[Relative validity of the 10-year cardiovascular risk estimate in a population cohort of the REGICOR study].