Objective: To evaluate, at a population level, whether total cholesterol (TC) is a risk factor of mortality. To verify whether or not this is true for both genders.
Design: Population-based, long-lasting, prospective study.
Setting: Institutional epidemiology in primary care.
Subjects: A total of 3257 subjects aged 65-95 years, recruited from Italian general population.
Intervention: None.
Main outcome measures: Total cholesterol was measured, analysed as a continuous variable and then divided into quintiles and re-analysed. For each quintile, the multivariate relative risk (RR) of mortality adjusted for confounders was calculated in both genders. Stratification of mortality risk by TC quintiles, body mass index and cigarette smoking was also performed in both genders.
Results: Total cholesterol levels directly predicted coronary mortality in men [RR being in the fifth rather than in the first quintile: 2.40 (1.40-4.14)] and any other mortality in women. It also inversely predicted miscellaneous mortality in both genders. This trend was more evident when low cholesterol was associated with malnutrition or smoking.
Conclusions: High TC remains a strong risk factor for coronary mortality in elderly men. On the other hand, having a very low cholesterol level does not prolong survival in the elderly; on the contrary, low cholesterol predicts neoplastic mortality in women and any other noncardiovascular mortality in both genders.