Cardiovascular risk functions are regarded as the best tools for establishing priorities in primary prevention. Since the original Framingham risk chart fell into disuse because it greatly overestimated the real risk, the adjusted REGICOR and SCORE functions have become widely available in Spain, although the REGICOR function is the only one that has been validated for use in the Spanish population. Risk estimates have been shown to be useful for decision-making, particularly on the treatment of hypercholesterolemia. However, the fact that the majority of cardiovascular events occurs in individuals classified as being at a medium risk is evidence for the poor discriminative ability of classical risk factors. Despite the use of new parameters proposed for estimating cardiovascular risk, such as the C-reactive protein level, the detection of coronary calcification, the carotid intima-media thickness and the ankle-brachial index, there has been no improvement in the predictive capacity of classical risk factors. The most promising alternative seems to be the identification of "vulnerable patients" using markers of vulnerable plaque (ie, unstable or high-risk plaque), vulnerable blood (ie, with a tendency for thrombosis), and vulnerable myocardium (ie, electrically unstable or with a tendency for arrhythmia). In this article, we discuss whether the combined use of cardiovascular risk functions, novel risk markers and noninvasive tests can be effective in increasing the accuracy of patient selection for the primary prevention of cardiovascular disease.