Several studies have demonstrated the importance of hypercholesterolemia as a cardiovascular risk factor and a direct correlation between the reduction in cholesterolemia and the reduction in cardiovascular mortality in populations younger than 65 years. This correlation is controversial in the elderly and, particularly, in the oldest old. The aim of our study was to evaluate the total cholesterol in the oldest old and to assess the eventual presence of correlation between total cholesterol levels and mortality in a group of nondemented oldest old. A subsample of 40 subjects was extracted from the 103 subjects aged over 84 years living in Troina, a rural village in Sicily. We excluded all subjects under therapy with lipid-lowering drugs, demented, with malnutrition or affected by acute or chronic diseases which might cause death in the short term. At the end, 23 subjects (15 males and 8 females) were included in the study. After 2 years, mortality data of all subjects included in the study were obtained from official registers. The statistical analysis was performed by means of the X(2) test. In all subjects the mean of total cholesterol was of 182+/-32 mg/dl (mean+/-SD) and the body mass index was above 20; 17 subjects were in the normal range, 3 were moderately over-weighed and 3 were severely over-weighed. Overall, mortality rate after 2 years was 30% (7 subjects, 4 males and 3 females). We evaluated the relationship between mortality and 4 factors: sex, age, body mass index (BMI) and serum total cholesterol. Mortality was significantly correlated (p<0.002) only with a low level of total serum cholesterol<or=160 mg/dl). Our data are in agreement with those already available in the literature, and demonstrate that in the oldest old levels of total cholesterol lower than those of young subjects are found, but when the total cholesterol decreases below a hypothetical critical limit of 160 mg/dl represent, an important index of frailty in the oldest old and a predictor of mortality.