Fifteen patients who had had a myocardial infarction before the age of 43 were compared with thirteen age-matched normal subjects. Twelve of the patients and three of the controls had a delayed glucose and insulin peak in the glucose and insulin areas than normal curves. When the measurements of the four patients with the largest areas under the glucose tolerance curve were separated, significant correlations were observed in the remaining patients and controls. The ratio in serum of the concentrations of estradiol-17beta to testosterone (E/T) correlated with serum glucose area (r equals + 0.69, P is less than 0.001), insulin area (r equals + 0.80, P is less than 0.001), and the ratio of insulin area to glucose area (I/G) (r equals + 0.64, P is less than 0.005) in the glucose tolerance test. Serum cholesterol concentration correlated with E/T, insulin area, and I/G, and serum triglyceride concentration correlated with glucose area, I/G, and serum cholesterol concentration. The hypothesis is presented (i) that in men who have had a myocardial infarction, an abnormality in glucose tolerance and insulin response and elevation in serum cholesterol and triglyceride concentrations are all part of the same defect (glucose-insulin-lipid defect), (ii) that this glucose-insulin-lipid defect when glucose intolerance is present is the "mild diabetes" commonly associated with myocardial infarction but is based on a mechanism different from that of classical diabetes, (iii) that this glucose-insulin-lipid defect is secondary to an elevation in E/T, and (iv) that an alteration in the sex hormone milieu is the major predisposing factor for myocardial infarction.