select search filters
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

expert reaction to study looking at benefits and harms of aspirin in people without cardiovascular disease

A study, published in Annals of Internal Medicine, reports that the benefits of aspirin for patients without cardiovascular disease may outweigh the risks.


Prof Christopher Garland, Professor of Vascular Pharmacology, University of Oxford, said:

“The paper aims to identify whether people without cardiovascular disease may derive a net benefit from taking aspirin.  It is an interesting study, but there is a caveat.  A worry is the criteria for selection.  The target group are people without established cardiovascular disease as defined by PREDICT, but we don’t know whether any of these people were assessed for small artery disease.  They may well not have been, as this is difficult.  It is now recognised that endothelial cell dysfunction in the microcirculation of the heart occurs before large artery disease appears.  So it is possible those showing benefit with aspirin in this study may actually be a disease cohort, even though they do not yet have established cardiovascular disease.  This makes the results difficult to interpret, so we cannot be sure some adults without established cardiovascular disease would gain a net benefit from taking aspirin.”


Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This is a careful piece of research, based on good previous work.  But care is needed in interpreting it, I’d say.  There’s evidence that a daily dose of aspirin can reduce the risk of cardiovascular diseases (CVD, such as heart attacks and strokes), but, because one of the things aspirin does is to make it rather harder for the blood to clot, there’s also evidence that it increases the risk of major bleeding disorders in the body – particularly bleeding in the stomach and gut, but also bleeding inside the skull (including some types of stroke).  Both CVD and these major bleeding disorders are often serious and can be fatal.  So how do the two risks balance out?  People who already have CVD, so are at higher risk of having another heart attack or stroke, are often prescribed aspirin because the reduction in CVD risk would generally outbalance the increased risk of bleeding.  But the position for people who are not known to already have CVD is not clear, and that’s what this new study addresses.  Some doctors have suggested that everyone over a certain age should take a daily aspirin to reduce their CVD risk, but does that mean that too many of them will have a serious bleed?  In the UK, the advice from NICE is that aspirin should not be routinely prescribed to people who aren’t known to have CVD, although there can be grounds for doing so in people at particularly high risk of a heart attack or stroke.

“This new research uses the results of some major work in New Zealand on the risk of CVD and of major bleeding to put numbers on the chances of these problems in the New Zealand population.  Somewhere around a third of the relevant New Zealand population have already had their absolute CVD risk assessed, and their absolute risk of major bleeding can also be assessed using the same data.  So, using anonymised data from the people who had not had CVD, and information from international clinical trials of aspirin, the researchers worked out the change in the risks of CVD and of major bleeding if people took daily aspirin.  For every 1000 people with a given level of CVD and bleeding risk, they worked out how many fewer cases of CVD there would be and how many more major bleeds, if they all took aspirin.  For some levels of CVD and bleeding risk, taking aspirin would reduce the number of CVD cases by more than the increase in cases of serious bleeding, and for other risk levels, it’s the other way round.  In the main analysis they found that (over a 5-year period) substantial numbers of men, and to a lesser extent women, would have a net benefit from taking aspirin, when the net benefit is measured by ‘trading off’ a case of CVD with a case of major bleeding in this way.  For both men and women, the numbers who would be disadvantaged by taking aspirin are larger – that is, the increase in the chance of a major bleed, if they take aspirin, would be greater than the decrease in the chance of CVD.

“On the whole, this looks as if it is in line with advice not to prescribe aspirin routinely.  More people would be disadvantaged than would benefit.  But really that’s not the point at all.  The point is that, in the New Zealand population, the research predicts that substantial numbers of people would benefit from aspirin, even if these people are actually a minority in the population.  So the question is, would it be possible to prescribe aspirin to those people, and not to the larger numbers for whom taking aspirin would give a net disadvantage?  Well, the tools to make that distinction do exist – there are statistical models, for New Zealanders, to assess the risk of CVD and of major bleeding, and to see how those risks would change if they take aspirin, and a decision whether to take aspirin can be made on that basis.  But it’s not that easy, not entirely.  First, all these assessments are based on statistical models and statistical data, and so inevitably the numbers are uncertain to some extent.  The researchers did some calculations taking into account the statistical uncertainty in how much aspirin changes the risk of CVD and major bleeds.  That can make a very substantial difference to the percentage of people who are predicted to get net benefit from taking aspirin.  For women, they calculated that the percentage who would have a net benefit could plausibly be as low as 0.1% or as high as 60%, and for men, as low as 2% or as high as 80%.  The researchers did not do similar calculations to take into account statistical uncertainties about the absolute risks of the diseases in people who don’t take aspirin – those uncertainties will be smaller, but they exist.  So despite using the best data available, and using it in an appropriate way in my view, there’s still a lot of uncertainty about which people would benefit from taking aspirin.  In any case, even if we could somehow know everyone’s risk levels exactly, it would always be possible that someone with a high CVD risk and a low bleeding risk could still have a major bleed rather than a heart attack, because that’s how chance works.

“There are other issues too.  These main calculations treat a case of CVD as equally ‘bad’ as a major bleed – but both CVD and major bleeds cover a wide range, from relatively minor cases to fatal cases, and it’s not clear that everyone would want to trade them off in exactly this way.  Indeed, what the appropriate trade-off is might depend on individual patients’ choices.  The researchers do look at a different trade-off – one CVD case for two major bleeds – and again that makes quite a difference to their assessment of who would benefit from aspirin.  Also, there’s some evidence that taking aspirin might reduce the risk of certain cancers, particularly bowel and prostate cancers, and that might tip the balance of whether aspirin should be prescribed for individual patients – but the evidence on the size of cancer risk reductions seems to be even less certain than for CVD or bleeding risk.  This research took no account of possible effects of aspirin on cancer risk.

“Finally, it’s important to note that the numerical details in these results apply only in New Zealand.  The methods of estimating the risk of CVD and of major bleeds are based on data from New Zealand, and might give inappropriate figures for the risks in other populations.  The calculations of how many people would be harmed or would benefit from aspirin are again based on the New Zealand population and wouldn’t apply elsewhere, not in all their detail anyway.  Statistical tools for assessing the risk of CVD and of other diseases such as major bleeding do exist for other populations, and might well be useable in a similar way to what’s been done in New Zealand, but it would be wrong simply to take over the New Zealand results to other populations without further research.”


‘Personalized prediction of cardiovascular benefits and bleeding harms from aspirin for primary prevention: A benefit–harm analysis’ by Vanessa Selak et al. was published in Annals of Internal Medicine at 22:00 UK time on Monday 16 September 2019.

DOI: 10.7326/M19-1132


Declared interests

Prof Christopher Garland: “None.”

Prof Kevin McConway: “Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.”

in this section

filter RoundUps by year

search by tag