Serum cholesterol and mortality from all causes, cardiovascular disease, and cancer.: A cohort study in 220 000 individuals.

Serum cholesterol level distribution and associations with all cause, cardiovascular, and cancer mortality were studied in 125 513 men and 96 301 women aged 16 to 90 years. Serum cholesterol increased with age until 50-60 years in men and 60-70 years in women and decreased afterward, very significantly in men and moderately in women. A U-shaped relation was found between age-adjusted total serum cholesterol and all cause mortality. In men, cardiovascular mortality increased significantly with age-adjusted serum cholesterol, even after adjustment for body mass index, blood pressure, smoking history, and gamma-GT; in women, the relation was not significant. Associations between serum cholesterol and mortality in men were significantly confounded by weight, body mass index, vital capacity, forced expiratory volume in one second, smoking history, serum urea and albumin, hematocrit, hemoglobin, alkaline phosphatase, gamma-GT, mean corpuscular volume, and erythrocyte sedimentation rate. Age had a very substantial influence. Thus, serum cholesterol in men younger than 50 years at baseline showed a trend toward a positive association with all cause mortality and a highly significant positive correlation with cardiovascular mortality; conversely, both correlations were negative in men older than 65 years. Nonsignificant trends in the same direction were found for mortality from cancer. These differences were largely ascribable to the higher prevalences among older subjects of malnutrition, poor general health, and patent or silent morbidity (e.g., liver disease). Thus, separating age groups and causes of mortality is essential when studying correlations between serum cholesterol and mortality. The data reported here suggest that low serum cholesterol may be a susceptibility marker for, or a consequence of, cancer rather than a causative factor.