There is a significant risk of gastrointestinal (GI) haemorrhage with long-term aspirin therapy even at low doses, researchers have found.
The incidence of GI haemorrhage was compared in patients taking aspirin or placebo in a meta-analysis of 24 randomised, controlled trials involving 65,987 patients. Participants were predominantly male (74 per cent) and middle-aged, and were taking doses of aspirin of between 50 and 1,500mg per day for a mean duration of 28 months, say Dr Sheena Derry and Dr Yoon Kong Loke (department of clinical pharmacology, Radcliffe infirmary, Oxford).
GI haemorrhage occurred in 2.47 per cent of patients taking aspirin compared with 1.42 per cent taking placebo, giving an odds ratio of 1.68.
The effect of low-dose aspirin was then assessed separately in 49,927 participants who took between 50 and 162.5mg per day. Of these patients, 2.30 per cent of patients on aspirin and 1.45 per cent on placebo had GI haemorrhage (odds ratio, 1.59). "Even at these lower doses, aspirin was associated with a significantly increased risk of GI haemorrhage compared with placebo," the researchers say.
Use of modified-release preparations did not reduce risk either — an odds ratio of 1.93 was found. The researchers say that there is insufficient evidence to support the view that modified release preparations are safer than standard formulations in terms of GI haemorrhage but that other symptoms, such as nausea and epigastric pain, might be reduced.
The authors comment that the risk of GI haemorrhage could be higher in patients who take over-the-counter aspirin without consulting a doctor (British Medical Journal 2000;321:1183).
In an accompanying leading article, Dr Martin Tramér (staff anaesthetist, Geneva university hospital) says that the study raises more questions than it answers.
He suggests that research needs to examine who should be given what dose of aspirin and for how long. He adds that innovative models are needed to estimate rare events with confidence and systematic reviews currently provide the best solution (ibid, p1170).