The ‘Cochrane Collaboration’ is an international collective of researchers whose self-proclaimed role is to provide accurate and robust assessments of health interventions. The group specialises in ‘meta-analyses’: the grouping together of several similar studies on interventions including drug therapies.
In 2011, Cochrane researchers assessed the evidence relating to statin use in individuals at low risk of cardiovascular disease (defined as a less than 20 per cent risk over 10 years), and concluded that there was limited evidence of overall benefit [1].
Then in 2013, the same Cochrane group updated their data and concluded thatoverall risk of death and cardiovascular events (e.g. heart attack or stroke) were reduced by statins in low risk individuals, without increasing the risk of adverse events (including muscle, liver and kidney damage) [2]. It seems the Cochrane reviewers had had quite some change of heart. A paper published in the BMJ in October 2013 questions the evidence on which this U-turn appeared to have been made [3].
The authors of the BMJ piece note that although the 2013 meta-analysis included four additional trials, these trials did not substantially change the findings. The change in advice was actually based on another meta-analysis, published in 2012, conducted by a group known as the Cholesterol Treatment Trialists’ (CTT) collaboration [4].
Among other things, the CTT authors concluded that, in low risk individuals, for each 1.0 mmol/l (39 mg/dl) reduction in LDL-cholesterol, statins reduce overall risk of death and heart attack by about 9 per cent and 20 per cent respectively. The conclusion was that statins have significant benefits in low risk individuals that greatly exceeded known risks of treatment.
However, the CTT authors took the odd step of calculating the benefits of statins according to a theoretical reduction in LDL-cholesterol levels. A much more relevant appraisal would be simply to calculate if, compared to placebo, statins actually reduce the risk of health outcomes.
The BMJ authors used the data from the CTT meta-analysis and found thatrisk of death was not reduced by statins at all. So, the CTT authors had extrapolated the data in a way that showed a benefit that actually does not exist in reality.
They also draw our attention to the impact of statin treatment on ‘serious adverse events’. This outcome can be improved by statins as a result of, say, a reduced number of heart attacks, but worsened through side effects such as muscle or liver damage. The BMJ authors note that the CTT review did not consider serious adverse events (a major omission).
Without knowing more about this, though, we simply cannot make a judgement regarding the overall effect of statins, and whether the net effect is beneficial or not. Interestingly, of three major trials that were included in the CTT review that assessed overall serious adverse effects, none found overall benefits from statin treatment.
So, while the CTT authors seem to have over-hyped the benefits of statins, they seem at the same time to have been quite keen to steer clear of talk of their very real risks and the absence of evidence foroverall benefit.
The BMJ authors draw our attention to the fact that every single trial included in the CTT was industry funded. Such trials are well known to report results more favourably and perhaps downplay risks than independently funded research. The BMJ authors cite specific ways in which the adverse effects of drugs seen in clinical trials can be ‘minimised’. These include:
- The exclusion of individuals from trials with known health issues likely to be exacerbated by statins or signal susceptibility to statin side effects (such as liver, kidney and muscle disease).
- The use of a ‘run-in’ period before the study starts which detects and then excludes individuals who do not tolerate statins.
- The possibility that individuals ‘drop out’ from the study because of side effects, meaning that the incidence of some side effects can be ‘lost’ from the data.
- Failure of the study investigators to assess and monitor adverse events such as muscle damage and changes in brain function.
- Failure to properly ascertain or report adverse events.
It is noted that the Cochrane authors admit the reporting of adverse effects in studies is generally poor, but also state that it’s unlikely statins have major life-threatening hazards. The authors of the BMJ piece were not convinced, though, writing: “[The] large discrepancies between the frequency of adverse events reported in commercially funded randomised controlled trials included in the CTT meta-analysesand non-commercially funded studies show that determination of harms cannot be left to industry alone.”
The BMJ piece is accompanied by an editorial from the journal’s editor, Fiona Godlee [5], in which she writes
There is a concern underlying their critique that will be familiar to BMJ readers. It is that all of the trials included in the CTT meta-analysis were funded by the manufacturer of the statin being studied. They list the various ways in which these trials might have exaggerated the benefits of statins and minimised the harms, and they summarise what low risk patients need to know. Top of the list is the benefit of lifestyle change, something that the dominance ofindustry sponsored clinical trials too often obscures.
At one point, Fiona Godlee remarks:
None of this does much to bolster confidence in the published literature.
And now a curious thing has happened. On 27th January, the BMJ published an article co-authored by two of the original Cochrane authors (along with a general practitioner) [6]. The piece examines the risk of a hypothetical male patient, and the benefits and risks to him of taking a statin. The piece is relentlessly positive in its appraisal of the benefits of statins, and does much to attempt to allay any fears we may have about risks.
Here’s some excerpts (I have removed references for ease):
Adverse events and stopping treatment because of [statins] are common, but occur at similar rates in treated and control groups in clinical trials (17% and 12% respectively), making it difficult to ascribe these events to statins.
Reports of reduced energy, fatigue, depressed mood, and reduced quality of life show inconsistent findings.
Myositis (a ≥10-fold rise in normal levels of creatine kinase) is caused by statins, but myalgia (muscle pain without raised creatine kinase) is not linked with statin use.
The authors conclude:
Patients may expect not to be harmed in any way by preventive treatment with statins, and their views of trade-offs between benefit and possible harms will likely determine the wider use of statins.
I am not the only person to have noticed what seems to be quite a biased tone in the latest BMJ piece. Here’s a response from three New Zealand doctors who articulate well the issues [7] (as above, I have removed the scientific references for ease).
Enthusiasm for the use of statins comes through strongly in the paper by Ebrahim et al. It reads more as a “YES” in a head to head piece about the use of statins and does not convey the ongoing controversy around many aspects of when and how to use statins for primary prevention.
We note a number of references to the National Institute for Health and Care Excellence guidance. More recent American guidelines now recognise the absence of evidence to support any particular arbitrary LDL target in primary prevention. The authors’ suggestion of using ezetimibe in conjunction with a statin when LDL is not controlled is therefore redundant. More importantly, there is no evidence that ezetimibe improves meaningful clinical outcomes either with or without statins in any population.
We were also surprised at the authors’ presentation of statins’ adverse effects on muscle. Their statement that “Myositis (a >10-fold rise in normal levels of creatine kinase [CK]) is caused by statins, but myalgia (muscle pain without raised CK) is not linked with statin use” seems to minimise this adverse effect. For a patient who develops muscle pain whilst taking a statin, whether their CK is raised beyond an arbitrary threshold is unimportant, the question is whether or not the pain is linked to their medication. This adverse effect is well-established and has been described in more recent studies than the quoted 2008 review which in fact did not conclude that statins do not cause muscle pain without raised CK. The distinction between myositis and myalgia in this context is thus obfuscatory. There is ample evidence from cohort, adverse events report data, and case reports to suggest myalgia without enzyme alteration is a problem, and is not rare.
A person (in this context they aren’t patients until we start labelling them as high risk) considering whether to take a statin for primary prevention needs to understand their estimated cardiovascular risk, and be given a personalised estimate of the magnitude of the benefits and harms of the available risk-reducing non-pharmacological and pharmacological interventions. Our role is to provide appropriate and personalised advice on the range and size of possible benefits and harms for all interventions (lifestyle and medications) to inform these shared decisions. (10) A shared decision making approach facilitates informed choices based onassessment, understanding and prioritisation of benefits and harms. This approach may particularly benefit those penalised by previous threshold guidelines: young people with low 5 or 10 year, but high lifetime, combinedrisk can opt for earlier treatment, and conversely some healthy elderly people (whose main risk factor is simply age) may choose to opt out of rigid guideline-driven unilateral treatment decisions. None should face pressure by practitioners fearing professional or financial censure for not following a one size fits all cut-off approach to treatment. These principles have been adopted in advice recently released by the New Zealand Ministry of Health (11) which includes a recommendation to adopt a shared decision making model.
Given that the overwhelming majority of those who take statins for primary prevention will not derive any benefit from them, it is crucial that decisions to initiate lifetime therapy are informed by a clear presentation and discussion of the best available evidence.
Remember, this most recent pro-statin piece [6] appears in the BMJ, which only a few weeks before, carried a piece which highlighted the potential bias in the statin literature and the problems with how statin studies are often undertaken and reported. This most recent piece gives plenty of soothing reassurances about the safety of statins, and yet back in October the editor of the BMJ was alluding to the potential problems with relying on industry-derived data. Is the editor of the BMJ suffering from a sudden bout of statin-induced amnesia, I wonder?
References:
1. Taylor F, et al. Statins for the primary prevention of cardiovascular disease. Cochrane Database Syst Rev 2011;1:CD004816.
2. Taylor F, et al. Statins for the primary prevention of cardiovascular disease. Cochrane Database Syst Rev2013;1:CD004816.
3. Abramson JD, et al. Should people at low risk of cardiovascular disease take a statin?
BMJ 2013;347:f6123
4. Cholesterol Treatment Trialists’ (CTT) Collaborators. The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials. The Lancet 2012;380(9841):581–90.
5. Godlee F. Statins for all over 50? No BMJ 2013;347:f6412.
6. Ebrahim S, et al. Statins for the primary prevention of cardiovascular disease
BMJ 2014;348:g280
7. Hudson B, et al. Rapid response published 30 January 2014
