Dietary and endogenous cholesterol and human cancer.

Recent questions about the role of cholesterol (particularly blood cholesterol) in human cancer have assumed considerable scientific and public health importance. This paper has reviewed the evidence relating human cancer to cholesterol, in diet, blood, and feces. With respect to dietary cholesterol, there is moderately consistent evidence, both descriptive and analytic, of a small-to-medium increase in risk of cancers of the colon and female breast in association with increased dietary cholesterol. However, the close correlation of cholesterol with other foods and nutrients precludes causal inference. The association of fecal cholesterol with large bowel cancer, in both descriptive and analytic studies, is inconsistent. However, there is some evidence that individuals with reduced degradation of fecal cholesterol are at increased risk of colon cancer. Other bile-derived fecal components, particularly the acid sterols, show a somewhat more consistent relationship with large bowel cancer. It may be of relevance to the findings on blood cholesterol that bile is produced from hepatic cholesterol which derives, in part, from blood cholesterol. Studies of blood cholesterol and cancer have been either experimental (intervention) or observational (primarily follow-up). Deliberate lowering of blood cholesterol, by either drugs or diet, does not appear to alter the risk of cancer, either overall or of specific types. The findings from 20 published follow-up studies, each initiated as a cardiovascular disease study, have been more varied. In 12 studies, an inverse association was observed between blood cholesterol level and overall cancer risk. Eight of those 12 were mortality studies, and in six, the inverse association was confined to deaths that occurred early in follow-up; this observation is consistent with lowered blood cholesterol having occurred as a metabolic response to a preclinical cancer. However, the results of the other two mortality studies do not exclusively support this interpretation. Furthermore, in three of the four incidence studies that reported an inverse association, the inverse association persisted for 10 or more years. This relationship was most marked for colon cancer in men and showed some evidence of being maximal in the proximal colon. The biologic plausibility of these particular observations on colon cancer risk in relationship to an antecedent naturally occurring low blood cholesterol gains some support from a body of epidemiologic, clinical, and experimental evidence.(ABSTRACT TRUNCATED AT 400 WORDS)