In this review we will discuss the cerebrovascular consequences of dysglycemia and current evidence for therapy, making reference to recent work in the fields of neuropathology, epidemiology, and relevant clinical trial data. Prospective observational and clinical trial data show a clear association between diabetes mellitus and vascular disease, which extends to cerebrovascular disease. The benefits of intervention to lower blood glucose in terms of microvascular health are well established but benefit on macrovascular, especially cerebrovascular, health has been less apparent. Recent large-scale trials and metaanalyses have helped us to better define the role of glycemic control in macrovascular disease. Although few studies of glycemic therapy have used cerebrovascular disease as a primary endpoint, stroke-specific data can be derived. Associations between blood glucose and outcome are also apparent for acute stroke. A period of hyperglycemia is common, with elevated blood glucose in the periinfarct period consistently linked with poor outcome in patients with and without diabetes. The mechanisms that underlie this deleterious effect of dysglycemia on ischemic neuronal tissue remain to be established, although in vitro research, functional imaging, and animal work have provided clues. While prompt correction of hyperglycemia can be achieved, trials of acute insulin administration in stroke and other critical care populations have been equivocal. Diabetes mellitus and hyperglycemia per se are associated with poor cerebrovascular health, both in terms of stroke risk and outcome thereafter. Interventions to control blood sugar are available but evidence of cerebrovascular efficacy are lacking. In diabetes, glycemic control should be part of a global approach to vascular risk while in acute stroke, theoretical data suggest intervention to lower markedly elevated blood glucose may be of benefit, especially if thrombolysis is administered. Trials have been underpowered to demonstrate treatment effect and any intervention must be balanced against risk of hypoglycemia.