base a UL. This indicates a need for caution. It does not mean that high intakes pose no risk of adverse effects.

The major focus of the development of EARs and AIs has been the determination of the most appropriate indicator of adequacy, followed by the derivation, from available data, of the EAR or AI. A key question is “Adequate for what?” In many cases, a continuum of benefits can be ascribed to various levels of intake of the same nutrient. A specified marker or indicator may be deemed the most appropriate to determine risk of deficiency for a nutrient, while another indicator may be the best marker in determining risk of chronic degenerative disease for that nutrient.

Each EAR or AI is described in terms of the selected criterion or outcome. For example, the dietary intake recommended as the AI for vitamin D for older adults (> 70 years) is based on both a biochemical marker (circulating 25-hydroxyvitamin D) and a functional outcome marker (reduced fractures and bone loss). Using data from clinical studies, an intake of vitamin D associated with normal circulating 25-hydroxyvitamin D concentrations was derived. This intake was supported by clinical trials in which supplemental calcium and vitamin D were associated with a reduced risk of fracture over three years and a reduction in loss of bone mineral density at specific bone sites. Thus, two sets of data form the basis for the AI. Since the individual contributions of the added calcium and vitamin D to the attenuation of bone loss cannot be evaluated, an AI was established. Whether these higher intakes of vitamin D at younger ages will reduce risk of osteoporosis and fracture in later life remains to be determined.

Imbedded in the framework of DRIs is the following approach. When requirements are estimated to decrease risk of disease, particularly chronic degenerative disease where associations may not be easily identified in short-term studies, there must be a preponderance of epidemiologic evidence that is supported by clinical trials and biologically plausible mechanisms before such associations are used to establish recommendations. Given that chronic degenerative diseases and developmental abnormalities may not be detectable for significant periods of time, it is quite possible that individuals who have increased risk due to diet may not be identifiable,