Micronutrients (vitamins and minerals) as cancer-preventive agents.

Reduction of incidence and mortality of common cancers requires successful preventive interventions both in high-risk populations and in the population at large. Agents chosen for intervention trials must have been demonstrated to have significant and plausible biological activity in animal, in vitro, and short-term human studies, and these should be supported by observational epidemiological analyses. This chapter reviews evidence about vitamins and minerals as possible cancer-protective agents, emphasizing carotenoids (notably beta-carotene), natural vitamin A, vitamin E, selenium, vitamin C, calcium, vitamin D and folic acid. It is helpful to group agents according to apparent primary mechanism of effect, such as antioxidant (and/or pro-oxidant) effects for beta-carotene, vitamin C, vitamin E, selenium, zinc, copper, iron, manganese, phenolic compounds and isothiocyanates. Certain conditions associated with increased concentrations of free radicals and reactive oxygen species-like smoking, aging and reduced glutathione levels-must be assessed as confounders. However, these chemicals generally have multiple effects, which also may complicate reliance on intermediate end points. To choose target populations, dosages and durations for interventions, it is essential to distinguish between overcoming a relative deficiency of the micronutrient in the diet and the circulation and providing a pharmacological dose. Selenium and beta-carotene are important examples of agents that may be protective at the high end of the normal diet-based distribution of serum values, but may be toxic-somehow-at much higher levels. We are becoming more aware that it is a long leap from associations observed or deduced in epidemiological studies to the design and demonstration of successful chemopreventive interventions with single agents or with combinations or mixtures. When interventions fail to support the hypotheses from observational epidemiology, it may be necessary to reconsider the hypotheses. Furthermore, effects may vary remarkably at different tumour sites.