ABSTRACT
In both men and women, the relationship between all non-CVD deaths and serum cholesterol showed an upturn in the group with serum cholesterol levels below 160 mg/dl (4.0 mmol/l), but did not show an inverse relationship at higher levels.81 The increase in cancer deaths in the cohort studies largely disappeared during the early years of the studies. Neoplastic diseases such as carcinoma of the colon or prostate and leukaemia are well known to lower cholesterol;53 indeed, it may be prognostic. It may be concluded that much of the excess cancer mortality is associated with pre-existing subclinical cancer present at the time of entry into the trial. However, for some diseases such as lung cancer, there is a graded inverse relationship with cholesterol across its whole concentration range that appears to antedate any cholesterol-lowering effect of the cancer itself. This may be explained because the antecedent causes of both low cholesterol and cancer, suicide or accident may coexist well before the development of cancer. Examples are low body weight due to cigarette smoking, emphysema, mental illness such as depression, poor nutrition, alcoholism and low socioeconomic status leading to low cholesterol, but each of these is related to other causes of increased mortality. This is made likely by five lines of evidence. First, there is an important difference between the results from prospective studies in employed populations and those in whole communities.82
