Cardiovascular diseases continue to constitute a major public health problem in all industrialized countries, where they are the main causes of premature mortality. There is a large body of evidence suggesting that free radical production directly or indirectly plays a major role in cellular processes implicated in atherosclerosis. Here we present mechanistic data and results of epidemiologic studies on the relationship between antioxidant vitamin intake or biochemical status and the risk of cardiovascular diseases. Most epidemiologic data obtained on this topic have been based on observational approaches, i.e. ecological studies, case-control or prospective studies. All these studies indicate that a high dietary intake or high blood concentrations of antioxidant vitamins are associated with a reduced risk of cardiovascular diseases. Although the results of these studies are convergent, they merely suggest a relationship at the population and individual level, but do not affirm a causality link. Only intervention studies (randomized trials), by specifically changing antioxidant vitamin intake, can provide conclusive answers. The apparent discrepancies between the results of recently published trials may be explained by the type of population (general or high-risk subjects), the differing doses of supplementation (nutritional levels or higher), the number of antioxidants tested (one, two or more) and the type of administration (alone or in balanced association). It thus appears that low risk of pathologies may be related to multiple nutrients consumed, at nutritional doses, and in combination. Optimal effects may be expected with a combination of nutrients at levels similar to those found in a healthy diet. A single antioxidant vitamin given at high doses in subjects with high risk of pathologies may not have substantial benefits and could even have negative consequences.