The Paris Prospective Study is a long-term investigation of cardiovascular diseases in a population of 7164 working men, aged 43-54 years. The first annual follow-up session (1968-73) included a 0-2 hr 75 g OGTT with measurement of plasma insulin and glucose levels beside the major coronary heart disease (CHD) risk factors: arterial pressure, cigarette smoking, weight, cholesterol, triglycerides. After a mean 11.2 years follow-up, 651 deaths, among them 126 due to CHD, were recorded. The annual CHD mortality rates were respectively 1.4, 2.7 and 3.2 per 1000 for the 6055 normoglycaemic, 690 impaired glucose tolerance, and 293 new and known diabetic subjects (1980 WHO classification) (p less than 0.01). The annual risk was analyzed by the multivariate Cox model. It showed that the fasting plasma insulin was positively associated with risk independent of the other factors (p less than 0.05), whereas glucose tolerance, including overt diabetes, was not significantly associated. We conclude that high insulin levels may constitute a more sensitive predictor of CHD than the degree of glucose intolerance, it could be useful to avoid excessive insulin plasma concentration, and even to lower its level.