Several studies have reported an inverse relationship between serum cholesterol levels and the risk of cancer, especially of the colon (Seven Countries, Framingham, Chicago studies, London Whitehall Study, Paris prospective study, New Zealand Maori, Honolulu Heart Study, Hypertension Detection and Follow-Up Program, ...). For example, in the Multiple Risk Factor Intervention Trial (361 662 men), the global mortality graph was J-shaped, higher at either side of the 4.6-5.1 mmol/l value of serum cholesterol. This increased mortality with lower serum cholesterol levels was due to increased numbers of death from cancer. However, when the relationship is studied with respect to the time elapsed between the cholesterol measurement and death from cancer, the relative risk of death in the lowest decile with respect to the average of the following deciles, decreases with the period between measurement of the serum cholesterol and time of death. The negative relationship between serum cholesterol and death by cancer, very significant for deaths occurring within the first 5 years, disappeared almost completely for deaths occurring after 5 years. Other trials designed mainly to examine cardiovascular risk, and concerning smaller numbers, have not demonstrated this inverse relationship between serum cholesterol and cancer. This negative relationship between serum cholesterol and cancer must be acknowledged. It is weak and concerns mainly colonic cancer, especially in men in the elderly age groups. Several explanations have been put forward: influence of the combination of factors, competition of risk of death by other causes, chance, alteration of normal biological function of the cell membrane.(ABSTRACT TRUNCATED AT 250 WORDS)