Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Low dose aspirin - harm and benefits

Risk of bleeding with low-dose aspirin
Drug interactions
Duration effects
Aspirin for primary prevention

By now most people are aware that aspirin and NSAIDs are associated with an increased risk of upper gastrointestinal bleeding, and the risk with NSAIDs has been carried before ( Bandolier 52 , 79 ). A question still asked though, concerns the risk associated with low dose aspirin (75 mg up to 300 mg a day) used in the secondary or primary prevention of heart disease or stroke. A new epidemiological study [1] and meta-analysis [2] may help.

Risk of bleeding with low-dose aspirin

A case-control study was conducted using the UK General Practice Research Database that contains details of patient demographics, medical histories and diagnoses, referrals and prescriptions. The UKGPRD is becoming an important research tool, used frequently by researchers from all over the world. The authors of the aspirin study were based in Madrid.

The source population were patients aged 40 to 79 years with the same GP for at least two years. Cases were identified as those with upper gastrointestinal bleeding or perforation, excluding those with cancer, varices, liver disease, or alcoholism, or with other obvious reasons why they may have had a propensity for gastrointestinal bleeding. For inclusion a case had to have a specific site of bleeding located in stomach or duodenum or a clinical diagnosis of peptic ulcer, and be referred to a specialist or admitted to hospital. Paper records confirmed the accuracy of the computer diagnoses. Controls were selected at random and matched for age, sex and year, and with the same exclusion criteria.

Current users were defined as use within 30 days, recent users between 31 and 180 days, and past users longer than 180 days before the incident date. Exposure to other drugs was noted and analysed. Results were adjusted for a variety of covariates using logistic regression to compute a relative risk.


There were 2,105 cases and 11,500 controls. They were well matched for age and sex. Two-thirds were men older than 60 years. The indication for aspirin use was predominantly for secondary prevention of coronary or cerebrovascular disease (about 80% of the time) in cases and controls. The most common dose prescribed was 75 mg (48%), 150 mg (27%) or 300/325 mg (19%). Only in 6% of people was the prescribed dose more than this.

For current users of low-dose aspirin the relative risk of upper GI bleeding was 2.0 (95% confidence interval 1.7 to 2.3). The risk diminished when aspirin had not been used for 180 days or more (Figure 1). The risks were the same for fatal or non-fatal bleeding, for men or women, and for age groups 40-59 years, 60-69 years and 70-79 years. There was no difference for risk associated with site of bleeding.

Figure 1: Upper GI bleed with low-dose aspirin according to recency of use


Enteric-coasted aspirin had a similar risk associated with upper GI bleeding or perforation to plain aspirin (Table 1).

Table 1: Upper GI bleed with low-dose aspirin according to formulation

  Adjusted relative risk (95% CI)
All sites Plain 1.9 (1.6 to 2.3)
  Coated 2.3 (1.6 to 3.2)
Gastric Plain 2.0 (1.5 to 2.5)
  Coated 2.2 (1.4 to 3.6)
Duodenal Plain 1.6 (1.3 to 2.1)
  Coated 2.2 (1.4 to 3.4)


There was a hint of a dose-response over a limited dose-range with a higher risk of upper GI bleeding at doses above 300 mg daily, but this was based on a very small number. There was no appreciable dose response below 300 mg.

Drug interactions

Using aspirin together with high-dose NSAID increased the risk of upper GI bleeding substantially, with a relative risk of 13 (9 to 21), beyond the sum of the independent effects. Interaction were only additive at low or medium doses of NSAIDs. There was no apparent interaction between aspirin and paracetamol or aspirin and steroids.

Duration effects

There was an indication that the risk of upper GI bleed with low-dose aspirin was greater in the first two months of use than with use over a longer time period (Figure 2).

Figure 2: Upper GI bleed with low-dose aspirin according to duration of use

Aspirin for primary prevention

A question frequently asked of Bandolier concerns the appropriateness of using low-dose aspirin in the primary prevention of cardiovascular disease. Which patients are likely to benefit? A new meta-analysis from Sheffield, with some clever thought behind it, goes most of the way to telling us.

The problem, simply put is this. Aspirin may reduce the number of heart attacks, but it can, at the same time, increase the number of bleeds, both gastrointestinal bleeds and perhaps strokes. Is there some point at which the benefits outweigh the risks, or vice versa?


The meta-analysis was of randomised trials of aspirin for primary prevention. Studies had also to give information on total cardiovascular events, myocardial infarction, stroke, bleeding complications, and all cause mortality.

There were four studies with 48,500 people, 25,000 of whom were given 75-500 mg aspirin daily. Only one trial included women. Trial duration was 3.8 to 6.8 years.


Table 2 shows the weighted mean results from the four trials for a variety of outcomes, as the absolute risk per year of the event in the control group, the absolute benefit or harm, and the odds ratio. Clearly low dose aspirin reduced cardiovascular events and heart attacks but increased haemorrhage, non-cerebral bleeds and non-minor bleeds (all non-cerebral bleeds not classed as minor). It was interesting that the odds ratio for the increase in bleeding events caused by low dose aspirin in this meta-analysis (odds ratio about 1.7) was similar to the increase found in the case-control study [1] (relative risk 2.0).

Table 2: Benefit and harm from low dose aspirin in a meta-analysis

Event Absolute risk in control group (Percent/year) Absolute benefit from aspirin (Percent per year) Odds ratio (95% CI)
Cardiovascular events 0.92 0.13 0.85 (0.78 to 0.94)
Myocardial infarction 0.52 0.15 0.70 (0.62 to 0.79)
Strokes 0.29 -0.02 1.06 (0.91 to 1.24)
All cause mortality 0.73 0.05 0.94 (0.85 to 1.04)
  Absolute risk in control group (Percent/year) Absolute harm from aspirin (Percent per year) Odds ratio (95% CI)
Best overall estimate of haemorrhage 0.13 0.09 1.69 (1.38 to 2.07)
Major non-cerebral bleeds 0.05 0.04 1.73 (1.14 to 2.63)
Non-minor bleeds 0.22 0.18 1.77 (1.40 to 2.25)

The authors then set out to calculate the net benefit by calculating the number of heart attacks prevented net of bleeding events at different levels of baseline risk for cardiac events. This was done on the basis of 100 patients treated for five years with low dose aspirin at annual risk levels of 0.5%, 1% and 1.5% a year. The assumption was that low dose aspirin reduced the risk of MI by 30%, but that the risk of bleeding was constant. They also give the number of patients needed to be treated with low dose aspirin over five years to prevent one event.

These results are in Table 3. At a low risk, 0.5% a year, 133 people have to be treated with low dose aspirin for five years to prevent one myocardial infarction. But as we take into account the increased risk of cerebral haemorrhage, major bleeds and non-minor bleeds, the NNT increases until eventually it becomes an NNH so that eventually one person is harmed for every 500 treated. The results look better with increasing cardiovascular risk.

Table 3: Benefit and harm from low dose aspirin: overall absolute benefits and NNTs

  CHD event risk
  0.5%/ year 1.0%/ year 1.5%/ year
Myocardial infarcts prevented Absolute reduction (%/year) NNT Absolute reduction (%/year) NNT Absolute reduction (%/year) NNT
Total 0.75 133 1.5 67 2.25 44
Net of 0.19 cerebral haemorrages 0.56 179 1.31 76 2.06 49
Net of 0.17 major bleeds 0.39 256 1.14 88 1.89 53
Net of 0.76 non-minor bleeds -0.20 -500 0.55 182 1.3 77
Numbers are calculated on basis of treating 100 patients for five years. NNT is the number of patients needed to be treated for five years to prevent one event. Negative values show number of patients needed to be treated to harm one patient


Together these two papers help us think our way through the problems around the use of aspirin in primary prevention. We may know that aspirin is a good thing for preventing heart attacks, but the risk of harm from bleeding is not negligible, even at the low doses used. The balance tips to benefit over harm when the annual risk of a cardiovascular event is 1% or above (the authors say > or = 0.8%), and the Sheffield Tables for Primary Prevention allow the risk to be calculated.

Their conclusion was that aspirin for primary prevention is safe and worthwhile at a coronary risk of 1.5% a year or more, safe but of limited value at a coronary risk of 1% a year, and unsafe at a coronary risk of 0.5% a year. Those with Sheffield Tables to hand will know that this means that primary prevention with aspirin makes no sense unless the patient is "bad enough" to be "on" the Sheffield tables. We have to remember that aspirin, like all other drugs, is a poison [3].


  1. FJ de Abajo, LA García Rogríguez. Risk of upper gastrointestinal bleeding and perforation associated with low-dose aspirin as plain and enteric-coated formulations. BMC Clinical Pharmacology 2001 1:1 ( )
  2. PS Sanmuganathan et al. Aspirin for primary prevention of coronary heart disease: safety and absolute benefit related to coronary risk derived from meta-analysis of randomised trials. Heart 2001 85: 265-271).
  3. Tramèr MR. Aspirin, like all other drugs, is a poison. BMJ 2000 321: 1170-1171.
previous or next story in this issue