Several news sources have reported that aspirin’s benefits in preventing heart attacks may be outweighed by the risk of stomach bleeding.
These results come from an analysis of data from a number of trials looking at whether aspirin could reduce the risk of heart attacks and strokes (vascular events). These studies featured data on 95,000 people with no history of vascular disease and a low risk of vascular events (the primary prevention group) and data from 17,000 people who had previous vascular disease and a high risk of vascular events (the secondary prevention group). Although aspirin use reduced the relative risk of vascular events in both groups, the lower absolute risk of such events within the primary prevention group means that the absolute benefit was smaller. This means that the benefits of the aspirin may not outweigh its associated risks of bleeding in this group.
The findings are a good illustration of the fact that the balance of benefits and risks of drugs may differ in different types of people. It also raises questions about whether aspirin should be prescribed to those with no past vascular disease on a patient-by-patient basis, rather than through blanket prescription.
Where did the story come from?
The study was carried out by the Antithrombotic Trialists’ (ATT) Collaboration, which is a large group of researchers from universities and research centres in the UK, US and Europe. The secretariat group organising the research was based at the Clinical Trial Service Unit and Epidemiological Studies Unit at the University of Oxford. This unit receives or has received funding from the UK Medical Research Council, the British Heart Foundation, Cancer Research UK and the European Community Biomed Programme. The study was published in the peer-reviewed medical journal The Lancet.
What kind of scientific study was this?
This was a study pooling statistics from a number of trials (a meta analysis) to look at whether regular aspirin use reduced the risk of vascular events such as heart attacks and strokes.
In particular, the authors were interested in what the balance of benefits and harms was in people who had never had vascular disease (disease associated with blood-vessel blockage). Previous meta analyses did not use individual patient data, and had not shown a clear overall benefit for aspirin in this group of people. This previous research was also unable to look at subgroups, such as elderly people, separately.
The researchers identified 16 published randomised controlled trials (RCTs) that compared aspirin use against no aspirin. These trials were either:
- primary prevention trials, which were in people who had no previous disease associated with blood-vessel blockage (for example, heart attack or stroke) at the start of the study, or
- secondary prevention trials, which were in people who had previously experienced these conditions by the start of the study.
The researchers included primary prevention trials that enrolled at least 1,000 non-diabetic patients who were scheduled to receive treatment for at least two years. They included secondary prevention trials that enrolled people with previous heart attack, stroke or mini-stroke. (These trials had been included in a previous meta analysis by the ATT group.)
The researchers only included trials where they could obtain information on what happened to individual patients, rather than those where only the overall result across all patients was available. Two RCTs were excluded because individual patient data could not be obtained. RCTs where anti-clotting drugs similar to aspirin (anti-platelet drugs) were used were excluded.
The researchers identified the first time any participant experienced a “serious vascular event” during the study. This was defined as heart attack, stroke, or death from these or other vascular (blood-vessel related) causes. They also looked for major coronary events (heart attack, death from a heart-related cause or sudden death), any stroke, death from any cause and any bleeds outside of the brain or skull (extracranial). Extracranial bleeds were usually defined in the individual studies as bleeds requiring a transfusion or leading to death, and usually occurred in the stomach.
The researchers used statistical methods to pool the data from all participants and look for differences between the aspirin and no-aspirin groups. Primary and secondary prevention trials were analysed separately. The researchers also looked at whether they could identify factors that affected the risk of having a range of vascular outcomes in people in the primary prevention trials. These factors included age, gender, body mass index, smoking, diabetes, blood pressure and blood cholesterol levels.
The researchers also grouped trial participants according to their predicted risk of coronary heart disease, based on what proportion of the control group experienced coronary heart disease events during the study. These groups were very low risk (five-year risk less than 2.5% without aspirin), low risk (2.5-5%), moderate risk (5-10%) and high risk (10% or more).
What were the results of the study?
The researchers included:
- Six primary prevention RCTs, featuring data on 95,000 people without previous heart attack or stroke and 3,554 serious vascular events.
- Sixteen secondary prevention RCTs featuring data on 17,000 people with previous heart attack or stroke and 3,306 serious vascular events.
In the primary prevention trials aspirin reduced the per-year risk of a serious vascular event from 0.57% to 0.51%, which was an absolute reduction of 0.06% per year. This equated to a relative risk reduction of 12% compared with no aspirin (relative risk [RR] 0.88, 95% confidence intervals [CI] 0.82 to 0.94). None of the factors investigated (such as gender, age, cholesterol, high blood pressure or predicted risk of coronary heart disease) significantly altered this relative reduction in risk. These primary prevention trials used a range of doses of aspirin, including one that used a daily dose of 500mg, a higher dose than is currently recommended for use in preventing vascular events.
In the secondary prevention trials aspirin reduced the per-year risk of a serious vascular event from 8.19% to 6.69%, which was an absolute reduction of 1.49% per year. This equated to a relative risk reduction of 19% compared with no aspirin (RR 0.81, 95% CI 0.75 to 0.87).
There was no significant difference in the relative reduction in risk of events between primary and secondary prevention trials. However, because the absolute risk of events was higher in the secondary prevention trials this was a greater reduction in terms of absolute risk.
When the researchers further broke down serious vascular events occurring in the primary prevention trials they found that aspirin did not significantly reduce risk of stroke or death from vascular causes, but it did significantly reduce the per-year risk of non-fatal heart attack from 0.23% to 0.18%.
In the primary prevention trials aspirin increased the per-year risk of major extracranial bleeding from 0.07% to 0.10%, an absolute increase of about 0.03% and a relative increase of 54% (RR 1.54, 95% CI 1.30 to 1.82). This was mostly through an increase in non-fatal bleeds.
Aspirin also increased the risk of major extracranial bleeds in the secondary prevention trials (RR 2.69, 95% CI 1.25 to 5.76). However, there were few such bleeds in the secondary prevention trials (only 29 cases), so the pooled result may not have been very reliable.
What interpretations did the researchers draw from these results?
The researchers conclude that the overall value of aspirin is uncertain in people who have not had previous vascular disease because the benefits of a reduction of vascular events need to be weighed against any increase in major bleeds.
They say that their results could help make appropriate individualised decisions about whether a person should use aspirin, and that their results “do not seem to justify general guidelines advocating the routine use of aspirin in all apparently healthy individuals above a moderate level of risk of coronary heart disease”.
What does the NHS Knowledge Service make of this study?
The pooling of a very large amount of data has shown that, for people with previous vascular disease, the risks associated with aspirin use seem to be outweighed by the benefits, but that this is not necessarily the case for people without vascular disease.
One particular strength of this study was that it had access to data about individual patients, which allowed the authors to look at the effects of individual people’s characteristics, such as their age, gender and body mass index. This helps the researchers to determine whether there might be any specific subgroups that might benefit more than others. This is important, as the subgroups they analysed included men over 65 years and those with a five-year risk of coronary heart disease greater than 10%. These are groups that have a high risk of future heart attacks and also a slightly higher risk of major bleeds with aspirin. These results may help inform patient-by-patient judgements about whether to take aspirin or not.
The authors suggest that there may yet be a subgroup of people without vascular disease that shows substantial overall benefit with aspirin, for example, those with diabetes. They report that two larger trials are recruiting people with diabetes to investigate this possibility further. They also say that further trials are ongoing in people without vascular disease who are at moderate-to-high risk of coronary heart disease, a group that was well represented in the trials carried out so far.
These results will no doubt further debate about the wisdom of blanket usage of aspirin in people without vascular disease, and whether there are subgroups of these people who may benefit.
Analysis by Bazian
Edited by NHS Website
Links to the headlines
The Daily Express, 29 May 2009
Daily Mail, 29 May 2009
The Daily Telegraph, 29 May 2009
BBC news, 29 May 2009
Links to the science
The Lancet, May 30 2009