Skip navigation

Gastrointestinal harm with aspirin


Clinical bottom line

There will be about one additional bleed in every 100 people taking daily aspirin using a wide definition including melena and haemetemesis, and for the most serious definition of bleed requiring hospital admission perhaps 1 or 2 extra cases per 1,000 patients.


LA Garcia Rodriguez et al. Association between aspirin and upper gastrointestinal complications: systematic review of epidemiologic studies. British Journal of Clinical Pharmacology 2001 52: 563-571.

S Derry, YK Loke. Risk of gastrointestinal haemorrhage with long term use of aspirin: meta-analysis. BMJ 2000 321: 1183-1187.

M Tramèr. Aspirin, like all other drugs, is a poison. BMJ 2000 321: 1170-1171.

The gastrointestinal adverse effects of aspirin are well understood. Higher doses are associated with gastric erosions, some of which can become ulcers, and some of which can bleed. The issue that excites people is the risk of this with low dose aspirin given prophylactically to prevent vascular events. We now have several meta-analyses that try and put a number on it, though there are differences in study architectures used, and definitions of adverse events.

Epidemiologic studies

A systematic review (Garcia Rodriguez) found 17 original epidemiologic studies (about 60,000+ patients) published between 1990 and 2001. Articles had to be case control or cohort studies on aspirin use and upper gastrointestinal complications defined as bleeding, perforations or other serious gastrointestinal events leading to admission to hospital or visit to a specialist.


The overall relative risk was 2.6 (95% CI 2.4 to 2.7). Studies that examined dose of aspirin tended to have lower relative risk for lower doses. No pooling was done on any but the overall risk at any aspirin dose.

Randomised studies

A systematic review (Derry & Loke) found 24 randomised studies with about 66,000 patients. Trials had to compare aspirin (any dose) against placebo or no treatment, and be of at least 12 months duration. The outcome included melena or haematemesis, but not selected outcomes like major bleeds, hospital admission, or transfusion.


The pooled odds ratio for bleeding in studies with aspirin doses of 50-162 mg/day was 1.6 (1.4 to 1.8). The odds ratio for doses of 163 to 1500 mg/day was 1.7 (1.5 to 1.9).

With control and aspirin doses of 50-162 mg/day, bleeding rates were 362/24963 (1.5%).

With control and aspirin doses of 163-1500 mg/day, bleeding rates were 462/32365 (1.4%).

Is there a dose-response with aspirin?

Tramèr, in an accompanying editorial to the Derry & Loke review comments that the definition of bleeding with aspirin may be important in determining a dose response. He points out that using a more exacting definition of serious bleed that did not include haematemesis or melena, studies could show a dose response with daily aspirin (Figure 1).

Figure 1: Dose response of serious bleed with aspirin


It is pretty clear that aspirin is related to increased risks of gastrointestinal bleeding, which are increased by about twice for daily aspirin use. The risk of serious harm is probably dose-related. For the less serious definition (about 90% of cases) there will be about one additional case in every 100 people taking aspirin, and for the most serious definition perhaps 1 or 2 extra cases per 1,000 patients.