Serum cholesterol and mortality in a Japanese-American population: the Honolulu Heart program.

The authors have examined the relation of baseline serum cholesterol level to subsequent 9-year mortality in a cohort of Japanese-American men. The baseline serum cholesterol level was positively related to coronary heart disease mortality. It was inversely related to total cancer mortality and to mortality from cancers of the esophagus, colon, liver and lung and to malignancies of the lymphatic and hematopoietic system. When mortality in the first two years after examination was removed from consideration in an attempt to allow for existing clinical or subclinical disease, the inverse relation to cancer persisted but was statistically significant only for colon cancer and lung cancer among the common sites as well as for malignancies of the lymphatic and hematopoietic system. When known prevalence cases of cancer were eliminated from the computation, the inverse relation between cholesterol level and cancer mortality persisted significantly only for colon among the common cancer sites. The relation of baseline serum cholesterol level to total mortality in this cohort was quadratic; that is, there was an excess of deaths associated with serum cholesterol level at the high end of the distribution (mainly due to coronary heart disease) and at the low end (mainly due to cancer).