Long-term prediction of coronary heart disease mortality in two rural Greek populations.
In 1960-61 two pooled Greek rural populations totalling 1215 men aged 40-59 years were followed-up for 25 years. A Cox model analysis of fatal coronary events over 15 years showed that serum cholesterol in men aged 40-59 years, cholesterol in men aged 45-64 years, and systolic blood pressure in men aged 50-69 played a predictive role. The coefficient of age became more significant with advancing age and that of cigarette smoking only at 25 years follow-up. The coefficient of cholesterol decreased stepwise and became negative for men aged 50-69; body mass index was without effect in any follow-up of these cohorts. Systolic blood pressure and serum cholesterol increased in these populations by 5.4 mmHg and 23.5 mg.dl-1 (0.61 mmol.l-1), respectively between the years 0 and 10, whereas cigarette consumption decreased minimally. These changes were used to test the predictability of coronary events occurring between years 10 and 25 of follow-up when added to the model containing the factors at entry. Of these changes only systolic blood pressure significantly increased the predictability of coronary deaths. It is concluded that even minor alterations in systolic blood pressure above or below the entry levels can be associated with marked modifications in coronary mortality above or below those occurring naturally in the 15 years after the changes occurred.