a1 U557 Inserm (UMR Inserm/Inra/CNAM), Institut Scientifique et Technique de la Nutrition et de l'Alimentation/CNAM, 5 rue Vertbois, F-75003 Paris, France
a2 Unité de Surveillance et d'Epidémiologie Nutritionnelle (USEN), InVS/CNAM, Paris, France
A voluminous body of epidemiological research concerning the potential role of antioxidant nutrients in the prevention of cancers has accumulated over the past few decades. However, results of large recent intervention trials do not support a preventive effect against cancer for supplementation with antioxidant nutrients. Seemingly contradictory results between observational studies and randomised trials can be explained by the fact that doses used in clinical trials were much higher than the highest levels attained by the usual dietary intake which, in observational studies, were found to be associated with the lowest risk of cancer. Recently, the Supplementation en Vitamines et Mine´raux Antioxydants (SU.VI.MAX) study, a randomised, double-blind, placebo-controlled primary prevention trial, tested the efficacy of supplementation with a combination of antioxidant vitamins and minerals, at nutritional doses, in reducing the cancer incidence in a general population not selected for risk factors. After 7·5 years, low-dose antioxidant supplementation lowered the total cancer incidence in men only. This may be explained by a lower baseline status of certain antioxidants in men compared to women. Finally, the effect of antioxidant supplementation on the incidence of cancer could depend on baseline antioxidant status (which differs from gender and/or nutritional status) and the health status of subjects (healthy v. cancer high-risk subjects). Antioxidant supplementation may have a beneficial effect upon cancer incidence only in healthy subjects who are not exposed to cancer risk and who have a particularly low baseline status. High doses of antioxidant supplementation may be deleterious in subjects in whom the initial phase of cancer development has already started, and they could be ineffective in well-nourished subjects with adequate antioxidant status.