Aspirin: don’t believe the hype. It’s still not a miracle drug

From yesterdays PM, Radio 4: Professor Peter Elwood, epidemiologist at Cardiff University:

“I have always held that it is for the individual himself or herself to decide whether or not he or she would  take aspirin – they should be told the risks or the benefits and it should be the patients or subjects value that they put  on  on the possible outcomes . The way I see it myself is that  a heart attack or stroke cause disability,  have a very high risk of death, the downside is a bleed that at worst is going to lead to me being rushed into hospital for a blood transfusion but no sequalea, no after effects – I find it easy to value those, and to take aspirin. Some have exaggerated the risk of a bleed and one has to recognise that doctors practice defensive medicine but I think that the individual should make that decision.”

He’s talking about a study published over at the Lancet, where some esteemed people have analysed the deaths due to cancer,  in trials where people were being randomised to treatment with aspirin or not.

We already know that aspirin is of definite benefit for people at high risk of cardiovascular disease – namely those people who have already had a heart attack or stroke. But preventing the first event is rather different. Doctors used to recommend that people with diabetes  but who hadn’t previously had a heart attack or stroke  (but who nevertheless were at higher risk because of the diabetes)  took aspirin as a precaution. But then a study came along that said it made no difference to mortality, and u-turns were made.

For this is the problem with aspirin. It comes with side effects. Elwood as quoted on Radio 4 today is incorrect. People can die from the side effects of aspirin as used therapeutically. Bleeds into the brain or into the gut are the two most serious side effects. One may survive these. But they may be complicated by a pneumonia or a fractured hip and an individual may not recover. This means that we need to know what the all-cause mortality is – not just the effects of aspirin on the chosen parameter (eg, cancer mortality) but everything else too. There’s no point saving a cancer death if you precede yourself with another, potentially aspirin linked cause.

So, to  the Lancet study. We already know that aspirin is capable of preventing certain cancers. This study confirms that for people taking aspirin for 4 years or longer, there was a reduction in certain cancers. But the crunch is – when you look at the all-cause mortality, the results are not as impressive as the media reports have made out.

The graphs in the Lancet paper show that the reduction in cancer deaths achieved with aspirin is of the order of magnitude of a couple of percent off whatever the normal prevelance is. This is important when it comes to population measures to prevent cancer. But it has to be considered in terms of other population measures too – weight loss, for example, would reduce cancer rates, and so to would more exercise. Aspirin has side effects: for example, one study describing the use of aspirin for secondary prevention described it like this: ” aspirin treatment for a mean of 6.4 years resulted in an average absolute benefit of around 3 cardiovascular events prevented per 1000 women and 4 cardiovascular events prevented per 1000 men. This was offset by an additional 2.5 major bleeding events per 1000 women and 3 major bleeding events per 1000 men”.  Or try this, from the US Task Force on Preventive Medicine : “For 1000 patients with a 5% risk for coronary heart disease events over 5 years, aspirin would prevent 6 to 20 myocardial infarctions but would cause 0 to 2 hemorrhagic strokes and 2 to 4 major gastrointestinal bleeding events. For patients with a risk of 1% over 5 years, aspirin would prevent 1 to 4 myocardial infarctions but would cause 0 to 2 hemorrhagic strokes and 2 to 4 major gastrointestinal bleeding events.”

From the paper itself, I quote:

“The reduction in cancer deaths on aspirin during the trials resulted in lowered in-trial all-cause mortality (10·2% vs 11·1%, OR 0·92, 0·85–1·00, p=0·047, webappendix p 4), even though other deaths were not reduced (0·98, 0·89–1·07, p=0·63” (Page 4)
“In patients with scheduled duration of trial treatment of 5 years or longer, all-cause mortality was reduced at 15 years’ follow-up (HR 0·92, 0·86–0·99, p=0·03), due entirely to fewer cancer deaths, but this effect was no longer seen at 20 years (0·96, 0·90–1·02, p=0·37).”
So there is a small reduction in all-cause mortality; but this disappears at 20 years. Additionally, the researchers report a 40% drop out rate for the aspirin taking groups, indicating that there may have been side effects reported very commonly – and which has implications if one wants to know the likely uptake in a large population.

So aspirin isn’t a wonder drug, and the nuances of taking it are more complex than portrayed on the radio.  At the very heart of medicine is explanation of the possible courses  of interventions and assistance to make sense of the positives and negatives – and explain the uncertainties. The worst that can happen is not a little bleed of no consequence.  I’m keen to ensure that my patients have access to good unbiased information. There have been numerous researchers in the press over the last two days revealing that they are not taking aspirin as part of their endorsement of their studies. But they have that wrong, too. The studies that the Lancet have presented were not originally designed to assess whether or not cancer was reduced when taking aspirin. We still don’t know enough about the risks and benefits in a general, unselected population. The way forward is a better trial, not a presumption that doctor knows best.

8 Responses to “Aspirin: don’t believe the hype. It’s still not a miracle drug”

  1. Anne Marie Cunningham December 9, 2010 at 4:54 pm #

    We certainly do need better trial evidence. I’ve seen Peter Elwoon talk aboutr aspirin many times in Cardiff. I don’t think he is suggesting at all that doctor knows best- in fact he’s suggesting that every person has to weigh up the risks for themselves at the moment on existing evidence. Doctors may worry more about causing harm to a patient- defensive medicine- than the potential benefits. It’s a hard call and I look forward to the emergence of some good decision aids so that we can explain the risks and benefits as we know them now.

  2. Margaret McCartney
    margaretmccartney December 10, 2010 at 2:22 pm #

    please do listen again Anne Marie. He’s suggesting that the side effects of aspirin are always minor. Doctors have to give unbiased evidence about interventions , the pros and the cons. I don’t find it helpful, as a clinician, to be biased in either direction; one has to be absolutely clear that a doctor should give disinterested information. I have changed my practice many times on the basis of what the evidence tells us. For years I’ve been trying to do this with screening tests, which it usually appears that to encourage active participation in decision making is to encourage dissent – not true, but it’s how it appears.
    This isn’t to me about ‘defensive medicine’ – it’s about accurate information and fair and unbiased perspective on risk.

  3. Derek Tunnicliffe December 10, 2010 at 6:49 pm #

    On a closer reading of the stats, like you Margaret, I wonder why there has been so much hype about this so-called “miracle”. And, on listening to the webcast of Peter Elwood, I believe he is seeking to add to the hype.

    The (mythical) Hippocratic oath demands that doctors “do no harm”. Encouraging the use of Aspirin goes against that, in my opinion.

  4. Eileen December 12, 2010 at 8:46 pm #

    Why is there so much hype? Easy – how many of the journalists have ever been taught to read, understand and critique what they are reading? Without a good – not a basic – knowledge of how to do that, they just pick up the bit that sounds great. And anyway – what journalist would be praised for a boring “it doesn’t do anything”. Now if the research had said that taking your aspirin would kill you – now that would have been a great front page headline.

  5. Margaret McCartney
    margaretmccartney December 12, 2010 at 9:16 pm #

    I have a thick file, alas, of university press releases and researchers publishing their work in terms of relative rather than absolute benefit, and generally making sure that they get every bit of media impact they can out of their work. They argue that they need to ensure future funding and publicity helps them do that. In a way I feel a bit sorry for some journalists who are merely reporting on what the ‘experts’ have to say. What you want, you’re right, is someone who can critique press releases and the ‘expert’ briefing – and yes, that would be often at the expense of duller, but more useful, headlines. But that wouldn’t be so necessary at all if researchers didn’t operate in a system where they feel that they must big-up all their findings – or were professional enough to report their work as showing marginal benefits, or no benefit at all.
    That kind of information is essential for patients and clinicians, after all…..

  6. Anne Marie Cunningham December 14, 2010 at 5:48 pm #


    There is a clear reduction in all-cause mortality within 10 years. In the three trials with long-term follow-up the median duration of treatment was 6, 4.4 and 6.9 years. It would really be very remarkable if there continued to be a reduction in all-cause mortality 15 years after people stopped taking the treatment? And as the authors point out many of the those in the control arm probably did start aspirin during follow-up but they chose to perform an intention-to-treat analysis anyway.

    Anne Marie

  7. Margaret McCartney
    margaretmccartney December 29, 2010 at 10:22 pm #

    I don’t know: what’s being proposed is an intervention to take for decades; I’d want some pretty good long term data on that – including side effects. The study this data came from was not designed for the purpose of gathering it specifically: we all know how difficult it is to rely on data which is not being specifically collated for that purpose. I’m not arguing that there was no benefit in this data analysis. What I am concerned about is the message from enthusiasts claiming that there are no ill consequences from using aspirin in this way. and a happy new year….

  8. Margaret McCartney
    margaretmccartney December 30, 2010 at 9:23 am #

    editorial in today’s bmj

    “The most rigorous approach to assessing this is to evaluate all cause mortality. To their credit, Rothwell and colleagues analysed this in some detail and report a statistically significant reduction in all cause mortality.4 Crucially, the effect is only marginally significant (relative risk of death in the aspirin arm 0.93, 0.87 to 1.00; P=0.045) (figure⇓). This is not particularly robust, especially as six of the eight trials were in subjects at high risk of cardiovascular disease, where the benefits of aspirin should be greatest. Furthermore, the authors included one RCT where the data had been destroyed….

    The data are not robust enough to recommend aspirin for the general population and we still do not know the optimum dose and length of treatment.”