Where’s the evidence that treating younger women with statins will do anyone any good?

Back in February, the National Institute for Health and Care Excellence (NICE) published draft proposals concerning the use of statins for the prevention of cardiovascular disease. Previously, NICE has advised doctors to recommend treatment in those with a calculated 10-year risk of cardiovascular disease of 20 per cent or more. Now, NICE is proposing that this threshold be reduced to 10 per cent.

As I explained here though, what the evidence shows is that the vast majority of people at relatively low risk of cardiovascular disease are extremely unlikely to benefit from statins, while risk of side-effects (including diabetes, liver damage, kidney damage, fatigue and muscle pain) are relatively high (around 20 per cent). There’s a valid argument that in low-risk individuals, the harms of stains outweigh the benefits, and I think that NICE’s proposals are utterly misguided.

With this background I was interested to read this piece in the Daily Mail earlier this week. It informs me that health charities (including the British Heart Foundation) are proposing that more women, including those in the 40s, be treated with statins. You see, according to the head of the British Heart Foundation, Professor Peter Weissberg, while the 10-year risk of some women “with stonkingly high cholesterol, very high blood pressure and who [smoke] is still fairly low”, “…her lifetime risk is huge.” The implication is that such women should be started on statins nice and early.

When pronouncements such as these are made, I sometimes like to see what evidence there is to support them. So, do we have studies that show that early treatment of women with statins leads to meaningful improvements in health outcomes?

The treating of individuals with no history of cardiovascular disease (such as a previous heart attack or stroke) with statins is an example of what is termed ‘primary prevention’. A previous review of six primary intervention studies found that reducing cholesterol did not reduce the risk of heart disease related events (e.g. heart attacks), death from heart disease or overall risk of death [1]. This evidence does cast some doubt on the wisdom of treating women with no established cardiovascular disease with statins.

But, I suppose, maybe Professor Weissberg will not care much about the research. After all, he is the person to steadfastly refused to engage with the evidence that suggests that cholesterol-reducing sterols (which the BHF recommends) may well have adverse effects on health (you can read more about this here). He is also the person who has previously used evidence to support the use of the statins that simply does not support the use of statins (see here for more about this).

This article also included input from Professor Rory Collins, who tells us that for those treated earlier “[their] potential benefit over the longer term is greater.” Potential is one thing, but how about providing the evidence that demonstrates this potential is realised in terms of actual benefits? ‘Experts’ and ‘scientists’ can pronounce all they like, but should their ideas not be based on actual evidence rather than assumptions?

Perhaps, though, we should not be too surprised that Professor Collins appears to play fast and loose with the evidence. After all, he is one of the research group (the Cholesterol Treatment Trialists) that regularly makes very positive noises about statins and reassurances about their safety based on data that is held close and no-one else is allowed to see. See here for more about this.

There’s a saying that “Extraordinary claims require extraordinary evidence.” I reckon the only extraordinary thing about the likes of Professors Weissberg and Collins is the just how often they appear (perhaps unwittingly) to rely on rhetoric rather than solid, verifiable research.

References:

1. Walsh JME, et al. Drug treatment of hyperlipidemia in women. JAMA 2004;291:2243-52

21 Responses to Where’s the evidence that treating younger women with statins will do anyone any good?

  1. Sally Wickenden 25 April 2014 at 1:49 pm #

    What is more worrying about prescribing statins to younger women is the risk of pregnancy.

    • Stephen Rhodes 25 April 2014 at 7:14 pm #

      Do you mean the side effects noted on the included leaflet that specifically proscribes their use in women of childbearing age?

      As in this

      ‘You should not take a statin if you have active liver disease, if you are are pregnant or intend to be pregnant, or if you are breast-feeding. You should stop a statin if you develop liver disease.’

      from http://www.patient.co.uk/health/statins-and-other-lipid-lowering-medicines

      Studies have been commissioned – presumably by the statin manufacturers – to disprove the teratogenic effects of statins. But you would have to be an idiot to accept their weaselly conclusions.

  2. Mark 25 April 2014 at 2:57 pm #

    Where’s the evidence that the risk of side effects is about 20%?

    • Stephen Rhodes 26 April 2014 at 1:05 pm #

      http://www.medscape.com/viewarticle/758377_2

      >18% close enough for you?

      “The voluntary withdrawal of cerivastatin from the worldwide market in 2001, mainly due to reports of fatalities attributed to drug-related rhabdomyolysis and subsequent renal failure, has raised concerns about the safety of statins as a of drugs, especially their muscle-related side effects. Myalgia is the most common side effect reported by patients receiving statins. Incidence of myalgia ranges from 18% in clinical practice [Joy and Hegele, 2009]. The clinical presentation of myalgia can be subtle or nonspecific and clinicians may have difficulty in deciding whether myalgia is statin-induced or caused by other musculoskeletal conditions commonly seen in the elderly. In comparison, myositis is much less prevalent but comparison of incidence rates among studies proves to be difficult due to the inconsistent definition of myositis. The clinical features, underlying mechanisms and management of statin-related muscle toxicity have already been addressed by a number of reviews [Joy and Hegele, 2009; Sathasivam and Lecky, 2008; Thompson et al. 2003] and will not be repeated here. Many risk factors, including advanced age, have been found to be associated with statin-related muscle toxicity (Table 2). High doses of statins are known to increase the risk of muscle toxicity; both the MHRA in the UK [Medicines and Healthcare Products Regulatory Agency, 2010] and the FDA in the USA [Food and Drug Administration, 2011] have recently issued recommendations on the use of high-dose (80 mg) simvastatin in relation to muscle toxicity. The prevalence of hypothyroidism was reported to increase with age [Empson et al. 2007] and undiagnosed hypothyroidism is also a risk factor for severe statin-induced muscle toxicity such as rhabdomyolysis. Concomitant administration of medications that may interfere with the pharmacokinetic profiles of statins can also increase the risk of potential muscle toxicity due to statin therapy (see below).”

      • Mark 28 April 2014 at 8:46 am #

        No, >18% isn’t close enough at all. You’ve shown much bias in what you’ve done:

        1) You’ve cherry-picked the type of evidence – choosing one quote from an unsystematic review, while ignoring systematic reviews of thousands of patients e.g. the recent Cochrane review:

        http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004816.pub5/abstract

        which reports:
        Myalgia and rhabdomyolysis: 3551 of 37,939 participants in nine trials developed myalgia, but there was no evidence of excess risk [compared with patients receiving control] with a pooled estimate [risk ratio] of 1.03 (95% CI 0.97 to 1.09] with some heterogeneity (I2 = 41%).

        2) You’ve cherry-picked the outcome, only referring to myalgia.

        3) To cap it all you even cherry-picked within the sentence. The full sentence actually reads: “Incidence of myalgia ranges from 18% in clinical practice”.

        So, you presented a thoroughly biased assessment.

        • Mark 28 April 2014 at 8:53 am #

          Point 3 should read “Incidence of myalgia ranges from 18% in clinical practice”. I’ll forgive you point 3, as there’s clearly a copy and paste issue with that sentence when it gets submitted as a comment!

          • Mark 28 April 2014 at 8:56 am #

            I’ll try for a final time. I think there’s an issue with greater than signs (or is this site just allergic to term randomized trials!):

            “Incidence of myalgia ranges from less than 5% in randomized trials to greater than 18% in clinical practice”.

        • Dr John Briffa 29 April 2014 at 6:00 am #

          Mark

          If you’re relying on randomised controlled trials to provide all the answers, you are perhaps missing the fact that the data here is far from complete and not without a sort of bias of its own.

          See this post: http://www.drbriffa.com/2014/03/13/researchers-pronouncing-statins-are-safe-are-undermined-by-their-own-observations/ for just some reasons why the ‘official line’ on statin adverse effects is not to be trusted.

          • Mark 29 April 2014 at 8:41 am #

            Thanks for replying. It’s certainly true that results from some RCTs may often not be generalisable to broader populations of patients, so they can’t provide all the answers. However, they do normally provide the strongest, most reliable evidence (when conducted properly).

            The thing I object to, is that you stated that the risk of side effects from statins is around 20%, as if that’s based on strong evidence. One major problem with using clinical experience as evidence in this area is that you can’t eliminate the nocebo effect; you can with RCTs.
            http://jama.jamanetwork.com/article.aspx?articleid=1104968

            The authors of the study on your previous blog at least listed the possible limitations of their results – you should also qualify your statements, especially when based mainly on clinical experience e.g. “Although the side effects of statins are reportedly very low in randomised trials, my experience suggests the rate could be around 20%”. Statements such as this would be far less misleading for patients reading these blogs – you shouldn’t just ignore the large meta-analyses.

  3. john mason 25 April 2014 at 3:30 pm #

    Lowering cholesterol does not reduce the risk of heart disease, so Statins may benefit as an anti inflammatory? So does aspirin.

    When I questioned my Lipidologist regarding the very low NNT of Statins, he replied that Statins are prescribed in volume for the community benefit, rather than for an individual’s benefit.

    So the greater the population on Statins, then, in absolute terms the numbers may look impressive.

  4. Maureen Berry 25 April 2014 at 3:52 pm #

    I am less concerned about the side effects of statins than I am about the effects of statins. Who knows what the effect will be of putting half the population on drugs, for decades, that have clearly not been tested, or assessed for the dangers, in people who have taken them for decades. Surely, if Cholesterol was so harmful to the human body then virtually every cell in the human body would not be able to manufacture it. I am convinced that, since the human brain is largely constructed of cholesterol, taking any medication to aggressively reduce the level of cholesterol in the bloodstream will have the potential of harm to the brain. It only seems to be common sense. The cholesterol drive has all been based on smoke and mirrors, poor research and poor evidence – more to do with drug company profits than promoting good health. There are no circumstances where I would consider taking these dreadful drugs.

  5. Stephen Rhodes 25 April 2014 at 7:15 pm #

    Please keep poking Rory Collins with your literary ‘pointy stick’ – he may eventually provide the ‘evidence’

  6. Jenny B 25 April 2014 at 8:07 pm #

    At 31, I was told my cholesterol was high and that I needed to go on statins right away. I asked for how long and was told for the rest of my life! I asked if i could improve it with diet and exercise, and was told “not likely” by my doctor. The more questions I asked the more angry my doctor became. When I told him I wouldn’t take them, he looked at me like i was crazy! I had no other risk factors, was not over weight and had a healthy heart. I changed my diet only slightly, and within 1 month my cholesterol levels were down! But hey, why bother with diet and exercise when you can just take poison for the rest of your life!

  7. Dave P 26 April 2014 at 10:32 am #

    As the man once said ” if the porky is big enough and told often enough eventually it will become the accepted truth ”

    It would seem the Proff knows more about Media impact than Medicine.

  8. Lindy Loo 27 April 2014 at 1:24 pm #

    Dave,

    Who was ‘the man’ who once said about porkies?

  9. Dr John Briffa 29 April 2014 at 9:51 am #

    Mark

    “Thanks for replying. It’s certainly true that results from some RCTs may often not be generalisable to broader populations of patients, so they can’t provide all the answers. However, they do normally provide the strongest, most reliable evidence (when conducted properly).”

    The problem is, Mark, is that the RCTs in this area have not been ‘properly conducted’, have they: There are gaping holes in the data, and quite some evidence of bias in terms of how studies have been conducted (and reported).

    “The thing I object to, is that you stated that the risk of side effects from statins is around 20%, as if that’s based on strong evidence.”

    Actually, I think you’ve misrepresented me here. My claim is not nearly as strong as you suggest, though I do utterly stand by it.

    “One major problem with using clinical experience as evidence in this area is that you can’t eliminate the nocebo effect; you can with RCTs.”

    As we know, the RCT data is not always to be relied on. But, anyhow, here’s a thought experiment for you, Mark: A patient comes to see you and tells you, “About a month after starting statins I started to feel unusually tired and achy. I thought it might be the statins, so I stopped and the symptoms went away over a couple of weeks. When I resumed, the symptoms came back again. I think maybe the statins are not suiting me.”

    In this situation, how do you know if the patient’s experiences are due to the nocebo response or not?

    If you were a doctor (assuming you are not), how would you manage and advise this patient, do you imagine?

    • Mark 29 April 2014 at 1:10 pm #

      “The problem is, Mark, is that the RCTs in this area have not been ‘properly conducted’, have they: There are gaping holes in the data, and quite some evidence of bias in terms of how studies have been conducted (and reported).”

      I think it’s always useful to question evidence, but I think your doubts with respect to statins are too strong. The methodological quality of statin trials, particularly the large ones, is generally acknowledged as being very good, e.g. see the risk of bias assessment results in this Cochrane review:

      http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004816.pub5/figures

      Or the quality assessment results in this review (see appendix, p11):

      http://www.sciencedirect.com/science/article/pii/S0735109711049898

      It’s true that such impressive ‘internal validity’ isn’t everything. However, there’s an important distinction between bias within a trial, and (lack of) generalisability to other populations. People sometimes mistakenly interpret limited applicability (of trial results) to other populations as investigator bias. It’s not – most trials clearly state the populations to which their results relate to.

      Another key issue to consider when thinking about proper conduct of trials is conflicts of interest. The above paper makes the point that: “In the majority of the studies, an independent data center was responsible for the data and analysis.” I guess a lot of this debate comes down to a question of trust – if you don’t trust the independent data centres then there’s nothing more I can really add to convince you, other than to say that in many other areas of research the trials aren’t conducted independently (e.g. anti-virals for flu). Personally, I think that all trials which are industry-funded should be conducted by independent organisations.

      “Actually, I think you’ve misrepresented me here. My claim is not nearly as strong as you suggest, though I do utterly stand by it.”

      That’s fine, but I think a patient taking statins would interpret what you wrote in a similar way as I did. Nevertheless there’s an obvious irony that, in writing an article bemoaning lack of evidence, you haven’t provided robust evidence yourself. Without the randomised trial evidence, we would not know that statins cause a slight increase in the risk of developing diabetes. For anyone interested: treatment of 255 patients with statins for 4 years resulted in one extra case of diabetes.

      http://www.sciencedirect.com/science/article/pii/S0140673609619656

      “….In this situation, how do you know if the patient’s experiences are due to the nocebo response or not? If you were a doctor (assuming you are not), how would you manage and advise this patient, do you imagine?”

      I think your final question is less about statins, and more about issues regarding how much advice clinicians should give about possible side effects: too little information would be unethical, too much caution could result in excessive nocebo effects. But my answer would be to check their CK levels (assuming you had a pre-statin measurement) and to try reducing the dose (if possible). Ultimately though it’s the patient’s decision about whether to stop the statin. They must make a judgement on the trade-off between their level of discomfort and the likely benefit for them as an individual.

      • Stephen Rhodes 1 May 2014 at 7:33 pm #

        “But my answer would be to check their CK levels”

        Is that the same as what is referred to as CPK levels in the US?

        If so it is not necessarily a good enough check.

        See http://www.webmd.com/cholesterol-management/news/20090706/muscle-damage-from-statins-may-evade-blood-test

        Quote: July 6, 2009 – Statin users with prolonged statin-related muscle pain may also experience muscle damage, even when a blood test used to identify muscle injury is normal, new research shows.

        Studies suggest that between 10% and 15% of patients who take cholesterol-lowering statin drugs like Crestor, Lipitor, Lescol, Mevacor, Zocor, and Pravachol experience muscle pain as a side effect of treatment.

        Most do not end up with muscle damage, and a simple blood test is routinely performed to identify patients who do.

        But the new study suggests the test for elevated levels of an enzyme associated with muscle injury, known as creatine phosphokinase or CPK, may be less accurate than widely believed.

        “The patients in our study were unusual in that they had experienced weeks to months of persistent muscle problems,” Richard H. Karas, MD, PhD, tells WebMD. “We found that these patients can have evidence of microscopic damage to their muscles even with a normal CPK.”

        Perhaps you would care to explain what is your interest in the mass medication of the populace with statins?

        Could you perhaps clear up the issue of which side effects are beneficial and which are not, considering that the stated purpose of Statins was/is to reduce cholesterol levels, something that has absolutely nothing to do with CVD or stroke prevention.

      • Z.M. 2 May 2014 at 2:01 pm #

        Mark: “The methodological quality of statin trials, particularly the large ones, is generally acknowledged as being very good”

        Hmm…not true at all. Most statin trials are poorly designed and conducted.

        For example, in the Cochrane review you cited, a bunch of small trials that were actually designed as surrogate marker trials (with methodological problems) were included, three large trials were stopped early for no good reason (which is usually a BS ethical argument) with allocation concealment being unclear in one, and another used an inappropriate design and wasn’t even blinded (MEGA). Of course, conflicts of interest and lack of transparency also doesn’t help.
        .

  10. NM 1 May 2014 at 3:24 pm #

    Relying on the nocebo crutch to this degree morphs into something tastelessly akin to victim-blaming.

    When one dispassionately analyses the biases in the data and their collection, one errs toward the presumption of under-reporting. One cannot honestly prod wildly at a usefully-figmented flock of Munchausen hysterics.

    Further, I’ve not had an answer to this point: throw a spanner into the Mevalonate pathway, and all research indicates and *predicts* a high frequency of the very side effects we have seen reported. Statins, of course, throw just such a massive spanner into the Mevalonate pathway as their primary mechanism.

    It is thus beholden on those who would claim general benignity to suggest the novel factor which would *mitigate* against the such mechanistically-predicted side-effects. The near-inevitability of side-effects should be the null hypothesis here!

    The appropriate sceptical question here should therefore NOT be “why would statins cause such a high rate of side effects” but, rather, “why wouldn’t they?”

Trackbacks/Pingbacks

  1. Should statin ‘key opinion leaders’ be allowed to just make stuff up? | Dr Briffa's Blog - A Good Look at Good Health - 8 May 2014

    […] 2 weeks ago, one of my blog posts was dedicated to assessing the call from certain ‘experts’ that more  younger women should be […]

Leave a Reply