equally dismissive. Conceding that the major error had been corrected, he added that several other serious errors had not yet been withdrawn—despite the finding by the panel that all the other numerical claims made in the Abramson paper were statistically sound.

McPherson said (in an email to Collins) that he found this response “quite shocking.” Other Oxford colleagues, including Iain Chalmers, the founder of the Cochrane Collaboration, have tried to broker peace between The BMJ and Collins but without success.

Calls have been made for all the patient level data held by the CTT to be released and (by Chalmers) for the statins to be reassessed in a withdrawal trial, where people who complain of side effects are randomly allocated to receive either placebo or to continue the active drug and the outcomes measured. For the moment, however, it is plain that trying to close down this particular debate by asserting superior knowledge is unlikely to succeed.

McPherson says of the situation, “My view is that it’s absolutely unsatisfactory as it stands. As more and more people go on to statins, the side effects will become more and more manifest, if they’re real, and we’ll get to the point where they won’t be the drug of choice.”

That, of course, would be the reverse of the objective sought by Collins.

Notes

Cite this as: BMJ 2014;349:g5007

Footnotes

  • Competing interests: See bmj.com/about-bmj/freelance-contributors/nigel-hawkes.

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

  • thebmj.com Editorial: Statins and The BMJ (BMJ 2014;349:g5038, doi:10.1136/bmj.g5038)

  • The independent panel’s report and all related materials, including a timeline of events leading up to the report, are at thebmj.com/statins.

References

View Abstract