Evaluation of dietary intake data using the tolerable upper intake levels.

We discuss the problem of assessing nutrient intake relative to the tolerable upper intake levels (UL) for the nutrient proposed by the Institute of Medicine and focus on 2 important topics: the estimation of usual nutrient intake distributions and the extent to which intakes above the UL can be considered risky. With the information that is currently available for most nutrients, it is not possible to estimate the proportion of individuals in a group with intakes that place them at risk. This is because the shape of the dose-response curve needed to carry out a risk assessment is unknown for most nutrients. Thus, intakes above UL cannot be declared to be unsafe. Intakes below the UL, however, are likely to pose no risk to individuals in the group. Because determining the proportion of individuals with intakes below the UL requires estimation of an upper-tail percentile of the intake distribution, the use of 1-d intake data or otherwise unadjusted intake data are likely to lead to severely biased estimates. It is important to remove within-individual variance in intakes from daily intakes so that the tails of the usual intake distribution are accurately estimated. Underreporting of the amount of nutrients consumed will tend to shift the estimated usual nutrient intake distribution downwards. In this case, the true proportion of individuals with intakes below the UL is likely to be overestimated.