A recent study which claimed that aspirin was a useful prophylactic for stroke in patients with atrial fibrillation (BMJ, 1999; 319: 958-64) has been resoundingly criticised by several medics.

The study by Helle-mons et al concluded that aspirin was the prophylactic choice in primary care for atrial fibrillation if there was no clear indication for more powerful anticoagulants such as warfarin. However, a great deal of criticism has been levelled at the study's use of aspirin compared with low dose anticoagulant therapy, suggesting that Hellemons et al may have been comparing apples and oranges.

Simon Ellis, consultant neurologist at the North Staffordshire Royal Infirmary, suggested that the question asked by the researchers whether low anticoagulation or aspirin should be used was the wrong one to ask since aspirin is more reasonably compared with standard anticoagulation therapy.

Several others, including Andy Evans, clinical lecturer, Department of Stroke Medicine at Guy's, King's and St Thomas's School of Medicine, have commented that excluding patients with chronic heart failure and patients aged 78 years or over from the anticoagulation therapy limited the study's ability to correctly detect the number of events experienced by those on warfarin.

Finally, criticism was levelled at the selection method for the patients who participated in the study, as many low risk patients were included, again potentially confounding the results.

The furore over Hellemons et al's conclusions illustrate neatly the folly of uncritically embracing the results of every new study. Often, the conclusions of studies like these are taken on board as gospel for clinical practice (BMJ, 2000; 320: 1008-9).