Multivariate prediction of coronary deaths in a 10 year follow-up of an Italian occupational male cohort.

3395 men aged 46-65 belonging to occupational groups in Rome were examined in 1980 for the measurement of a number of potential coronary risk factors and of other personal characteristics. They represented the 76.5% of the enrolled roster. In 2853 men free from previous major coronary events and with all the measurements available, 98 died from CHD in the next 10 years. The multivariate analysis performed by the Cox model showed the significant predictive role of age, systolic blood pressure, non-HDL cholesterol, blood glucose and heart rate. The relative risk (for a distance of 2 standard deviations from the mean) was of 2.3, 2.1, 1.4, 1.7 and 1.4 respectively. Smoking habits, body mass index, xanthelasma, leisure physical activity, serum triglycerides, uric acid and a stress score did not contribute to the prediction of coronary deaths. HDL cholesterol produced a negative but not significant coefficient. As compared to the previously available Italian risk functions the role of non-HDL cholesterol and of blood glucose represents the most remarkable contribution.