BACKGROUND:
This study aims at identifying determinants of all-cause mortality in elderly populations of different countries.
METHODS:
Men ages 65-84 years from defined administrative areas were enrolled in Finland (rural areas of east and west Finland; N = 693), in the Netherlands (the town of Zutphen; N = 851), and in Italy (the rural areas of Crevalcore and Montegiorgio; N = 682). They were survivors of cohorts studies for 25 years within the Seven Countries Study with the addition of a subgroup of the same ages in the Netherlands.
RESULTS:
Five-year death rates from all causes were higher in Finland (297 per 1000), intermediate in the Netherlands (231 per 1000), and lower in Italy (191 per 1000). Five-year all-cause mortality was studied in relation to measurements taken at entry (age, systolic and diastolic blood pressure, non-high-density lipoprotein (HDL) and HDL cholesterol, body mass index, heart rate, smoking habits, and presence of coronary heart disease manifestations). Univariate and multivariate analyses were performed (in the latter models, both linear and quadratic terms were used for most variables) with all-cause mortality as endpoint. Results suggested significant predictive power of age (direct relationship) and, in most cases, U-shaped relationships of risk factors to mortality. Non-HDL cholesterol showed significant relationships with mortality in Finland and the Netherlands, HDL cholesterol in all three countries, systolic blood pressure only in Finland, body mass index in Finland and the Netherlands, smoking habits only in Finland, and heart rate in none. Levels of risk factors associated with the lowest death rate in the pool of all countries were 183.3 mg/dl for non-HDL cholesterol, 59.8 for HDL cholesterol, 177.5 mm Hg for blood pressure, and 30.2 kg/m square for body mass index.
CONCLUSIONS:
In these elderly men the association of traditional risk factors with all-cause mortality is reduced, U-shaped, or even inverted. This is probably due to selection due to previous mortality, to comorbidity, and to changes in homeostatic mechanisms.