It's a good thing all possible CVD risk factors were measured without measurement error, included in the analysis, and had exactly correct coefficients despite sampling error.
It's a good thing observational studies like these are not notorious for never replicating and for motivating enormously costly (financially, medically, and personally) fiascos such as the wars on salt and cholesterol.