Epidemiologic studies indicate that hyperglycaemia is responsible for microangiopathy, but that its role in macroangiopathy is more controversial. The relative risk of coronary heart disease (CHD) is 2- to 3-fold greater for diabetic men and 3- to 5-fold greater for diabetic women. It is greater for lower limb arteriopathy (4- to 6-fold) and amputations (10- to 20-fold). Although relative risk is rather constant for different populations, absolute CHD risk depends on other risk factors and the rate of risk in the non-diabetic population. Yet hyperglycaemia is also a causal factor for CHD risk, as demonstrated in cohort studies of Type 1 diabetic patients without diabetic glomerulopathy or any associated CHD risk factors, and especially in diabetic Pima Indians who are genetically protected against hypercholesterolaemia and hypertension. Finally, according to WESDR and UKPDS data, the 10-year risk of cardiovascular mortality increases by 10% for every 1% increase in HbA1c value. Hyperglycaemia can be linked to atherogenesis through several pathways: gluco-oxidation of the extracellular matrix inducing accelerated atherosclerosis, endothelial dysfunction, with a decreased production or inactivation of NO, a thrombogenic tendency, with increase in PAI1, Willebrandt factor and platelet aggregation, and last (but not least) dyslipidaemia subsequent to lipoprotein glycooxidation and increased production of VLDL. Hyperglycaemia was associated with hyperlipoproteinaemia and high plasma triglyceride levels, low plasma HDL levels and high plasma levels of small and dense LDL in three high risk populations: diabetic women, Asian migrants, and the Finnish population of Kupsio. Moreover, impaired glucose tolerance appears to be a CHD risk marker indicative of insulin resistance apparently responsible for atherosclerosis related to an association of CHD risk factors rather than to hyperinsulinaemia. The precedence of insulin resistance over onset of Type 2 diabetes is consistent with the existence, at diagnosis, of clinical complications of atherosclerosis in 20% of cases, as confirmed in the UKPDS study. Finally, though blood glucose control by sulphonylureas and insulin does not appear to be deleterious, these drugs have not shown their efficacy in reducing macrovascular adverse events in Type 2 diabetes, either because the cumulative incidence of events is inadequate (DCCT) or the efficacy of long-term hypoglycaemic effects is not apparent (UKPDS). Moreover, these studies have shown that exogenous as well as endogenous hyperinsulinaemia can lead to increase in weight, with potentially atherogenic android fat distribution. In conclusion, though the correlation between hyperglycaemia and microangiopathy is linear and well-established worldwide (with every 1% increase or decrease in HbA1c resulting in a 30% increase or decrease in microangiopathic events), the same is not true for macroangiopathy, whose prevalence is variable among populations. Thus, CHD mortality due to diabetes is 5-fold lower in France than in Finland, though the Monica study indicates a disparity within the French community.