Subjects with fasting and/or postprandial plasma glucose levels that are higher than those in normal subjects but less than those in patients with manifest NIDDM are at increased risk of cardiovascular disease (CVD). This association between glucose intolerance and cardiovascular disease was first hypothesized in the 1950s but was not substantiated until the results of several long-term prospective studies became available. The results suggest that there is not a continuously variable association between blood glucose levels and the risk of CVD but that the risk of CVD becomes evident at the upper end of the distribution of glucose tolerance, though different studies yield different threshold levels. With the introduction of the descriptive term Impaired Glucose Tolerance (IGT), studies were undertaken to establish the relations between IGT and CVD. To date most data refer to people of European origin. In several populations CVD has been observed to be more prevalent in subjects with IGT and/or the incidence of CVD is higher in subjects with IGT than in normoglycaemic controls. A causal link between hyperglycaemia and CVD seems unlikely from the published evidence. The most probable explanation of the association is "a common soil', that is, a number of associated metabolic abnormalities which may lead to CVD and IGT/NIDDM or both.