Different dietary proteins determine different serum cholesterol levels if fed in a semisynthetic diet to some, but not all, animal species. In one species, the rabbit, this metabolic response is elicited without adding high sucrose or cholesterol supplements that have to be added to rat or pig diets in order to cause a similar response. Eleven out of 13 studies show that casein and soy protein do not induce different serum cholesterol levels in normal man. More important, protein-induced differences of serum cholesterol concentrations have not been reported when appropriate nutritional methodology has been applied. We conclude that no protein-induced hypercholesterolemia is observed in primates, particularly not in the human species. Dietary recommendations urging the general public to reduce consumption of animal protein because of a higher atherogenicity are not supported by the present data. The biochemical basis of the metabolic responses has been studied by many investigators, but no convincing unifying concept has yet been identified. The recent observation of higher serum thyroxine concentrations following soy protein consumption (and vegetable protein in general) when compared to casein shed new light on this problem. This endocrine response explains a wide array of metabolic features of soy-fed rodents: the lower hepatic VLDL secretion, the higher hepatic HMG-CoA reductase activity, the higher hepatic apo B, E receptor activity, the higher fecal bile acid excretion, and finally the lower serum cholesterol concentrations.