In those who have already survived myocardial infarction (MI) or stroke, or have had a transient ischaemic episode (TIA), daily low dose aspirin (ASA) reduces the risk of recurrences by an amount that greatly exceeds the risk of serious bleeding (secondary prevention). ASA is therefore recommended for these people. However, in primary prevention-reducing risk in those so far free of clinically manifest episodes-the benefit is of the same order as the bleeding hazard, (which is much the same in both primary and secondary prevention contexts). The use of other effective agents such as statins further emphasises the even balance between benefit and hazard in primary prevention. Six primary prevention trials are reviewed, first singly and then in a meta-analysis based on individual patient data. ASA reduced non-fatal myocardial infarction by about 25%. However, death from coronary heart disease (CHD) was not significantly reduced (by 5%), nor was any vascular death (3%). There was a non- significant reduction in strokes of 5%, this being the net result of an 8% reduction in non-fatal stroke and a 21% increase in stroke death (mainly from haemorrhagic events), both effects being non-significant. Serious vascular events (MI, stroke or vascular death) were significantly reduced by 12%, mainly due to the large effect on non-fatal MI. About 1650 people would need to be treated with ASA for a year to avoid one serious vascular event, which contrasts with the 10-20 events avoided in secondary prevention by treating 1,000 patients for a year. Other primary prevention trials not included in the meta-analysis have also reported no benefits in MI or stroke, but the findings of still unpublished trials are awaited. Recently, however, encouraging results have come from meta-analyses of the effects of ASA on cancer incidence and mortality and on its effects on cancer metastasis, particularly for adenocarcinomas. Typically, reductions in these measures have been around 30% following treatment for four or five years, but more in several instances. These results alter the balance in primary prevention between benefit and hazard as it appears for arterial events alone, tipping it towards the use of ASA. Consequently, new guidelines on advice and decisions on ASA in primary prevention are now needed. Low dose ASA, eg. 75 mg daily is as effective as higher doses for all the vascular and cancer benefits established in the meta-analyses, and it causes less serious bleeding than higher doses.