The Paris Prospective Study is a long-term, large-scale study of the factors predicting coronary heart disease in healthy middle-aged men. Subjects with impaired glucose tolerance or diabetes (not treated by insulin) at the first follow-up examination (n = 973) were selected from the total cohort for a separate analysis of the predictors of death from coronary heart disease. An index of body fat distribution, the iliac to thigh ratio, was entered into the list of potentially predictive variables, despite the fact that it had been measured one year before the first follow-up examination. After 15 years of mean follow-up, 41 of the selected subjects had died from coronary heart disease. Univariate analysis showed that these subjects differed from the subjects who died of another cause or who were alive at 15 years on the following variables: iliac to thigh ratio (p less than 0.0005), plasma triglyceride level (p less than 0.006), systolic blood pressure (p less than 0.01), and body mass index (p less than 0.04). In multivariate regression analysis using the Cox model, only iliac to thigh ratio and triglyceride plasma level achieved statistical significance as independent predictors. This result supports the current hypothesis that upper-body fat distribution, a characteristic trait of subjects with diabetes of glucose intolerance, plays an important role towards their high cardiovascular risk. However, it is unlikely that this role would be mediated through the lipid abnormalities that have been described as associated with upper-body fat deposition.