LAUSR

Abstract Objectives: To describe age at death (AD) in ten cohorts of middle-aged men followed-up until nearly extinction as a function of personal characteristics measured at baseline. Material and Methods: Cohorts of men aged 40–59 years in five European countries (Finland, the Netherlands, Italy, Serbia, Greece) were examined and then followed-up for mortality during 45 (3 cohorts) or 50 years (7 cohorts). Multiple linear regression (MLR) models were computed with AD as dependent variable and 35 personal characteristics as independent variables. Results: Out of 7047 men, 6798 (96.5%) died in 45–50 years. Average AD was 74.4 (median 75) years. Only five variables were significant in all the five countries when tested by MLR. By pooling together all countries and selecting variables by a stepwise technique, 20 were significant (p < .05): age at entry examination, being married, being a never smoker or an ex-smoker (versus current smokers), forced expiratory volume, high socio-economic status, arm circumference, subscapular skinfold, having moderate or vigorous physical activity (all with a positive relationship with AD); father early mortality, laterality/linearity index, high systolic blood pressure, high serum cholesterol, having a major cardiovascular disease, diabetes or chronic bronchitis (these with a negative relationship), while body mass index, subdivided into four classes, showed a parabolic relationship (versus a reference class). Each factor was associated with a difference of months or a few years to AD, while their combinations were associated with several years of greater or smaller AD. Conclusions: The availability of a relatively small number of risk factors measured in middle-aged men allows making rough estimates of AD as related to all-cause mortality during a follow-up of 50 years. AD is also a valuable metrics to describe past health in nearly extinct populations. Key messages We aimed at studying the age at death (AD) of middle-aged men of the European cohorts of the Seven Countries Study, followed-up 45–50 years until nearly extinction, as a function of personal characteristics measured at baseline. Multiple linear regression (MLR) models were computed with AD as dependent variable and 35 personal characteristics or risk factors as independent variables and only five were significant in all individual countries, whereas pooling together all countries, 20 variables were significant as selected by stepwise MLR. Each factor was associated with differences of months or a few years to AD, while their combinations were associated with several years of greater or smaller AD. AD is a valuable metrics to describe past health in nearly extinct populations: it might also be used to communicate years of life gained by preventive measures on modifiable variables.