While the use of aspirin in the secondary prevention of cardiovascular atherothrombotic disease is well established, many aspects of primary prevention are still unclear. Uncertainties mostly depend on a doubtful risk-benefit ratio, because of the low atherothrombotic risk of populations involved on the one hand, and the non-negligible bleeding risk of treatment on the other. Areas of specific doubt are those of diabetes and asymptomatic peripheral arterial disease, where neither single trials nor meta-analyses allow issuing high-grade specific recommendations at the moment. The present review aims at giving an account on this topic, highlighting areas for further studies, but also attempting at providing a rationale for what to do practically now, while awaiting more conclusive evidence. Based on the results of a number of clinical trials and meta-analyses, and especially considering the absolute figures of the benefit (major cardiovascular events avoided) and of the harm (major bleeding events occurred related to aspirin), the authors recommend to limit primary cardiovascular prevention with aspirin (in apparently healthy subjects with no previous cardiovascular events) to subjects with an estimated global cardiovascular risk ≥2 major cardiovascular events per 100 patients-year, as assessed by the risk score assessments proposed in the Italian "Progetto Cuore" (www.progettocuore.it). This cut-off should also be adopted for primary prevention in patients with type 2 diabetes and/or asymptomatic peripheral arterial disease.