[Food lipids, serum cholesterol and coronary heart disease (author's transl)]

A clear association is observed between saturated fats in the diet, mean serum cholesterol and the frequency of coronary heart disease (CHD) among human populations. Such an association led to a pathogenic interpretation: the diet heart hypothesis. But epidemiological studies have not put forward coherent relations between individual consumption of fats and either cholesterolemia or the risk of developing a CHD. No preventive trial has shown any definitely coherent and significant reduction of the incidence of CHD after a decrease of serum cholesterol following diet modifications. However changes of relative proportions of saturated and polyunsaturated fats induce predictable variations of the mean serum cholesterol of a group which can be maintained in free living populations by appropriate diet modifications on long periods of time. Such conflicting data are discussed with reference to methodology, etiological research and public health.