Objectives: To explore age at death (AD), overall and for different causes of death, in a cohort followed up to quasi-extinction.
Materials and methods: In 1960, in the Italian Rural Areas of the Seven Countries Study, 1712 men aged 40-59 years were enrolled, examined and then followed up for 50 years. AD was computed for all fatal events and compared across 12 groups of causes of death. Multiple linear regression model was used to estimate AD and Cox model to predict each of the 12 causes of death, as a function of 28 selected baseline risk factors.
Results: After 50 years, 97.5% of men had died. Mean AD was 75.0 years (median 76), while large variation was found across the 12 causes of death, with the highest levels for Heart Disease of Uncertain Etiology (HDUE) and Senility plus Causes Unknown (SNUNK), having means of 79.2 and 84.5 (median of 80 and 86) years, respectively. Many risk factors were directly associated with overall AD, the most significant being subscapular skinfold, arm circumference, Mediterranean diet, age at baseline examination, never smokers and vigorous physical activity. Systolic blood pressure (SBP) was inversely related. The relevant modifiable risk factors predicting single causes of death were SBP and the lifestyle behaviors of dietary, motion and smoking habits.
Conclusions: AD proved to be a useful indicator of previous health and aging of populations. HDUE and SNUNK seem the most "physiological" causes of death. SBP and lifestyle risk factors are the most relevant characteristics associated with AD.
Keywords: Age at death; Causes of death; Longevity; Prediction; Risk factors.