Statins were the 20th century ‘blockbuster’ drug. Almost everyone could benefit from turning a Euro-American cholesterol into a Japanese one according to their supporters … And now they are off-patent, statins look better and better value to those seeing health policy through the lens of cost effectiveness. But precisely because of their prominence, debates about the goods and bads of statins never go away. After the most recent battles in the ‘statin wars’, the BMJ has tried to close down the controversy with a ‘State of the Art Review’ (published last week alongside the latest NICE Guidelines on Lipid Modification, which proposed lowering the threshold for statin treatment).
From outside the profession, this attempt to end the debate seems unlikely to work. For example, the BMJ review addresses a wide range of ‘unanticipated consequences’ of statins, set against their ability to reduce cardiovascular risk, which roughly means that they may save two people a heart attack or stroke from 100 taking them, though 8 still suffer an event (see this animation by David Spiegelhalter).
Personally I prefer not to make a general statement about whether this is ‘large’ or ‘small’ effect. People may well wish to take statins for this level of (potential) benefit, and at a population level it appears to be accepted as cost effective, which is the current criteria for NICE. However if people are being offered these drugs they deserve to be informed about the potential bads of statins as well as the potential goods.
Yet, there is still considerable uncertainty about the risks as Ben Goldacre points out with some frustration. The BMJ review summarises reports from randomized controlled trials, meta-analyses, systematic review and even the odd observational study, but struggles to find consistent rates of claimed side effects or adverse events that could be meaningfully presented to a patient. Instead it concludes with questions for further research. The picture that emerges is complex and patchy. The authors list the findings and circumstances of each report, sometimes in favour of statins, sometimes against, like layering oil paints one upon the other, without a single image emerging. The review is full of gestures of epistemic modesty – a classic tactic to manage accusations of enthusiasm in medicine. Furthermore, it skirts around the question of statin risks, by addressing ‘non-cardiovascular effects of statins’ enabling a focus on potential positive as well as negative unanticipated consequences. Possible negatives include muscle pain, diabetes, liver toxicity, cataracts, cancer, fatigue and dementia. Possible positives include reductions in venous thromboembolism, kidney injury, erectile dysfunction and COPD. But surely it is the potential negative effects that need clarifying now? If a doctor is offering a drug to someone, they need to know what ‘harms’ to balance against the intended good (especially as both are uncertain and in the future). From the clinical perspective I think there is a difference between sins by commission and omission. Individual patients and doctors need clear information about the likelihood of the ‘bads’ of statin use and their reversibility.
Medically, ‘side effect’ is used to delimit negative effects that may be deemed to be an acceptable risk for a given chance of benefit. It is usually used for mild to moderate unwanted effects of drugs, while ‘adverse events’ is used for a serious reaction. We could debate the boundaries of the ‘serious’ or ‘mild’ according to our own preferences, or drawing on research data about patient preferences and experiences. However, such data is rarely incorporated into reviews of evidence before decisions about licensing for sale or reimbursement.
Sociologically, people not taking drugs in trials might be simply expressing a general dislike of medicines – what sociologists have described as ‘resistance to medication’ – though they would seem on the face of it to be less resistant than your average patient. They might also be responding to real side effects from statins. But both general and specific dislike is ‘data’ about a broader set of possible harms. When you ask people prescribed statins how they feel about them, some say it makes them feel as if they are ill. Their sense of self is disrupted by the prescription. They may worry that stomach trouble, aches and pains are signs that the drug is doing them harm. Given the recent high profile expert debates, that wouldn’t be surprising. Though trialists might not see this as their business, in practice doctors should be discussing this with their patients.
The BMJ review authors suggest that such discussions are desirable. Just as the NICE guidelines insist that ‘patients should have the opportunity to make informed decisions about their care and treatment, in partnership with their healthcare professional,’ the review concludes somewhat bizarrely given its limits that there should be ‘renewed emphasis on individualised cardiovascular risk assessment [that] promotes an evidence based discussion of the risks and benefits of statin use between patients and clinicians.’ This is all in keeping with Trisha Greenhalgh’s recent insistence that ‘real evidence based medicine has the case of individual patients as its top priority.’ But there is a great distance between the contents of both the published review and guideline and these visions of how evidence might be used in the clinic. In the meantime, I don’t see why tools attempting to show patients the benefits of statins could not attempt to put in a fuller range of possible harms (including side effects and adverse events) and even – quite simply – acknowledge that people don’t like taking pills by changing some of those smiley faces to something more sombre. While I’m manifestly not qualified to do this definitively, my suggestion would look something like this…