Diabetes mellitus is usually recognized as a major risk factor for coronary heart disease (CHD) morbidity and mortality. As chronic hyperglycaemia defines diabetes mellitus, it is logical to think that hyperglycaemia itself is related to these complications. But, in diabetic patients, there appears to be no relation between CHD and the specific characteristics of diabetes, namely the duration of the disease and the level of blood glucose. Moreover, in the UGDP trial, the group of patients with the best-controlled glycaemia failed to demonstrate a reduction in CHD death, in comparison with the group with poor blood-glucose control. These results are reinforced by the data from 15 prospective studies, which failed to demonstrate a homogeneous and significant increase in CHD risk with asymptomatic hyperglycaemia. Coronary heart disease mortality depends on a constellation of risk factors which are interrelated to some extent; one of the risk factors, namely blood glucose, is not an independent risk factor, nor is diabetes mellitus. In patients with abnormal glucose tolerance, the factors contributing to CHD risk are hypertriglyceridaemia and also a central fat distribution. This central fat distribution could explain the low CHD risk in nondiabetic women, which contrasts with the high risk in diabetic women, who are characterized by central adiposity, as are diabetic men. Björntorp hypothesized that intra-abdominal fat, which has an exceedingly sensitive lipid mobilization capacity, could play a key role in the development of the cluster of metabolic abnormalities that are present in the plurimetabolic syndrome. Hyperglycaemia, one of the anomalies of the syndrome, could be an 'innocent bystander' rather than a 'partner in crime', in the manifestations of CHD.