Wednesday March 3 2010
Aspirin: more harm than good?
“Healthy people who take aspirin in the hope of preventing a heart attack or stroke are doing themselves more harm than good,” The Daily Telegraph reported. It said that healthy people who take a low dose of daily aspirin to reduce their risk of a heart attack are also increasing their likelihood of major internal bleeding.
The news is based on a study in almost 30,000 men and women aged between 50 and 75 without known heart disease. It found that taking 100mg aspirin daily almost doubled the risk of dangerous internal bleeding compared to dummy pills (placebo), while having no effect on heart attacks or strokes.
The Telegraph report is accurate. This well-conducted study suggests that the risks and benefits of aspirin are both small in this group of patients at low risk of vascular disease. Though it is possible that the risks outweigh the benefits neither outcome reached statistical significance. There are other groups of patients who are at higher vascular risk, for example, those with high blood pressure, cholesterol and diabetes, who may benefit from aspirin. People taking aspirin following a heart attack or stroke should continue to do so as instructed.
Where did the story come from?
This research was carried out by F. Gerald R. Fowkes and colleagues for the Aspirin for Asymptomatic Atherosclerosis Trialists. The study was funded by the British Heart Foundation and the Chief Scientist’s Office in Scotland. Bayer HealthCare provided the aspirin and placebo tablets and funds for packaging, dispensing, and some statistical analysis.
The study was published in the (peer-reviewed) Journal of the American Medical Association.
What kind of research was this?
This research tested the effectiveness of aspirin at preventing cardiovascular events in people who were thought to be at risk of atherosclerosis and cardiovascular events through screening. This study design was a large, double-blind randomised controlled trial, which ran from 1998 to 2008 in Scottish communities. The researchers were interested in both good and bad outcomes. Initially they set out to see if fatal or non-fatal heart attacks, strokes or deaths were reduced by aspirin, but they also monitored the participants for side effects of aspirin, such as bleeding.
The study was well-designed and carefully conducted.
What did the research involve?
The screening consisted of the ankle brachial index (ABI), which is a simple, inexpensive test. It involves participants lying down for five minutes, during which the blood pressure in their feet is compared to that in their arms. Blood pressure is measured using a typical blood pressure cuff and an ultrasound probe to detect the pulse in two arteries of the feet. The ratio of blood pressures is recorded (above 0.95 is thought to be normal and below 0.95 is thought to indicate narrowing of the arteries to the legs).
The researchers wanted to see if the ABI test could be used in population screening programs to identify people who might benefit from preventive treatments. There is some uncertainty about the benefit of the test in screening, with some US guideline development groups saying that screening should be considered in primary care among certain high-risk groups, and others not recommending screening at all.
The participants were recruited from a community health registry of people living in central Scotland. Invitations to ABI screening were sent to 165,795 people aged 50 to 75. Of these, 28,980 men and women were screened. The researchers then excluded anyone who already had diagnosed vascular disease, were already taking medication such as aspirin or warfarin, or were unwilling or unable to participate. This left 3,350 people with an ABI of 0.95 or less for randomisation to either aspirin or placebo.
The participants were split into two equal groups. 1,675 participants received aspirin at a dose of 100mg daily and 1,675 received a placebo (dummy pill). The researchers followed all except 10 participants for over eight years on average. Participants were seen at intervals of three months, one year and five years in the clinic and were then contacted annually by telephone. They also received a mid-year letter, enquiring generally about any problems, and an end-of-year newsletter.
The researchers monitored fatal or non-fatal heart attacks, stroke or revascularisation (such as angioplasty or bypass grafts). They also looked for all deaths, angina, intermittent claudication (pain in the legs on walking due to narrowing of the arteries) and warning strokes (transient ischaemic attacks). Results were analysed appropriately in the groups to which the patients were originally allocated.
What were the basic results?
By the end of the trial, 357 participants had had a fatal or non-fatal heart attack, stroke or revascularisation; a rate of 13.5 events per 1,000 person-years (95% confidence interval [CI], 12.2 to 15.0).
There was no statistically significant difference between groups. There were 13.7 events per 1,000 person-years in the aspirin group compared to 13.3 in the placebo group (hazard ratio [HR] 1.03, 95% CI 0.84 to 1.27).
No statistical significance was observed between groups in other outcomes including death from any cause (176 deaths in the aspirin group compared with 186 in the placebo group).
A first major haemorrhage requiring admission to hospital occurred in 34 participants in the aspirin group (2.5 per 1,000 person-years) and 20 in the placebo group (1.5 per 1,000 person-years; HR in favour of placebo group, 1.71, 95% CI, 0.99 to 2.97).
How did the researchers interpret the results?
The researchers say that in this study “the administration of aspirin compared with placebo did not result in a significant reduction in vascular events.”
This trial has tried to answer an important question regarding who should be given aspirin to prevent a heart attack or stroke. It used a systematic method to screen people and followed a reasonably large group of patients for up to 10 years in some cases. The finding of “no statistical significance” can be an important result, and in this case suggests that any benefits from taking aspirin for this group of people are likely to be small. The risk of bleeding was also small and not technically of statistical significant.
- There is a non-significant trend in the results towards aspirin being harmful. As there is also a suggestion that the study may have been underpowered (planned for too few people), this implies that a larger study may have detected a significant increase in major bleeding in the aspirin group. However, the fact that the reported results were not statistically significant has been picked up by the newspapers.
- Although there were more bleeds in the aspirin group than the placebo group, they varied in their severity. and not all bleeds had the same implication for patients. For example, some episodes of bleeding from stomach ulcers were easily treated, while other cases of bleeding from haemorrhagic stroke were fatal. There were three fatal haemorrhagic strokes in both groups. Fourteen patients in the aspirin group required admission to control bleeding (reasons not given) compared to five in the placebo group. By combining the bleeding outcomes important information is lost.
- Considering this trial originally screened about 30,000 people, it is important to keep the small number of patients (9) who died from major haemorrhage in perspective.
Overall, this study has found that aspirin doesn’t appear to be of benefit in preventing cardiovascular disease in this group of patients at least, and suggests it could even increase bleeding. There are other groups of patients at higher vascular risk, for example, those with high blood pressure, cholesterol and diabetes who may benefit from aspirin. People taking aspirin following a heart attack or stroke should continue to do so, and others should consider being assessed for vascular risk.