Low serum total cholesterol concentrations and mortality in middle aged British men

To examine the relation between low serum total cholesterol concentrations and causes of mortality.Cohort study of men followed up for an average of 14.8 years (range 13.5-16.0 years).One general practice in each of 24 British towns.7735 men aged 40-59 at screening selected at random from the 24 general practices.Deaths from all causes, cardiovascular causes, cancer, and non-cardiovascular, non-cancer causes.During the mean follow up period of 14.8 years there were 1257 deaths from all causes, 640 cardiovascular deaths, 433 cancer deaths, and 184 deaths from other causes. Low serum cholesterol concentrations (< 4.8 mmol/l), present in 5% (n = 410) of the men, were associated with the highest mortality from all causes, largely due to a significant increase in cancer deaths (age adjusted relative risk 1.6 (95% confidence interval 1.1 to 2.3); < 4.8 v 4.8-5.9 mmol/l) and in other non-cardiovascular deaths (age adjusted relative risk 1.9 (1.1 to 3.1)). Low serum cholesterol concentration was associated with an increased prevalence of several diseases and indicators of ill health and with lifestyle characteristics such as smoking and heavy drinking. After adjustment for these factors in the multivariate analysis the increased risk for cancer was attenuated (relative risk 1.4 (0.9 to 2.0) and the inverse association with other non-cardiovascular, non-cancer causes was no longer significant (relative risk 1.5 (0.9 to 2.6); < 4.8 v 4.8-5.9 mmol/l). The excess risks of cancer and of other non-cardiovascular deaths were most pronounced in the first five years and became attenuated and non-significant with longer follow up. By contrast, the positive association between serum total cholesterol concentration and cardiovascular mortality was seen even after more than 10 years of follow up.The association between comparatively low serum total cholesterol concentrations and excess mortality seemed to be due to preclinical cancer and other non-cardiovascular diseases. This suggests that public health programmes encouraging lower average concentrations of serum total cholesterol are unlikely to be associated with increased cancer or other non-cardiovascular mortality.