Higher levels of physical activity have been related to higher concentrations of high density lipoprotein (HDL) cholesterol and lower concentrations of triglycerides. To test the hypothesis that the association between physical activity and the lipoprotein profile is mediated at least in part through increased insulin sensitivity, the authors measured fasting serum levels of HDL cholesterol, triglycerides, insulin, and glucose in 87 men and 83 women (aged 20-24 years) from a population-based survey in Beaver County, Pennsylvania, in 1981-1982. An insulin sensitivity index was calculated as the reciprocal of the insulin and glucose product multiplied by 10,000. Univariate analysis among men indicates that HDL cholesterol was positively related to insulin sensitivity (r = 0.24, p less than 0.05) and to the physical activity score as assessed with Paffenbarger's questionnaire (r = 0.21, p less than 0.05). Insulin sensitivity and physical activity score were positively related (r = 0.14), although not significantly (p = 0.21). Triglycerides were inversely related to both physical activity (r = -0.22, p less than 0.05) and insulin sensitivity (r = -0.19, p = 0.07). No significant findings among women were noted. Multivariate results indicate that the relation between physical activity and the male lipoprotein profile is reduced after controlling for the effects of insulin sensitivity (p greater than 0.10). The authors conclude that in these young men the beneficial effect of physical activity is likely to be partially mediated by increased insulin sensitivity. The lack of findings among women suggests that sex hormones may influence the association between insulin sensitivity and lipoprotein lipids.