[Blood cholesterol and mortality].

Mortality rates vary with serum cholesterol levels: the causal nature of this relationship is studied by prospective studies (analysis of associations) and unifactorial primary prevention trials (experimentation to determine causality). In prospective studies all cause mortality is often increased at low and high cholesterol levels because of the inverse relationships of this factor with cardiovascular mortality (positive) and non-cardiovascular mortality (negative). Coronary death increases proportionally to serum cholesterol levels in all populations, including those with low cholesterol levels, in both sexes and at all ages. The relationship with cerebrovascular mortality seems to depend on the clinical feature: increased mortality rate due to cerebral haemorrhage in patients with low serum cholesterol and a positive correlation with cerebral thrombosis. Mortality due to cancers is generally negatively correlated to serum cholesterol levels with a significant increase in mortality in patients with a low serum cholesterol. This relationship often becomes less significant as the time between measurement and death increases. Mortality due to violent causes is not usually related to serum cholesterol. The multitude of possible causes of confusion makes any causal interpretation of data illusory. Experimentation by unifactorial primary prevention trials is essential for any etiological research but none of the trials performed to date was designed to analyse the effect of lowering serum cholesterol on global or coronary mortality.(ABSTRACT TRUNCATED AT 250 WORDS)