Background: Cardiovascular risk estimation in people over 70 years is problematic. Most scores have been derived from cohorts of middle-aged people, with older persons under-represented. The predictive power of classical cardiovascular risk factors reduce with age. The aim is to develop a specific score for the elderly populationMethods: Population-based cohort established in 1995. Setting: Three geographical areas of Spain (Madrid, Ávila and Lugo). Participants: 3,729 people older than 64 years, free of any cardiovascular diseases (CVD), at baseline. Measurements: Suspected fatal and nonfatal CVD (both coronary heart disease and stroke) were yearly investigated and confirmed using the WHO-MONICA criteria. All participants were followed until the occurrence of the first CVD event, until death or until December 31th 2015 if alive.Results: Age was the strongest predictor of 10-year CVD both in men and women. In men, variables associated to CVD were high blood pressure treatment (HR: 1,35; 95% CI: 1,067 , 1,710), diabetes (HR: 1,359; 95% CI: 0,997 , 1,852) and smoking (HR:1,207; 95%CI: 0,945 , 1,541), and in women smoking (HR: 1,881 95%CI: 1,356 , 2,609) and diabetes (HR:1,285; 95%CI:0,967 , 1,707). Total-cholesterol did not increased the risk of CVD either in males or females. However, total-cholesterol level (>200mg/dl) were negatively associated both in men and women.Conclusions: In old Spanish men, 10-year total CVD is significantly increased by age, diabetes and antihypertensive treatment, and in old women by diabetes and smoking. Total-cholesterol levels did not increased the risk of CVD either in males or females.