In nondiabetic individuals, a poor response to an endothelium-dependent vasodilator in coronary vessels has been shown to increase the likelihood of a future cardiovascular event. Such prospective data are not as yet available in patients with type 1 or type 2 diabetes. However, consistent with the greatly increased cardiovascular risk in these patients, endothelial dysfunction has been almost universally found to characterize patients with type 2 diabetes particularly. Endothelial dysfunction frequently coexists with features of insulin resistance, such as the presence of small dense low-density lipoprotein (LDL) particles even in nondiabetic individuals. This association is independent of obesity and other causes of endothelial dysfunction, such as LDL cholesterol, hypertension, and smoking. In patients with type 1 diabetes, endothelial dysfunction has been found in approximately half of the studies. In some but not all studies, endothelial dysfunction has been especially severe in patients with poor glycemic control. Reversal or amelioration of endothelial dysfunction has been documented by many commonly used therapeutic agents such as successful insulin therapy, fibrates, and angiotensin-converting enzyme inhibitors, but also with some but not all agents that act as antioxidants. Long-term studies addressing the prognostic significance of endothelial dysfunction and its reversal are urgently needed to determine whether measurement of endothelial function could be used to identify individuals at risk better than can be done at present using classic risk factor assessment among patients with type 2 diabetes especially.