The use of epidemiologic approaches and meta-analysis to determine mineral element requirements.

With a new focus on long-term health effects of nutrient intake as well as the traditional concerns of preventing deficiency and maintaining tissue saturation, epidemiologic studies which identify diet and health relationships have become important to debates on defining nutrient requirements. Epidemiologic studies include descriptive surveys, case-control studies, cohort studies, and intervention studies. National dietary intake survey data have always played a role in defining intake amounts which are feasible and which are associated with no clearly apparent health problem. A growing number of large epidemiologic cohort studies have provided new evidence of preventive effects of specific nutrients on chronic diseases. The question of how that information should be used in setting requirements is still unresolved. Because one epidemiologic study alone cannot prove a causal relationship, and because small studies often lack sufficient power to clearly establish statistical significance, meta-analysis is sometimes used to combine results from many studies to obtain an overall quantitative result. It has been most effectively used with sets of clinical trials. Meta-analysis of multiple observational studies is usually difficult because of variations in design, methodology of data collection and form of the variables; its use has been seriously challenged. Careful review of the totality of epidemiologic studies completed on specific nutrient and health or functional outcome is important to defining future research priorities and designs, and can inform the definition of nutrient requirements when used carefully and in conjunction with the wider body of scientific evidence.