The balance of evidence from observational studies suggests that elevated levels of homocysteine are associated with increased risk of carotid artery disease and stroke. There is, however, a paucity of prospective studies. There are also concerns regarding confounding caused by factors associated with hyperhomocysteinaemia, including renal impairment, an atherogenic diet and cigarette smoking. Homozygosity for a defective thermolabile variant of methylene-tetrahydrofolate reductase, a common genetic polymorphism which results in hyperhomocysteinaemia, has not been consistently linked with stroke or other vascular diseases. Additional prospective studies are required, with sufficient power to characterise the form of the association between homocysteine concentrations and stroke risk, whether linear or threshold, and to study interactions between homocysteine, other dietary markers and established stroke risk factors such as smoking and hypertension. Ultimately, the case for a causal role for elevated levels of homocysteine in vascular disease, including stroke, will depend on data from randomised controlled trials of homocysteine-lowering interventions. Given the high prevalence of hyperhomocysteinaemia in apparently well-nourished populations and the tendency for homocysteine concentrations to increase with age, modest effects of homocysteine on stroke risk will have profound implications for public health.