Atrial fibrillation (AF) is found in 0.4% of the adult population and is a common condition in the elderly. Its prevalence increases with age to affect 5 to 14% of those over 74 years. Recent evidence indicates that, compared with sinus rhythm, AF is associated with a 4-to 7-fold increase in the risk of stroke. However, there is strong evidence from randomised trials that full anticoagulation with warfarin substantially reduces the risk of stroke. Elderly patients are among those at higher risk and stand to gain the most from such treatment. They are also at higher risk for complications related to anticoagulant therapy and this sometimes makes clinical decisions difficult. There is a strong rationale for prescribing warfarin for all patients with AF who are over 65 years and free of contraindications. Some concerns exist about the benefit: risk ratio of warfarin in patients aged > 75 years. The answer is probably to use low intensity anticoagulant therapy (international normalised ratio 2.0 to 3.0), which is safer but no less effective than higher intensity regimens. Few data are available in the literature on physicians' attitudes to anticoagulation in elderly patients with AF. Although the results of randomised clinical trials in AF seem to suggest that anticoagulants and/or aspirin (acetylsalicylic acid) are underused in the elderly, over 90% of the patients initially screened were excluded from randomisation, making the sample highly selected. Compared with randomised controlled trials, some observational studies seem to indicate a higher likelihood of using anticoagulation and have targeted the intensity of anticoagulation according to age and clinical scenario.