A single risk factor measurement predicts 35-year mortality from cardiovascular disease

Two Italian rural cohorts of men aged 40-59 years, were examined in 1960 within the Seven Countries Study of Cardiovascular Diseases and a total of 1712 men were enrolled (participation rate 98.8%). Cardiovascular risk factors were measured and 35-year follow-up made for vital status, mortality and cause of death. Cardiovascular diseases represented the first cause of death (46.2%), cancer the second (29.9%). The association between risk factors measured at baseline and the occurrence of cardiovascular deaths was tested by the use of multivariate functions (proportional hazards model in particular) which predict an event as a function of many possible factors. The predicted fatal events were, among men initially free of cardiovascular disease, coronary heart disease-restricted criteria (CHD-RC), coronary heart disease-broad criteria (CHD-BC), strokes (STR), and all cardiovascular diseases (CVD). The predicting variables were 21 risk factors of different nature. All models were highly discriminant between cases and non-cases. The predictivity of risk factors was assessed by testing the statistical significance of their multivariate coefficients, and by computing relative risks (expressed as hazards ratios) for standard differences in their levels. Age and systolic blood pressure produced significant coefficients and large hazards ratios in solutions for all end-points; cholesterol and cigarette smoking in three (not for STR); vital capacity (inverse relationship) and gerontoxon in two; physical activity (inverse relationship), forced expiratory volume (inverse relationship), urine glucose, family history of heart attack, and xanthelasma in one each. Marital status, family history of hypertension or diabetes, body mass index, skinfold thickness, arm circumference, shoulder-pelvis shape, laterality-linearity index, trunk-height ratio, and heart rate never provided a significant contribution to prediction. As an example, a difference of 20 mmHg in systolic blood pressure corresponds to a relative risk (excess risk) of 1.50 for CHD-RC, 1.46 for CHD-BC, 1.42 for STR and 1.43 for CVD; a difference of 40 mg/dl of serum cholesterol corresponds to relative risks of 1.38, 1.33, 1.13 and 1.25 respectively for the four end-points; a difference of 10 cigarettes smoked per day corresponds to relative risks of 1.19, 1.21, 1.06 and 1.17 respectively for the four end-points. The findings indicate that some cardiovascular risk factors measured once in middle age retain a long term association with prediction of future cardiovascular events, up to 35 years follow-up.