Cardiovascular risk factors predicting all causes of death in an occupational population sample.

A group of 768 men aged 40-59 at entry examination and belonging to an occupational sample of railroad employees in Rome have been examined for the measurement of some risk factors and followed-up for 20 years. In all 676 men, free from life-threatening diseases and with all measurements available, produced 166 fatal events in 20 years. Out of the 27 different personal characteristics considered only six contributed significantly to the multivariate prediction of all causes of death in the Cox proportional hazards computed by the forward stepwise technique. The factors predicting all causes of death were age, cigarette smoking, diabetes, blood pressure, mother's vital status and being on a diet prescribed by a doctor. The relative risk of those located in the upper decile of the estimated risk as compared to the bottom decile was 8.2. The results do not differ much from those obtained in a demographic sample studied in the same way.