Low concentrations of HDLs in plasma are a strong predictor of risk for coronary as well as other cardiovascular diseases. There is increasing evidence that this relation is causal and that interventions that change HDL concentrations also change risk. One such intervention is exchanging fat and carbohydrate. In controlled trials, low-fat, high- carbohydrate diets decrease HDL concentrations. The effect is strongest when carbohydrates replace saturated fatty acids, but is also seen when carbohydrates replace mono- and polyunsaturated fatty acids carbohydrates. The effect is seen in both short- and long-term trials and therefore appears to be permanent. This finding is supported by epidemiologic studies in which populations eating low-fat, high- carbohydrate diets were shown to have low HDL concentrations. Weight losses with consumption of low-fat diets could theoretically counter effects on HDL, but in published trials weight losses have been modest and insufficient to offset the decrease in HDL concentrations induced by carbohydrates. Thus, replacement of saturated fat by carbohydrates adversely affects plasma HDL concentrations; replacement of saturated fat by unsaturated fatty acids deserves consideration as an alternative.