There has been much concern and discussion whether reduction of serum cholesterol by diet or drugs, while preventing coronary heart disease, might cause an increase in the risk of dying from noncardiovascular disorders, in particular cancer. A review of the evidence from ecological, cross-cultural comparisons, prospective epidemiological cohort studies and intervention trials provides no indication for a cause-and-effect association between low serum cholesterol or serum cholesterol lowering and the risk of cancer, except possibly at very low levels. An inverse relationship between noncardiovascular-noncancer mortality and serum cholesterol level does exist but the link is unlikely to be causal; for better understanding, there is a need for studying the biological, lifestyle and social correlates of this association. Continuing observations should pay attention to the accuracy of information on specific causes of noncardiovascular deaths, especially mortality not due to disease. Further research needs to include investigations into the determinants of low serum cholesterol levels and their possible influence on biological functions. It is concluded that the current recommendations for coronary heart disease prevention, aimed at lowering elevated serum cholesterol levels in individuals (the 'high-risk strategy') and in populations at high risk (the 'population strategy') require no basic modification in the light of present knowledge on the relationships between cholesterol and noncardiovascular disease mortality.